Hey everyone, thanks for all the comments. There's been a couple recurring comments so I'd like to go over a clarification and correction here. The first is a clarification on what I meant by "exposed" blocks. For me, there are two types of blocks which I'll call costly and free. Costly blocks are blocks that I have to spend time/effort on. So if I'm calculating exposed blocks for various methods, I want to know how many free blocks each method gives me. Basically, how many bonus blocks do I get to see that I didn't have to spend time/effort touching. I explicitly excluded blocks that cost me, or blocks that I have to mine anyway. This results in the 4:1 ratio for the 1x1 and 3:1 ratio for the 2x1 tunnels I got in the video. A few people have commented that the 1x1 is actually a 5:1 ratio because you are exposing the block that you will mine next in the tunnel. They've explained their view well and I totally agree that the 5:1 ratio is valid based on what they count as 'exposed'. I just should have done a better job at explaining what exactly I was calculating so I appreciate people pointing this out. The overall trend of the results are the same, so feel free to use whichever makes more sense for you. The second is a correction on the limits shown in the video. When talking about how many exposed blocks there are per block mined, I took the limits of (4x+1) and (3x+1), which are only the "exposed blocks' part and by themselves go to infinity. I forgot to put them over the 'per block mined' part, x, which brings them back down to 4 and 3. So the correct limits are lim((4x+1) / x) and lim((3x+1) / x). Thanks for catching that!
Somebody probably pointed that out already but did you weight in exposing double on the layers you're actually mining? (When you mine a 2x1 corridor you exposed two wall blocks and 1 ceiling or 1 floor.
What about a 1x1 method where you mine your first row at Y-57 then with one block gap between the exposed blocks / glass but at Y-56 This would result in a method with a spacing between the 1 and 2 block gaps which might be more efficient
@@jackweslycamacho8982 That is the result, yes, but the derivative doesn't actually explain why it matters, while proposing it as a limit actually translates to the more you mine the closer you will get to this ratio, while it is identical, it's not pedagogically equivalent.
The seconds/block exposed metric doesn't actually matter, unless you go for the torches in which case the 1x1 tunnel is superior because you don't need torches because nothing can spawn. But most players just use allbright or gamma=100 and wouldn't bother with the mobs, and considering you can constantly mine regardless of what method you are using(with the branching ones there is a slight inefficiency if you are not paying full attention and skipping the blocks with perfect tick skipping, essentially breaking every second block on your way looking out of the branch, to ensure you aren't missing any breaking ticks by turning, but most players have this slight inefficiency even without branches)
@@DanielLCarrier if you are making a tunnel, you are going to mine that fifth block even if it wasn't ore, so it would be considered blocks mined instead of blocks exposed.
@@Big_boie It's both. Imagine instead of using any of these efficient mining techniques, you just mine out a solid cube. Using your method, the limit is 6x^(2/3)/x, which approaches zero. And yet the diamonds you find goes with the volume you mine, not the surface area.
Something that might be important is that in a 1x1 tunnel, you'll be moving more slowly. This matters for the return trip, after you're done mining, and it would matter earlier if you could mine blocks faster than you could move to reach new ones
That’s a good point, but for me I just mind a tunnel back lol so if I mine for 20 minutes I’ll go 10 out and 10 back. Otherwise, yeah I’d agree it would waste a lot of time
Hi! Nice video, but in the end you are overlooking an important thing: You said that the standard deviations don't matter, but actually, they are extremely important in the interpretation of your results in this case. They strongly indicate that your results are not significant because the variation (your error) is too high. The testing area of 100x100 is just too small to account for the huge deviations between locations, particularly because diamonds are so rare. 75 seconds per diamond per layer is way within 1 standard deviation of all your results for the 1x1 tunnels, actually within half a standard deviation. Assuming normal distributed diamond amounts per layer, for example, this means that the hypothesis that the average time to mine for a diamond for all 1x1 tunnel techniques is 75 seconds cannot be rejected with more than 38% confidence ( half a standard deviation) for all the techniques, or in other words based on these results its more likely that all techniques yield the same result than not. If anything, than this says that the techniques don't matter unless you go on a very large scale, and luck plays a much more important role. But as others have pointed out, we could try to extend the efficiency calculation by adding in the time to go back to base, which means that more dense techniques or ones where you can walk or run back more quickly will have an advantage, and try to write some scripts to automate the exposed diamond calculations for larger areas to get more accurate results. Thanks for providing great ground work here!
i think the running back to base thing is also an interesting dilemma. With a 1x1 tall mining, it would take forever to run all the way back to base, so i turn around when my inventory is halfway full. that way the effective time to get back to base is 0s, since half of my mine is away, and half is towards. Also late game when I have shulker boxes, i find it easier to just stache everything in the boxes, and mine for well over 1000 blocks, then to get back i just mine straight up and fly home with an elytra (or /kill if I have keepinventory on). flying home with an elytra will be way faster if im going super far away, but slower if I'm mining less, because it would take a while to resurface. I think a more in depth video on this topic would be really interesting.
Great point and I definitely agree. There's just no way I'm doing any more manual testing for this lol But I am planning on doing a follow up code-based video so we can get a much larger sample size for more accurate data. I also want to test a lot more methods (maybe even cave mining but that might deserve it's own video). Lastly, I didn't include return time for any of the methods tested as I assumed the player would be mining a tunnel back. So if the 1x1 does end up better, I'd always recommend snaking instead of backtracking.
In the newer versions of Minecraft, strip mining techniques are completely irrelevant. Deepslate is far too slow to mine, even with a haste beacon and Eff 5 pick. The fastest way to get diamonds is to drink a night vision potion and fly around a cave, or swim around a submerged cave. You'll have way more fun, find other cool stuff along the way, and get a crap load of diamonds.
Yeah I agree. I mentioned this near the end, but I chose to test out strip mining cause that's how I like to mine for diamonds lol. It's a lot more chill for me that way. But I have been looking at ways to test cave mining, so I might do a video about that in the future. Let me know if you've got any ideas for that!
that has the downside of needing to deal with mobs, strip mining you can pretty much just zone out/autopilot and if your world is pre-1.17 swamp clay diamond finding is always the best, if a bit cheesy
@@ThylineTheGay I agree, strip mining is just a good way to chill and play. I hadn’t even considered people playing on pre 1.18 versions, good point. But I think it’d still be an interesting challenge to find a way to accurately test cave mining. Implementing mob interactions would be difficult unless I just manually cave mined for hours and hours but that doesn’t seem like a sustainable method lol
I feel I am usually quite unlucky caving. Someone should probably do a "seconds per diamond" test run for that... also its a matter of how far you are in the game, quite early you might not have access to enough gear and potions to do that comfortably.
17:38 When you said "I appreciate it I know its been a pretty long one." I looked at the time and realised the length of the video but was so enthralled (because maths and minecraft) that I said out loud "I appreciate you too" and my wife gave me a funny look. Thanks for the video, all the work and effort you've put in ( and you made your excel sheets nicely formatted too 👍 )
The way I saved this thinking it was about real life mining - something I’m very interested in - started watching, realised it was about Minecraft, and then continued to watch because it also collides with my interest in Minecraft
One thing I was confused by until I went back to double check is the terminology "no spacing", "1 block spacing", "2 block spacing", etc. I usually do tunnels with 3 blocks in between them which seems the best but showed terrible stats, but then the "no spacing" idea made no sense at all, since no spacing would just be mining the whole layer. Your spacing though is referring to blocks between exposed blocks, so "1 block spacing" is the method I usually use with 3 blocks between tunnels and it is the best. While this might make sense with the way you were showing the spacing with the glass and the wool, I think referring to it by how many blocks are between each tunnel is more clear since that is what people will be counting when mining.
I have to say. Very well made and explained video. I'm in the middle of developing my Masters Thesis and while watching this I realized that here you've essentially done a thesis on diamond mining. You're solving the problem of people not knowing how to best find diamonds, you present your research questions, you explain your research and the metrics you used to prove said research in your experiments, and you answer the mentioned questions throughout the video, before finally, showing your results and conclusions. Its funny to think about how anyone working on a thesis should unironically watch this video as reference. Amazing job.
If you're developing your masters, I challenge you to find the critical error in the calculations being done on screen at 19:15. Hint: Think about what the standard deviations mean and why them being so high implies something is wrong. My other comments reveal the answer, so don't spoil yourself!
standard deviation is still interesting, because it's how consistently you're finding diamonds, if you have a method with a huge sd that would mean you find a lot of diamonds at once, but sparsely, so you could find no diamond vein for a while and then multiple back to back, while if you had no deviation you could basically put a timer and you'll be sure you will get diamonds every time. I'd be interested in how other methods would compare, like hilbert curve and such. also I still like 2 high farming because it's easier to get around, I think I might switch to having like 9 rows of 1 high, and a row of 2 high. though an advantage of 1 high is that you don't need torches to avoid monsters. If I have the motivation I might make a script that calculates the actual efficiency for a given method.
Also sd matters A LOT. The 1 block gap method had the lowest sd, which means we can trust the number from that one the most. The rest were really inconsistent, which means with a small data set like this, the numbers probably aren't accurate. 2 block gap could be better than 1-block gap, but we will never know because it has such a large standard deviation and a small data set.
@@simonwillover4175 yeah, I'm currently looking and the generation, according to a modder there's : - 7 veins per chunk with a size of 1-5 diamonds. - 1 vein per chunk (1/9 chance) with a size of 1-23 diamonds. - 4 veins per chunk with a size of 1-10 diamonds. and with 1.20.2 added: - 2 veins per chunk with a size of 1-10 diamonds. The first and last vein type has a 50% chance to discard if exposed to air, the second vein type has a 70% chance to do so, and the third vein type has a 100% chance to do so. which means there's probably a more efficient conditional way to farm (using a chunk's expected yield, if you got more diamonds than average you probably should skip the chunk.)
@@satibel Yeah I’ve been trying to think of ways to program this for more data, mainly to test cave mining easily, so let me know how it goes! I definitely want to get those standard deviations down in a future video
@@SeaJayPlays my idea is to pre gen like 10k chunks for each biome then convert the world with amulet and write a custom script that iterates my mining functions. Desert should yield slightly more diamonds because of fossils.
The problem with 1x1 block mining is that your limit isn't your pickaxe, it's how fast you can crawl. This didn't come up in your testing because you were using an unenchanted diamond pickaxe, which doesn't mine faster than you crawl, but most of the time I imagine mining with Eff 5, which does. It also takes much longer to get back home once you're done, but that's a hard thing to include in testing
@@sorin_markov this is a good point and I’ll have to test out enchantments in the next one. As for the return time, I’d recommend always mining a tunnel back so you don’t have to spend time backtracking. I probably should have mentioned that lol
@@SeaJayPlays Mining back requires inventory space, pick durability, and torches. I usually stop mining and make a return trip because one of those has run out.
I wrote a datapack a while back to test this, basically just simulating different methods of mining physically as if I were actually mining myself. Glad to see our results matched up in the end, but what really would be more interesting now would be comparing it to caving, if only caves were more consistant
18:26 your expected time to mine and actual time to mine deviate so much becuase you forgot to take into account the 0.3s delay between mining 2 blocks. Your expected time to mine should be 50% higher. Amazing vid though!
@@eladnlg I’ll have to double check how I calculated it, but that sounds right. Thanks for pointing that out! If I do a follow up video I’ll make sure to correct that
Though you can circumvent that one tick focus delay by rebinding the mine block button from left mouse to a keyboard key. Useful for digging long straight tunnels and would further improve your efficiency.
One thing you didn't account for is the return to base time, when you mine one long tunnel you'll take longer to return, would be nice to see if it ends up changing the results
I didn’t account for return time cause when I mine diamonds I just mine a return tunnel back. So if I mine for 20 minutes I’ll go 10 out and 10 back. A few people have mentioned this now, so in the follow up I’ll make sure to cover it. Thanks!
Yesterday I used a simple TNT duping quarry for about an hour and a half and got 2 stacks of diamond ore, with fortune III it ended up giving me 426 diamonds. It also only took about 10 minutes to build the machine
@@tfk_001 I’ve got a pinned comment where I make the correction to what I meant to do, but the equations should have been over x corresponding to how we are calculating exposed blocks per block mined
Nice video. I loved to hear your process and the sample size was way larger than I expected. One critique I'd offer. I believe when 1x1 mining, if your pickaxe has high enough effeciency, your crawling movement speed starts to throttle how fast you can dig. . . This is a small, but possibly meaningful difference between the 1x1 and 2x1 methods as a whole. I would love to see if that upsets your scales if you are late game mining with an effeciency-5 pick.
@@benjaminsipe9217 Thank you! This is a good point, and I think someone did mention it early on. I’d say that based on just how much better (efficiency-wise) the 1x1 is over the 2x1, my guess would be that the crawl speed wouldn’t be enough to change which is faster. But I would like to do an even more thorough testing between these two methods to really see. A lot of in-game factors could play a part like placing torches, placing trapdoors, actually mining the diamonds, running into lava, mobs, food, and I’m sure others. I’m not sure how speed potions affect crawling, but that could be a solution (especially end-game) if it does end up making the 1x1 slower overall.
Practical Tip: If you are worried about the time to travel back to your starting point after a 1x1 mining session, when you decide you are halfway done mining, make a u turn (including that 3 block width). You'll end where you started, while not having to waste hardly any time crawling through tunnels.
@@dtplayers I honestly thought this was common practice, but based on the comments, see that I should have brought it up. Thanks for the good explanation!
I can’t describe how happy I am to find a channel that takes an analytic approach to farm / mining. I’ve built and optimized some smaller farms which is super fun. It’s not a game unless I’m breaking out excel😅
So I'm guessing since Y-56 and Y-57 have the most diamonds, and 1x1 with 1 block gap is the most efficient mining technique, it could be good to stack your Y-56 tunnels on top of the 1 block gap in Y-57 Maybe by using this technique it would even be better to have a 3 block gap
I just read comments that video felt so short and then I realise that it is actually over 20 minutes!!! Wow, it was extreamly interesting, well organized and useful! Looking forward for more such type of content!
Good stuff! I had done some crappy tests back in the day and found out that branch mining with a spacing of 6 (or is it 4? 6 blocks in between holes) and digging 100 long tunnels of 2x1 yielded nice results. I tested it by replacing common blocks with air to simulate digging and then manually counted the diamonds on the way.
@@bs_blackscout Yeah I had always used that method from Xisumavoid’s video, so I was surprised too. I really didn’t want the 1x1 to win lol That’s pretty similar to what I did, and in the future I want to find a way to automate it to so I can get more data
@@SeaJayPlaysthat Amulet tool is written in python and it's open source... you can use their library to open a world and program in a mining method to see how it would fair over the long term without it actually taking a long term
I mine 2 tall tunnels because I want the deepslate for building. The diamonds, for me, are just a bonus that I mine around and leave in place unless I have a silk or fortune pickaxe with me. If I have a haste beacon I'm going to be mining the entire circle of affected area. This stuff is good to know though.
Worth the wait? YES! I *LOVE* these mathematical experiments as done with Minecraft. When I took my Computer Architecture course, I had a BLAST figuring out how to translate on-paper math into Minecraft. This is so much more fun. I'd love to see more content like this. New sub, man!
Tiny problem: if you mine in a 1×1 tunnel you still need to return the same way you mined, meaning you need to painfully stare at your screen while crawling back to your base
Nah, just only mine halfway(whatever that means to you, half full inventory, half your torches are gone, half your food, whatever) then turn around and go three blocks over and mine your next strip back the way you came.
3:40 I think this isn't quite accurate. You should probably have weighted your sum based on the fact that you're exposing two faces each for layers ~ and ~-1, and only one face each for ~-2 and ~1
Edit: I forgot some crazy people mine one-high. So in that case it's 2 for ~ and 1 each for ~1 and ~-1. I don't know if that's what you did, but you didn't say that's what you were doing
@@noway2831 Yeah, it’s been so long I don’t remember how I did it. So I’ll have to go back and look at how I calculated the stuff to give a good reply. But thanks for brining this up cause I want to do a better/more thorough follow up based on all the comments so far
i've noticed you're saying ~1 for the block above your head and ~-2 for the block below your feet, but that doesn't work anymore because you're position is now calculated from where you're feet are, rather than from where your head is like it used to be. i'm not sure why they changed that but ~-1 is now the block under your feet and ~2 is the block above your head
I think it is still accurate because this isn't how many exposed diamonds, but rather which layer has the most raw diamond output. So this specific calculation is essentially if you mined out the entire layer, you'd fully see the layer you mined, as well as the above and below layer. The exposed diamond test came later where he tested the gaps, but this one was just because the largest total amounts of diamonds are in those three layers combined, giving the best likelihood of running into diamonds there as opposed to other layers, not other mining styles.
Great video! Only thing I’d like to see added is a test on spacing the 1x1 with height variance. Like a wine-rack spacing on the 3 block spacing at levels -56 and -58 or something.
"Wine racking" should definitely be the word for this. The concept makes sense, and I always do it. I'm just not sure if it actually works out given the width of the best diamond density depths, and the size of veins. Would love to see it tested
@@theswankypotato This is a really interesting idea and I’d never heard of it or seen it done before. So I’ll make sure to test it out in the follow up video!
Hey, this is so underrated. A Video which would be loved by every single minecraft player but sadly only 22K Views. Great Work tho! Thanks for calculating everything, i think its very interesting to know the maths behind mining diamonds. One other point: I actually always mined a 2x1 tunnel with 1x1 branches going off of them. Surely a bit more complicated to calculate but i would be interested to know the results behind this method. Keep up the great work! Thanks from Germany!
@@etwasflauschiges Thank you! I’m planning on doing a more in depth follow up video in the future (code based this time) so hopefully I can test a lot more methods
Maybe you will bring it up but at 17:14 I see the Error about the "hidden diamonds" that have been kept, as these type of diamonds require more effort, they need to be treated as a separate ground. We essentially have 3 groups of diamonds: visible (easy), hidden (more effort), invalid (0). Adjusting this could make a difference in the general methods depending how the hidden to visible diamonds ratio changes. The effort of the hidden blocks need to be statistically proven too, by how many blocks on average you would need to dig out to expose the hidden ones in the most efficient way, which may also mean that there is a method that would allow some hidden diamonds to be missed. Anyway, we could then compare if the effort of searching for hidden diamonds is itself worth it, or if it is best to not go for them at all. Lots of calculations..
I absolutely love these kinds of scientific/statistical approaches to video games. Very fun! One thing I'd suggest for your future vid: do an actual trial run of these tunnels in survival. I think there's a lot of small details that start to add up that may change the efficiency of each tunnel. One being the time to return that others have mentioned. Another being the collection of the diamonds themselves. Wouldn't you stand up every time you collect 2-tall diamonds, and then have to recrawl to get back in the tunnel?
I think it's nice to demonstrate how real mining times will always be slower than simply calculated time to mine. 2x1s should even be more than twice as slow, because unless you plan to cover the entrance to the tunnel and circle back in a separate tunnel every time, you have to light it to prevent monsters greeting you when you return. In a 1x1, torches are optional because no monster will spawn in those conditions, so you can get much closer to the optimal mining rate.
My personal favorite mining strat is just going down to y -56, mining a 1 x 2 stripmine for about 100 or so blocks, placing tnt blocks every 2 blocks going backward. Once all thats done light it off and enjoy.
Not sure if you noticed this, but I have when I was using X-Ray for ore research I noticed Diamonds are diagonally adjacent to other ores, either vertically or horizontally, or sometimes parallel with other ores. My mining technique involves this homebrew research and I've noticed you can find usually hidden diamonds just by mining a couple staircases off of redstone ore or gold ore down at peak Diamond Y-Level. And if you want ores in general, "staircasing" off of ores can help you find more veins. I'm not too sure if this is per-chunk or in general, and it would be neat if this is a secret ore generation cheat code for getting lots of ore or diamonds just based on running into non-diamond ore when strip mining! :o
Set a timer for an hour. Explore caves at diamond level. Pause timer whenever not in diamond level cave, as we are only measuring the efficiency of actively hunting for diamonds
This was absolutely awesome. Absolutely awesome. I’m not an incredible math person and only took the first Calculus in college but you made it very accessible. Nice work👌
Half the community: Figuring out the most mathematically efficient way to find diamonds. Other half: Quarrying from surface to bedrock because zen. I'm the other half. But I appreciate how efficient mining can be if you're willing to put in the effort.
To be fair, this experiment was trending towards efficiency in getting diamonds _faster,_ but the 1x2 branch method is most efficient when it comes to wxtracting the highest possible concentration of diamonds from a given space, regardless of time, while the 1x1 straight tunnels leave a lot of "waste" diamonds behind.
@@ruanholtzhausen4000 I do want to do a cave mining video, but I haven’t been able to figure out how to test it well. With a lot of comments lately, I might also have to do a follow up to this one and I’ll make sure to compare these methods!
now i understand why my teatcher said that engineers over enginner everything, 2x1 with 3 block spacing where always my go to choice, and i never needed to explain why to my friends, analysing something rather than go test by test, data by data sometimes is the best But great video man
Yes! This is because it's not a normal distribution. It's hyperbolic. You cannot realistically get zero seconds per diamond. Also, if you found zero diamonds, you'd have infinite (undefined really) seconds per diamond.
Your limits at 6:50 should be lim(x->inf, (4*x+1)/x). As you have it now it just goes to infinity. You should show that the average blocks exposed per blocks mined tends to ignore the extra 1
I have an objection to your conclusion about which layer is best to mine on. The layers on which you're mining your tunnel expose twice as many blocks as the layers which are your floor and ceiling (and you're actually checking 3 times as many since diamonds in mined and exposed blocks both count). So the floor and ceiling layers of tunnels are weighted too strongly on your spreadsheet, because you're in practice only checking 1/3 as many blocks on those layers. This is made more complex by poking holes to either side as you tunnel, since the ratio of blocks exposed+mined is not so neat any longer, so you'd have to redo the math yourself for the right ratios there (though I doubt the difference is significant.) Based on your data, this makes level -56 by far the best to mine on across all samples for a 1x1 tunnel. For 2x1 the difference between -56 and -57 is essentially non-existent (-56 is worse than -57 by 0.05%). Here are my results for 1x1 (numbers are in the order 160K; 490K; 810K; 1M): -55 -- 419.3; 1302.3; 2119.7; 2594 -- short by 147 blocks -56 -- 437.3; 1334.7; 2157; 2653.3 -- BEST (0 blocks short) -57 -- 417; 1276; 2094.3; 2607.7 -- short by 187.3 blocks -58 -- 418.7; 1227.3; 2100.3; 2613.7 -- short by 222.3 blocks (notably, the 490K sample is somewhat of an outlier, and so it's possible that -58 is on average better than -57!) Here are my results for 2x1: -55 -- 657.3; 2001.7; 3287.3; 4049.3 -- short by 401.7 blocks of sum of best; short by 382.3 blocks of overall best -56 -- 682; 2099.3; 3406.7; 4185.3 -- short by 24 blocks of sum of best; short by 4.7 blocks of overall best -57 -- 683; 2085; 3401.7; 4208.3 -- BEST (short by 19.3 blocks of sum of best) -58 -- 670; 2007.3; 3362.3; 4180.3 -- short by 177.3 blocks of sum of best; short by 158 blocks of overall best
I appreciate the effort and detail that went into this video but I think you might be missing something. It is true that it's a lot quicker to just *find a diamond* when 1x1 strip mining, but what I think you're not taking into account is how long it takes to crawl back to the start considering 1. how slow crawling is and 2. how much more distance you cover compared to branch mining. I think that should be taken into account and I think it would immediately even things up when comparing, well, all methods (1x1, 1x1 branch, 2x1, 2x1 branch). Even ignoring 1x1 vs 2x1, just by using branches rather than pure strip mining, you're drastically cutting down your crawl/walk-back time per block exposed.
I was initially also thinking of this since headhitter sprint jumping is one of the fastest ways to move, and crawling is one of the slowest. However, it could be argued that you should just mine a few blocks over and then start a new tunnel back towards where you started instead of coming back the same way you went in (now you are mining the whole time).
@@johnenright9859 That's one way to do it, but then you end up with way more tunnels covering way more area, which is fine if that works for your setup but I prefer space efficiency personally. using branches is certainly the best as far as space efficiency goes.
I honestly didn’t consider backtracking cause I always just mine a tunnel back, so I’d recommend that for 1x1s. In terms of branch mining, I’d say that’s less area and time efficient than just 2x1 tunnels, so if you’re wanting to backtrack, then 2x1s would be the way to go
water bucket makes it pretty fast in the crawl tunnels, as well as the fact that on long mining sessions it may be faster to go to the surface or nether and travel back that way (especially faster with elytra)
I think this might bring in too many variables that assume the player makes perfect plays- if we're going to bicker about where you end up at the end of the mine and travel methods through tunnels, we may as well start weighting and accounting for other complicated player actions or movements that aren't based just on movement, but interface interaction and your game state ahead of the mining session (efficiently dumping your resources, storage access, biomes that might hinder lighting or breathing, transportation of materials to the mining site such as food or tools, etc.)... seems like this experiment gets us some instructive numbers regardless of these other optimizations or mining style
My only question is, considering you'll be returning periodically to the surface from your mining journey (to empty your inventory, craft new tools, etc), how long can you mine a 1-block shaft for before the time it takes to go back to nullify the gains in relation to other methods?
That’s a good question, but I don’t think it’d be a problem. First, I’d only recommend the 1x1 for diamonds. In this case, the only resources I use/need are picks, maybe torches, food, and diamonds. I have plenty of slots to spread diamonds and food in. For the other two, if you need to replenish those, I’d recommend mining tunnels out and back in so that way when you return to your main starter tunnel, you can easily pick up/drop off resources without losing any mining time. Then any travel time to the surface is negated as all methods will be the same.
Since there are 4 layers with decent distribution, I would probably do a staggered offset 1x1 grid pattern (3 over, 1 up/down), as that would seem to be the most efficient way to expose most of the blocks in those 4 layers (since diamond veins not exposed to air almost never spawn as a single ore, you don't need to see every block, just most of them)
One issue I can see with this is that when you find a diamond block on a given layer, say the ceiling of your 2x1 tunnel, then you will mine the exposed block but also any connected diamond blocks . In other words, if you mine the diamond block above you, you usually mine the other diamond blocks in that vein. So the total per layer isn't quite right in application. There's also the average vein dimensions. If your cross tunnels are too close, then even if you are exposing the ideal number of blocks, a previously discovered vein is very likely to bleed into the previous/next tunnel, whereas adding 1-2 more blocks of spacing would have all but guaranteed you found that vein while mining (potentially) half the blocks.
More than the video itself, which actually go into some interesting statistical analysis, the comment section brings in very insightful details about travel time and other pertinent variables. For Minecraft. I just want to say that you guys would make for good professionals in many research fields, if you're not in them already.
Awesome video, cool to see such thorough measured testing to provide concrete results, my only legitimate counter to 1x1 tunnel 1 gap vs 2x1 tunnel 1 gap (2x1 being my preferred method) is return time. I don't mind setting up a trap door, but it's not uncommon for me to dig a tunnel for 30 min to an hour and return to my entrance to get back to the surface rather it's ladders or a soul sand elevator. So return time then, would be how long does it take to return to the start of your tunnel, with a 2x1 I can just sprint and spam jump so even after mining in a straight line for an hour I'm back in a minute, with a 1x1 this return time would be much longer. so if you're really grinding out tunnel mining sessions, you might dig in a line for 30 minutes, drop off the valuables and then start a new tunnel with a 1 gap. So my logic is a 2x1 tunnel with 1 gap is still more efficient over your total world life span as your not losing time on the return trip. That being said you could just 1x1 tunnel, and than 1x1 tunnel your way back with the 1 gap. I think I'm just missing something here as the data does clearly show a 1x1 is best at diamonds per time invested. If you have a clarification to my counter argument or if this is a legitimate argument that was not tested, either way I would love to know the answer.
Great point here, and if return time mattered I'd say a 1x1 wouldn't be very efficient. In the testing, I didn't consider return time at all because I assumed the player would be mining a return tunnel back. This is how I mine, but I probably should have mentioned it in the video lol
6:15 I don't think you should count it like that. the block you mine is a block you exposed previously, so it should count as an exposed block, and so should every block mined.
@@pancito3108 A few people have brought this up, so I posted a pinned comment explaining how I determined ‘exposed’ blocks. Basically I don’t count the blocks we mine as exposed, but I’d say if you want to it’s a totally valid way to do it! The trends of the results remain the same, so either way works
Cave mining was actually nerfed pretty significantly in Caves and cliffs update. When the first snapshots came out, they realized that there were way too many diamonds being exposed, so they implemented the code to delete most of the exposed diamonds in caves. Which means that number of exposed stone in caves doesn't correspond to diamonds anymore.
@seanweb9051 Someone needs to do a test of minig and cave diving to know for sure, but it would definitely need to be an extensive test since new caves have so much variety in size and depth.
hey, i just noticed something right away that i was watching which maybe was overlooked. A simple sum of 3 or 4 layers might not properly capture the potential diamonds exposed for a 1x1 or 2x1 tunnel. The reason being, the middle layers of the tunnel will actually be 3 wide of exposed blocks, while the top and bottom layers will only be 1 wide of exposed diamond. To visualize what im trying to describe, think of the cross section of the tunnel as a "plus shape". So for the potential exposed diamonds you should actually use a formula that is something like bottom layer + (3 * middle layer) + top layer for a 1 wide tunnel, and a formula that is something like bottom layer + (3 * middle layer 1) + (3 * middle layer 2) + top layer for a 2 tall tunnel. This has some implications, because the middle layer spawning rates are favored over the top and bottom layers at a 3:1 ratio (1:3:3:1) for 2 tall tunnels, and a 3:2 ratio (1:3:1) for 1 tall tunnels. This trend is also less true as the tunnel gets wider. For a 2 wide 1 tall tunnel, there is a 1:1 (2:4:2) ratio, and for a 2x2 tunnel there is a 2:1 (2:4:4:2) ratio. My guess is that this wouldn't impact rates a whole ton anyway, as long as you're somewhere around y level -56 or -57, but if you're mining somewhere around y level -60 or y level -53 i imagine using a simple sum could begin to throw off numbers.
Yes. This is the second biggest error of the video. Didn't really affect the results much due to the relatively even distribution of diamonds at that level.
Wow! The best and most comprehensive video on this subject ive seen, by far! Im in bedrock so cannot use the 1x1 methods. But ive always mined on the layer or 2 above bedrock using 2x1 branch. Will certainly try Y57 2x1 tunnelling now.
@@killamackay Thank you! Based on the comments, there’s a lot more I could do and test so I want to do a follow up. I’ll make sure to include results for bedrock too!
Mining a 1x1 tunnel does expose five blocks per block mined. It's just that one of those blocks gets mined after you expose it. You get the diamonds whether or not it's in your mining path, so I see no reason not to count it.
@@DanielLCarrier yeah I think it’s fair to count it as 5 instead of 4. I’ve got a pinned comment explaining why I didn’t count the mined block if you want to check that out
Hi, hello I'm the gal who is interested in the standard deviation; I just wanted at add a little bit more analysis, because I believe that your conclusion is wrong due to your low sample size, and the difference in gap size is insignificant. For my analysis i compared 0, 2, 3-block gaps with the 1-block gap and used a confidence requirement of 95% to determine weather your results were accurate or not. I chose the null hypothesis H0: x_1 = x_{0,2,3}, and alternate hypothesis Ha: x_1 ~= x_{0,2,3}. I did three tests, comparing 1-Block gap, 2-Block gap, and 3-Block gap. The averages and standard deviation were used from 21:28. Since n = 10, the degrees of freedom = 10 - 1 = 9. Calculating the T test values yields critical values for a one tail test yields +-2.26 (either plus or minus, depending on the test). The two sample test scores were as follows: T_0 = (71.767 - 88.179 ) / sqrt(17.443^2/10 +43.700^2/10) = -1.1 T_2 = (71.767 - 95.158) / sqrt(17.443^2/10 + 48.227^2/10) = -1.44 T_3 = (71.767 - 108.76) / sqrt(17.443^2/10 + 75.516^2/10) = -1.509 since the values fall between +-2.26,6 we must fail to reject the null hypothesis, thus these methods are not measurably better or worse than each other
@@greatdanchik Thanks! Yeah it was kinda a lot and i definitely won’t be doing any more manual mining tests lol but I will be doing a follow up (code based) to hopefully get more and better data
@@grey521.mp4 I appreciate it! I did make a mistake on the limits though lol, I forgot to put the equations over x. Otherwise they’ll balloon to infinity
Magnificent analysis! Now do it again with water and lava diving (diamonds generate more frequently when exposed on all side, and that includes water and lava) >:D That and running around caves with or without a mob switch EDIT: You just mentioned that you wanted to test that, woops
i really loved this video, hope you will reach more people with you accurancy. And i would love to see a tunnel mining vs cave mining video, I've been thinking about this for a while now and I can't find an answer. Keep going!
@@danielepaschetta4742 thanks so much! I’ve been thinking about how to test cave mining for a while now, but can’t think of a good way to do it. And I really don’t want to actually cave mine in-game for a hundred hours 😂 I’ve got a couple ideas that I need to look into more, so hopefully I can make a video on it
@@TraitorousHomeworlder based on all the comments recently, I’ll definitely have to do a follow up lol I’ve never heard of moss mining, but I’ll make sure to check it out!
Fascinating video ! I too always thought that the 2×1 tunnel with branches was more efficient, I know I won't be doing that anymore. I wonder how feasible it would be to run that whole process programmatically in order to get the numbers for an even higher sample size, as well as test with efficiency 5 pickaxes.
@@witherflower42 yeah, I was a bit sad it wasn’t very efficient. I’ve been thinking about how to do this recently, especially with cave mining. I don’t know how Minecraft generates the diamonds so I’m not sure I could replicate it in python. I haven’t looked into mods yet or even making mods, so that could be a solution. If I do an updated video, I’ll be sure to test out efficiencies!
@@SeaJayPlays From my experience with datapack I'd say this would be a problem that could be fixed with those. In them you could iterate through your wool blocks and check each adjacent block whether it's a diamond ore
Also another point -- one of the important things you didn't really consider is space efficiency of each of these methods and how long it takes to reach untouched rock. If you can get super fast diamonds while you are mining then that's great, but if you're crawling through an infinitely long tunnel to get there then that's no real use. I'll probably start using 1x1 mining for this video, but I might try a layered version, and have major routes that I normally would go down every time as 2x1 as that's faster (without elytra/dolphin tunnels/etc).
Nice work! I know I'm late to the party but I'd like to ask about your choice of metric regarding blocks exposed per block mined vs blocks exposed per second. Considering that to go from one to the other you only have to multiply or divide by the average blocks mined per second, I think that blocks exposed per block mined is a better metric, since the other choice introduces a lot of subjectiveness/uncertainty. For example: * What is the time spent moving your cursor between mining blocks? * What gear do you have? Iron pickaxe? Diamond/netherite? Enchanted? * Do you have a haste 2 beacon or not? * How consistent you are between measurements All of these factors affect the metric, and seeing as your final standard deviations were quite large I'd wager that these uncertainties can't really be ignored, but I may be wrong Edit: Nevermind about the standard deviations, I missed that they were between layers and not measurement runs. But I think the point still stands regarding blocks exposed per block mined being the more objective metric
@@tomwaligorski9882 Thanks for this comment, you bring up a good point! I’d say I chose the diamonds/sec metric because it does try to account for ‘real world’ game play. I think some of the biggest differences between the mining techniques come from how much time is spent using them. For example, I wanted to include time spent placing torches (you technically don’t need to place torches with the 1x1 so that could increase its speed), and time spent turning for the branch mining methods. You’re right that it introduces a lot of inconsistency on my part in measuring that, and after getting feedback on this video, I’m convinced I need to do a follow up video/better version that measures things like that more consistently. I might have to get some community help in collecting time data for various tooling/effects lol
i'd love to see a test for other ways of getting diamonds. maybe use some code to ingest one thousand hours of someone caving and one thousand hours of someone end busting, and also i wanna see some calculation for the 3 most common designs of tunnel bores, with 1 module, 4 chunks, 8 chunks, and 12 chunks, and also checking how many times you should fire the tunnel bore before checking for diamonds. personally i went with the newest 2 block high tunnel bore, just high enough to not see bedrock, 8 chunks long, and 2 fires per scout. i've had some tests to see what's faster, caving in an ancient city with a warden switch, or using the tunnel bore, and got to the conclusion that if you include time between each ancient city, its roughly the same amount of diamonds per hour, but my data is only roughly 5 hours on each method, and isn't really scientific in nature.
Would love to see these stats with other popular mining patterns. Windmill especially, which attempts to optimize walking time, compared to straight branches, which typically involves some degree of backtracking. (I guess you could zig-zag a branch mine, but windmill also aims to optimize the walk-back to start by mining in all 4 directions, and creating a diagonal path back from any of the corners.)
This notation was really confusing to me at first, because I kept expecting tunnel height to be mentioned. Some of my mines have 1x2x1 and others are 1x3x1, because having more headroom lets you run faster. But it also means mining an extra block, so I usually skip it and make the tunnel only 2 blocks tall. So the 1x1 versus 2x1 confused me, because my tunnels never end up being double-wide. I can see how it would help, but 2x2x1 and 1x2x2 tunnels are only used to connect my mines to each other.
Btw, you should mine with a beacon, use 1x1 tunnels, BUT mine in a "horse circle", meaning you make a ⬛⬛⬜⬛ ⬜⬛⬛⬛ ⬛⬛⬛⬜ ⬛⬜⬛⬛ pattern. This makes your tunnels get a wider area. And you mine faster with the beacon, AND you get everything in those 4.5 layers mined. The only thing more efficient is a tunnel bore that is in a world without lava.
@@DevidCipher I do want to test how status effects affect the results. But I’d say that pattern probably isn’t the most efficient. You can expose all the blocks in a volume by leaving two blocks in between tunnels on the horizontals, then shifting that pattern for the vertical layers
I feel like since villager trades, diamonds are far less needed, instead i would really like a in depth netherite video, since there is no other way to obtain it rather than mining. Still a great video btw good job.
6:16 The blocks in fornt also count as exposed blocks because you dont know them. Also when you counted the blue glass and the blue wool and divided them, you didnt count the wool blocks them selves as exposed blocks. So the exposed blocks/ blocks mined ratio for 1x1 tunels is 5:1. The 2x2 ratio is 8:2, same as 4:1. I would appreciate you answering to confirm what i am saying. I also want to thank you for doing this types of videos that are actually interesting.
@@lucasbarroso2997 Hey, this is a good question, and I can definitely understand your logic. My answer would be that it depends on how you want to define exposed blocks. In my opinion, what matters is the blocks revealed as a result of your mining (so blocks that aren’t required to be mined). I’m gonna have to mine my tunnel no matter what, so to me that doesn’t count. Maybe a better way of phrasing it would be the blocks exposed after each block mined. Luckily, either way, the trends of the results stay the same!
Yes the results are the same. Now I can understand better the way you thought about the exposed blocks. But I keep thinking Its 1 more exposed than you say No matter what your way is valid in this context because as you said the results are still the same, not proporcionally but in the same order. And also, keep the amazing content.
@SeaJayPlays Hi. I like to find someone as nerdy as me when it comes to games. I see your point. By "exposed" you mean revealed for free, without mining time used. That's fair. When it comes to the "exposed/s" metric, it's quite misleading thought because people might think it means the (exposed + mined)/s metric, which it is not.
Blocks exposed per second was what u wanted to hear thank god. I found people stupid who crawl very slow just to get a better ratio on blocks exposed:mined blocks. Love u bro.
I think the #1 thing videos like this overlook is taking advantage of world gen behavior - we also need to look at each method when doing specific things with knowledge about world gen. Ex: as previously mentioned being between chunks, or the method of mining a chunk until you find the vein (and for this one, the optimal amount you should search before giving up & moving to the next), or how each method stacks up when leaving your current chunk as soon as you find a vein there ect... I think it makes a big impact overall
You absolutely should not being mining a chunk until it's drained, that'd be crazy slow. And mining on the borders of chunks is slim to none difference. The average is effectively the same, but there's a higher potential that 2 or even 4 diamonds veins spawn against eachother
@@kirasmith1147 This is a good point that I want to test further in a follow up video. I just haven’t been able to find any answer as to how diamonds are generated (specifically) so it’ll take some work to get that data
@@Nate-bd8fg You could use a method where you mine on the edge of one chunk, and then mine into the adjacent chunk five blocks in. You would then skip the rest of that chunk if you find a big vein of diamonds. I'd consider this too mind-intensive, keeping track of chunk numbers and whatnot. And if you're going to use mods to help with that stuff, might as well just use X-ray. Or creative mode.
Hey everyone, thanks for all the comments. There's been a couple recurring comments so I'd like to go over a clarification and correction here.
The first is a clarification on what I meant by "exposed" blocks. For me, there are two types of blocks which I'll call costly and free. Costly blocks are blocks that I have to spend time/effort on. So if I'm calculating exposed blocks for various methods, I want to know how many free blocks each method gives me. Basically, how many bonus blocks do I get to see that I didn't have to spend time/effort touching. I explicitly excluded blocks that cost me, or blocks that I have to mine anyway. This results in the 4:1 ratio for the 1x1 and 3:1 ratio for the 2x1 tunnels I got in the video. A few people have commented that the 1x1 is actually a 5:1 ratio because you are exposing the block that you will mine next in the tunnel. They've explained their view well and I totally agree that the 5:1 ratio is valid based on what they count as 'exposed'. I just should have done a better job at explaining what exactly I was calculating so I appreciate people pointing this out. The overall trend of the results are the same, so feel free to use whichever makes more sense for you.
The second is a correction on the limits shown in the video. When talking about how many exposed blocks there are per block mined, I took the limits of (4x+1) and (3x+1), which are only the "exposed blocks' part and by themselves go to infinity. I forgot to put them over the 'per block mined' part, x, which brings them back down to 4 and 3. So the correct limits are lim((4x+1) / x) and lim((3x+1) / x). Thanks for catching that!
Somebody probably pointed that out already but did you weight in exposing double on the layers you're actually mining? (When you mine a 2x1 corridor you exposed two wall blocks and 1 ceiling or 1 floor.
what you intended to calculate was a derivative, not a limit.
What about a 1x1 method where you mine your first row at Y-57 then with one block gap between the exposed blocks / glass but at Y-56
This would result in a method with a spacing between the 1 and 2 block gaps which might be more efficient
@@jackweslycamacho8982 That is the result, yes, but the derivative doesn't actually explain why it matters, while proposing it as a limit actually translates to the more you mine the closer you will get to this ratio, while it is identical, it's not pedagogically equivalent.
The seconds/block exposed metric doesn't actually matter, unless you go for the torches in which case the 1x1 tunnel is superior because you don't need torches because nothing can spawn. But most players just use allbright or gamma=100 and wouldn't bother with the mobs, and considering you can constantly mine regardless of what method you are using(with the branching ones there is a slight inefficiency if you are not paying full attention and skipping the blocks with perfect tick skipping, essentially breaking every second block on your way looking out of the branch, to ensure you aren't missing any breaking ticks by turning, but most players have this slight inefficiency even without branches)
Mfs get a PhD in statistics and CompSci and immediately use it for Minecraft. This community is so good
😂 you might offend some statistics and compsci majors, but thanks
Gotta keep the brain sharp for whenever a job interview happens (it never does)
If you can’t use your degree for Minecraft was it even worth it?
At least he is doing something useful with it.
Well, this level of statistics can be learned even in medical school 👀
6:34 I believe the limit is meant to be (4x+1)/x
The current formula just balloons to infinity.
@@TheKastellan You’re right! Another comment caught this too and I realized I didn’t put the full formula in. Thanks for correcting it
@@SeaJayPlaysamazing video man
Should be (5x+1)/x. Blocks don't retroactively become unexposed when you mine through them.
@@DanielLCarrier if you are making a tunnel, you are going to mine that fifth block even if it wasn't ore, so it would be considered blocks mined instead of blocks exposed.
@@Big_boie It's both.
Imagine instead of using any of these efficient mining techniques, you just mine out a solid cube. Using your method, the limit is 6x^(2/3)/x, which approaches zero. And yet the diamonds you find goes with the volume you mine, not the surface area.
Something that might be important is that in a 1x1 tunnel, you'll be moving more slowly. This matters for the return trip, after you're done mining, and it would matter earlier if you could mine blocks faster than you could move to reach new ones
That’s a good point, but for me I just mind a tunnel back lol so if I mine for 20 minutes I’ll go 10 out and 10 back. Otherwise, yeah I’d agree it would waste a lot of time
From my experience on bedrock, it actually causes you to go faster and thus boosting efficiency
@@SeaJayPlays holy shit why have i never considered mining backwards, that's genius
@@artemisSystembruh 😂
using swift sneak you can crouch while in crawling mode which will boost your player movement speed even though you are not visually crouching
Hi! Nice video, but in the end you are overlooking an important thing: You said that the standard deviations don't matter, but actually, they are extremely important in the interpretation of your results in this case. They strongly indicate that your results are not significant because the variation (your error) is too high. The testing area of 100x100 is just too small to account for the huge deviations between locations, particularly because diamonds are so rare. 75 seconds per diamond per layer is way within 1 standard deviation of all your results for the 1x1 tunnels, actually within half a standard deviation. Assuming normal distributed diamond amounts per layer, for example, this means that the hypothesis that the average time to mine for a diamond for all 1x1 tunnel techniques is 75 seconds cannot be rejected with more than 38% confidence ( half a standard deviation) for all the techniques, or in other words based on these results its more likely that all techniques yield the same result than not.
If anything, than this says that the techniques don't matter unless you go on a very large scale, and luck plays a much more important role.
But as others have pointed out, we could try to extend the efficiency calculation by adding in the time to go back to base, which means that more dense techniques or ones where you can walk or run back more quickly will have an advantage, and try to write some scripts to automate the exposed diamond calculations for larger areas to get more accurate results. Thanks for providing great ground work here!
Very good point
Very good point
I totally agree, seeing the SD equal or close to the mean was quite perplexing.
But I can totally see why he kept the samples relatively small xd
i think the running back to base thing is also an interesting dilemma. With a 1x1 tall mining, it would take forever to run all the way back to base, so i turn around when my inventory is halfway full. that way the effective time to get back to base is 0s, since half of my mine is away, and half is towards. Also late game when I have shulker boxes, i find it easier to just stache everything in the boxes, and mine for well over 1000 blocks, then to get back i just mine straight up and fly home with an elytra (or /kill if I have keepinventory on). flying home with an elytra will be way faster if im going super far away, but slower if I'm mining less, because it would take a while to resurface. I think a more in depth video on this topic would be really interesting.
Great point and I definitely agree. There's just no way I'm doing any more manual testing for this lol But I am planning on doing a follow up code-based video so we can get a much larger sample size for more accurate data. I also want to test a lot more methods (maybe even cave mining but that might deserve it's own video). Lastly, I didn't include return time for any of the methods tested as I assumed the player would be mining a tunnel back. So if the 1x1 does end up better, I'd always recommend snaking instead of backtracking.
In the newer versions of Minecraft, strip mining techniques are completely irrelevant. Deepslate is far too slow to mine, even with a haste beacon and Eff 5 pick. The fastest way to get diamonds is to drink a night vision potion and fly around a cave, or swim around a submerged cave. You'll have way more fun, find other cool stuff along the way, and get a crap load of diamonds.
Yeah I agree. I mentioned this near the end, but I chose to test out strip mining cause that's how I like to mine for diamonds lol. It's a lot more chill for me that way. But I have been looking at ways to test cave mining, so I might do a video about that in the future. Let me know if you've got any ideas for that!
that has the downside of needing to deal with mobs, strip mining you can pretty much just zone out/autopilot
and if your world is pre-1.17 swamp clay diamond finding is always the best, if a bit cheesy
@@ThylineTheGay I agree, strip mining is just a good way to chill and play. I hadn’t even considered people playing on pre 1.18 versions, good point.
But I think it’d still be an interesting challenge to find a way to accurately test cave mining. Implementing mob interactions would be difficult unless I just manually cave mined for hours and hours but that doesn’t seem like a sustainable method lol
caving is arguably the best compromise between speed and approachability, but probably not the fastest, especially not if you allow tnt duping.
I feel I am usually quite unlucky caving. Someone should probably do a "seconds per diamond" test run for that... also its a matter of how far you are in the game, quite early you might not have access to enough gear and potions to do that comfortably.
17:38 When you said "I appreciate it I know its been a pretty long one." I looked at the time and realised the length of the video but was so enthralled (because maths and minecraft) that I said out loud "I appreciate you too" and my wife gave me a funny look.
Thanks for the video, all the work and effort you've put in ( and you made your excel sheets nicely formatted too 👍 )
Dang, thanks for the kind comment
I feel like it’s a normal video length. Does he usually make shorter videos?
The way I saved this thinking it was about real life mining - something I’m very interested in - started watching, realised it was about Minecraft, and then continued to watch because it also collides with my interest in Minecraft
My pro tip is X-ray
@@linus0075 😂
Baritone
X-ray=?
Just play creative
@@vegtt8445if it doesn’t end up killing you then yeah
One thing I was confused by until I went back to double check is the terminology "no spacing", "1 block spacing", "2 block spacing", etc. I usually do tunnels with 3 blocks in between them which seems the best but showed terrible stats, but then the "no spacing" idea made no sense at all, since no spacing would just be mining the whole layer. Your spacing though is referring to blocks between exposed blocks, so "1 block spacing" is the method I usually use with 3 blocks between tunnels and it is the best. While this might make sense with the way you were showing the spacing with the glass and the wool, I think referring to it by how many blocks are between each tunnel is more clear since that is what people will be counting when mining.
Holy shit I didn't even notice this ☠️ wouldn't have without this comment
I have to say. Very well made and explained video. I'm in the middle of developing my Masters Thesis and while watching this I realized that here you've essentially done a thesis on diamond mining. You're solving the problem of people not knowing how to best find diamonds, you present your research questions, you explain your research and the metrics you used to prove said research in your experiments, and you answer the mentioned questions throughout the video, before finally, showing your results and conclusions. Its funny to think about how anyone working on a thesis should unironically watch this video as reference. Amazing job.
@@MartinDestro I appreciate the compliment! There’s definitely some errors in the video but I’m glad to see the structure of it came across well
If you're developing your masters, I challenge you to find the critical error in the calculations being done on screen at 19:15. Hint: Think about what the standard deviations mean and why them being so high implies something is wrong. My other comments reveal the answer, so don't spoil yourself!
standard deviation is still interesting, because it's how consistently you're finding diamonds, if you have a method with a huge sd that would mean you find a lot of diamonds at once, but sparsely, so you could find no diamond vein for a while and then multiple back to back, while if you had no deviation you could basically put a timer and you'll be sure you will get diamonds every time.
I'd be interested in how other methods would compare, like hilbert curve and such. also I still like 2 high farming because it's easier to get around, I think I might switch to having like 9 rows of 1 high, and a row of 2 high. though an advantage of 1 high is that you don't need torches to avoid monsters.
If I have the motivation I might make a script that calculates the actual efficiency for a given method.
That script would be really nice!
Also sd matters A LOT. The 1 block gap method had the lowest sd, which means we can trust the number from that one the most. The rest were really inconsistent, which means with a small data set like this, the numbers probably aren't accurate. 2 block gap could be better than 1-block gap, but we will never know because it has such a large standard deviation and a small data set.
@@simonwillover4175 yeah, I'm currently looking and the generation, according to a modder there's :
- 7 veins per chunk with a size of 1-5 diamonds.
- 1 vein per chunk (1/9 chance) with a size of 1-23 diamonds.
- 4 veins per chunk with a size of 1-10 diamonds.
and with 1.20.2 added:
- 2 veins per chunk with a size of 1-10 diamonds.
The first and last vein type has a 50% chance to discard if exposed to air, the second vein type has a 70% chance to do so, and the third vein type has a 100% chance to do so.
which means there's probably a more efficient conditional way to farm (using a chunk's expected yield, if you got more diamonds than average you probably should skip the chunk.)
@@satibel Yeah I’ve been trying to think of ways to program this for more data, mainly to test cave mining easily, so let me know how it goes! I definitely want to get those standard deviations down in a future video
@@SeaJayPlays my idea is to pre gen like 10k chunks for each biome then convert the world with amulet and write a custom script that iterates my mining functions.
Desert should yield slightly more diamonds because of fossils.
The problem with 1x1 block mining is that your limit isn't your pickaxe, it's how fast you can crawl. This didn't come up in your testing because you were using an unenchanted diamond pickaxe, which doesn't mine faster than you crawl, but most of the time I imagine mining with Eff 5, which does. It also takes much longer to get back home once you're done, but that's a hard thing to include in testing
@@sorin_markov this is a good point and I’ll have to test out enchantments in the next one. As for the return time, I’d recommend always mining a tunnel back so you don’t have to spend time backtracking. I probably should have mentioned that lol
@@SeaJayPlays Mining back requires inventory space, pick durability, and torches. I usually stop mining and make a return trip because one of those has run out.
I wrote a datapack a while back to test this, basically just simulating different methods of mining physically as if I were actually mining myself. Glad to see our results matched up in the end, but what really would be more interesting now would be comparing it to caving, if only caves were more consistant
That’s really interesting, I’ve never looked into making a datapack. I’ve got some ideas for how to test cave mining so stay tuned
18:26 your expected time to mine and actual time to mine deviate so much becuase you forgot to take into account the 0.3s delay between mining 2 blocks. Your expected time to mine should be 50% higher. Amazing vid though!
@@eladnlg I’ll have to double check how I calculated it, but that sounds right. Thanks for pointing that out! If I do a follow up video I’ll make sure to correct that
I thought so too: five ticks delay after mining a block and an extra tick delay for focusing on a new block
They need an item that halves that delay time. Some mid-gane bonus before insta-mining is easy
Though you can circumvent that one tick focus delay by rebinding the mine block button from left mouse to a keyboard key. Useful for digging long straight tunnels and would further improve your efficiency.
One thing you didn't account for is the return to base time, when you mine one long tunnel you'll take longer to return, would be nice to see if it ends up changing the results
I didn’t account for return time cause when I mine diamonds I just mine a return tunnel back. So if I mine for 20 minutes I’ll go 10 out and 10 back. A few people have mentioned this now, so in the follow up I’ll make sure to cover it. Thanks!
Well... it isn't efficient to return to base 😂
Yesterday I used a simple TNT duping quarry for about an hour and a half and got 2 stacks of diamond ore, with fortune III it ended up giving me 426 diamonds. It also only took about 10 minutes to build the machine
I’ve never looked into dupers or quarry mining so I’ll have to check them out! Thanks for the suggestion
@@SeaJayPlays That would be an awesome video to watch. I've always known TNT mining is faster, but it'd be fun to see the data behind it
can u give us the name or the design of the machine you built? thanks@@asiriomi
@@ray-vz4jp it was the one designed by raysworks ua-cam.com/video/1Ql6ioAF9CM/v-deo.htmlsi=wqP6tqM-nlP4iW61
6:50 that's the derivative with respect to x (which is whats being looked for anyway) not the limit as x approaches infinity
@@tfk_001 I’ve got a pinned comment where I make the correction to what I meant to do, but the equations should have been over x corresponding to how we are calculating exposed blocks per block mined
@@SeaJayPlays oh yeah missed that thanms
Nice video. I loved to hear your process and the sample size was way larger than I expected.
One critique I'd offer.
I believe when 1x1 mining, if your pickaxe has high enough effeciency, your crawling movement speed starts to throttle how fast you can dig. . .
This is a small, but possibly meaningful difference between the 1x1 and 2x1 methods as a whole.
I would love to see if that upsets your scales if you are late game mining with an effeciency-5 pick.
@@benjaminsipe9217 Thank you!
This is a good point, and I think someone did mention it early on. I’d say that based on just how much better (efficiency-wise) the 1x1 is over the 2x1, my guess would be that the crawl speed wouldn’t be enough to change which is faster. But I would like to do an even more thorough testing between these two methods to really see. A lot of in-game factors could play a part like placing torches, placing trapdoors, actually mining the diamonds, running into lava, mobs, food, and I’m sure others.
I’m not sure how speed potions affect crawling, but that could be a solution (especially end-game) if it does end up making the 1x1 slower overall.
The Swift Sneak enchantment also increases your crawling speed, so this would be a more viable option to counter the slow crawling pace
@@colywolygaming4643there is also a glitch speedrunners use to sprint while crawling
Practical Tip: If you are worried about the time to travel back to your starting point after a 1x1 mining session, when you decide you are halfway done mining, make a u turn (including that 3 block width). You'll end where you started, while not having to waste hardly any time crawling through tunnels.
@@dtplayers I honestly thought this was common practice, but based on the comments, see that I should have brought it up. Thanks for the good explanation!
I can’t describe how happy I am to find a channel that takes an analytic approach to farm / mining. I’ve built and optimized some smaller farms which is super fun. It’s not a game unless I’m breaking out excel😅
So I'm guessing since Y-56 and Y-57 have the most diamonds, and 1x1 with 1 block gap is the most efficient mining technique, it could be good to stack your Y-56 tunnels on top of the 1 block gap in Y-57
Maybe by using this technique it would even be better to have a 3 block gap
@@NathanHaaren That’s a really interesting method that I’ve never seen before. I’ll definitely test it out in the follow up video!
I just read comments that video felt so short and then I realise that it is actually over 20 minutes!!! Wow, it was extreamly interesting, well organized and useful! Looking forward for more such type of content!
Good stuff! I had done some crappy tests back in the day and found out that branch mining with a spacing of 6 (or is it 4? 6 blocks in between holes) and digging 100 long tunnels of 2x1 yielded nice results.
I tested it by replacing common blocks with air to simulate digging and then manually counted the diamonds on the way.
@@bs_blackscout Yeah I had always used that method from Xisumavoid’s video, so I was surprised too. I really didn’t want the 1x1 to win lol
That’s pretty similar to what I did, and in the future I want to find a way to automate it to so I can get more data
@@SeaJayPlaysthat Amulet tool is written in python and it's open source... you can use their library to open a world and program in a mining method to see how it would fair over the long term without it actually taking a long term
I mine 2 tall tunnels because I want the deepslate for building. The diamonds, for me, are just a bonus that I mine around and leave in place unless I have a silk or fortune pickaxe with me. If I have a haste beacon I'm going to be mining the entire circle of affected area. This stuff is good to know though.
Worth the wait? YES! I *LOVE* these mathematical experiments as done with Minecraft. When I took my Computer Architecture course, I had a BLAST figuring out how to translate on-paper math into Minecraft. This is so much more fun. I'd love to see more content like this. New sub, man!
Tiny problem: if you mine in a 1×1 tunnel you still need to return the same way you mined, meaning you need to painfully stare at your screen while crawling back to your base
Nah, just only mine halfway(whatever that means to you, half full inventory, half your torches are gone, half your food, whatever) then turn around and go three blocks over and mine your next strip back the way you came.
@@dallencorry yeah I just mine a tunnel back. I don’t have the patience to crawl 500 blocks back lol
3:40 I think this isn't quite accurate. You should probably have weighted your sum based on the fact that you're exposing two faces each for layers ~ and ~-1, and only one face each for ~-2 and ~1
Edit: I forgot some crazy people mine one-high. So in that case it's 2 for ~ and 1 each for ~1 and ~-1. I don't know if that's what you did, but you didn't say that's what you were doing
@@noway2831 Yeah, it’s been so long I don’t remember how I did it. So I’ll have to go back and look at how I calculated the stuff to give a good reply. But thanks for brining this up cause I want to do a better/more thorough follow up based on all the comments so far
i've noticed you're saying ~1 for the block above your head and ~-2 for the block below your feet, but that doesn't work anymore because you're position is now calculated from where you're feet are, rather than from where your head is like it used to be.
i'm not sure why they changed that but ~-1 is now the block under your feet and ~2 is the block above your head
I think it is still accurate because this isn't how many exposed diamonds, but rather which layer has the most raw diamond output. So this specific calculation is essentially if you mined out the entire layer, you'd fully see the layer you mined, as well as the above and below layer. The exposed diamond test came later where he tested the gaps, but this one was just because the largest total amounts of diamonds are in those three layers combined, giving the best likelihood of running into diamonds there as opposed to other layers, not other mining styles.
@@thirdeditman9494Removing beta ladder climbing. that's the main reason.
Great video! Only thing I’d like to see added is a test on spacing the 1x1 with height variance. Like a wine-rack spacing on the 3 block spacing at levels -56 and -58 or something.
"Wine racking" should definitely be the word for this. The concept makes sense, and I always do it. I'm just not sure if it actually works out given the width of the best diamond density depths, and the size of veins. Would love to see it tested
@@theswankypotato This is a really interesting idea and I’d never heard of it or seen it done before. So I’ll make sure to test it out in the follow up video!
Hey, this is so underrated. A Video which would be loved by every single minecraft player but sadly only 22K Views. Great Work tho! Thanks for calculating everything, i think its very interesting to know the maths behind mining diamonds. One other point: I actually always mined a 2x1 tunnel with 1x1 branches going off of them. Surely a bit more complicated to calculate but i would be interested to know the results behind this method. Keep up the great work! Thanks from Germany!
@@etwasflauschiges Thank you! I’m planning on doing a more in depth follow up video in the future (code based this time) so hopefully I can test a lot more methods
Dang you're a legend! I don't know many people with this type of patients.
Thanks so much!
Maybe you will bring it up but at 17:14 I see the Error about the "hidden diamonds" that have been kept, as these type of diamonds require more effort, they need to be treated as a separate ground. We essentially have 3 groups of diamonds: visible (easy), hidden (more effort), invalid (0). Adjusting this could make a difference in the general methods depending how the hidden to visible diamonds ratio changes. The effort of the hidden blocks need to be statistically proven too, by how many blocks on average you would need to dig out to expose the hidden ones in the most efficient way, which may also mean that there is a method that would allow some hidden diamonds to be missed.
Anyway, we could then compare if the effort of searching for hidden diamonds is itself worth it, or if it is best to not go for them at all. Lots of calculations..
I absolutely love these kinds of scientific/statistical approaches to video games. Very fun!
One thing I'd suggest for your future vid: do an actual trial run of these tunnels in survival. I think there's a lot of small details that start to add up that may change the efficiency of each tunnel. One being the time to return that others have mentioned. Another being the collection of the diamonds themselves. Wouldn't you stand up every time you collect 2-tall diamonds, and then have to recrawl to get back in the tunnel?
I think it's nice to demonstrate how real mining times will always be slower than simply calculated time to mine. 2x1s should even be more than twice as slow, because unless you plan to cover the entrance to the tunnel and circle back in a separate tunnel every time, you have to light it to prevent monsters greeting you when you return. In a 1x1, torches are optional because no monster will spawn in those conditions, so you can get much closer to the optimal mining rate.
My personal favorite mining strat is just going down to y -56, mining a 1 x 2 stripmine for about 100 or so blocks, placing tnt blocks every 2 blocks going backward. Once all thats done light it off and enjoy.
I saw what you mean but the formula is (4x+1)/x, without the denominator the limit goes to infinity
@@michelleslay2701 Good catch! I added a pinned comment correcting this, now I just need to figure out how to correct it in the video lol
You deserve more views.
I really appreciate it!
Not sure if you noticed this, but I have when I was using X-Ray for ore research I noticed Diamonds are diagonally adjacent to other ores, either vertically or horizontally, or sometimes parallel with other ores. My mining technique involves this homebrew research and I've noticed you can find usually hidden diamonds just by mining a couple staircases off of redstone ore or gold ore down at peak Diamond Y-Level. And if you want ores in general, "staircasing" off of ores can help you find more veins. I'm not too sure if this is per-chunk or in general, and it would be neat if this is a secret ore generation cheat code for getting lots of ore or diamonds just based on running into non-diamond ore when strip mining! :o
this is wild, back when I used to play mc 10 years ago the advice was to just strip mine at y 13 lol.
THANK YOU SO MUCH, everyone says the same thing but its so nice to see the data done with real work
Set a timer for an hour. Explore caves at diamond level. Pause timer whenever not in diamond level cave, as we are only measuring the efficiency of actively hunting for diamonds
Why dont you do it?
This was absolutely awesome. Absolutely awesome. I’m not an incredible math person and only took the first Calculus in college but you made it very accessible. Nice work👌
Thanks!
And I’m not an incredible math person either lol I wrote the wrong limits 😂 (correction in a pinned comment if you’re interested)
Half the community: Figuring out the most mathematically efficient way to find diamonds.
Other half: Quarrying from surface to bedrock because zen.
I'm the other half. But I appreciate how efficient mining can be if you're willing to put in the effort.
That's a really great in depth video! Thank you for sharing and creating it, this answers one of the most common debates for strip miners lmao
To be fair, this experiment was trending towards efficiency in getting diamonds _faster,_ but the 1x2 branch method is most efficient when it comes to wxtracting the highest possible concentration of diamonds from a given space, regardless of time, while the 1x1 straight tunnels leave a lot of "waste" diamonds behind.
Finally some good sinentofic backed answer to the question we asked for years
Next, Branch mining vs caving vs underground lakes swimming mining.
@@ruanholtzhausen4000 I do want to do a cave mining video, but I haven’t been able to figure out how to test it well. With a lot of comments lately, I might also have to do a follow up to this one and I’ll make sure to compare these methods!
I've always used the 2x1 branch method because I learned to mine before crawling was a thing, definitely gonna start using the 1x1 method from now on.
now i understand why my teatcher said that engineers over enginner everything, 2x1 with 3 block spacing where always my go to choice, and i never needed to explain why to my friends, analysing something rather than go test by test, data by data sometimes is the best
But great video man
Those standard deviations are HUGE! Though, I greatly appreciate a well-made analysis
Yes! This is because it's not a normal distribution. It's hyperbolic. You cannot realistically get zero seconds per diamond. Also, if you found zero diamonds, you'd have infinite (undefined really) seconds per diamond.
As soon as you used calculus I INSTANTLY subscribed lol
Your limits at 6:50 should be lim(x->inf, (4*x+1)/x). As you have it now it just goes to infinity. You should show that the average blocks exposed per blocks mined tends to ignore the extra 1
I have an objection to your conclusion about which layer is best to mine on. The layers on which you're mining your tunnel expose twice as many blocks as the layers which are your floor and ceiling (and you're actually checking 3 times as many since diamonds in mined and exposed blocks both count). So the floor and ceiling layers of tunnels are weighted too strongly on your spreadsheet, because you're in practice only checking 1/3 as many blocks on those layers.
This is made more complex by poking holes to either side as you tunnel, since the ratio of blocks exposed+mined is not so neat any longer, so you'd have to redo the math yourself for the right ratios there (though I doubt the difference is significant.)
Based on your data, this makes level -56 by far the best to mine on across all samples for a 1x1 tunnel. For 2x1 the difference between -56 and -57 is essentially non-existent (-56 is worse than -57 by 0.05%).
Here are my results for 1x1 (numbers are in the order 160K; 490K; 810K; 1M):
-55 -- 419.3; 1302.3; 2119.7; 2594 -- short by 147 blocks
-56 -- 437.3; 1334.7; 2157; 2653.3 -- BEST (0 blocks short)
-57 -- 417; 1276; 2094.3; 2607.7 -- short by 187.3 blocks
-58 -- 418.7; 1227.3; 2100.3; 2613.7 -- short by 222.3 blocks (notably, the 490K sample is somewhat of an outlier, and so it's possible that -58 is on average better than -57!)
Here are my results for 2x1:
-55 -- 657.3; 2001.7; 3287.3; 4049.3 -- short by 401.7 blocks of sum of best; short by 382.3 blocks of overall best
-56 -- 682; 2099.3; 3406.7; 4185.3 -- short by 24 blocks of sum of best; short by 4.7 blocks of overall best
-57 -- 683; 2085; 3401.7; 4208.3 -- BEST (short by 19.3 blocks of sum of best)
-58 -- 670; 2007.3; 3362.3; 4180.3 -- short by 177.3 blocks of sum of best; short by 158 blocks of overall best
You’re right, thanks for the thorough comment! I’ll make sure to point this out in the next video
I'm amazed with how well thought out this is.
I appreciate the effort and detail that went into this video but I think you might be missing something. It is true that it's a lot quicker to just *find a diamond* when 1x1 strip mining, but what I think you're not taking into account is how long it takes to crawl back to the start considering 1. how slow crawling is and 2. how much more distance you cover compared to branch mining. I think that should be taken into account and I think it would immediately even things up when comparing, well, all methods (1x1, 1x1 branch, 2x1, 2x1 branch). Even ignoring 1x1 vs 2x1, just by using branches rather than pure strip mining, you're drastically cutting down your crawl/walk-back time per block exposed.
I was initially also thinking of this since headhitter sprint jumping is one of the fastest ways to move, and crawling is one of the slowest. However, it could be argued that you should just mine a few blocks over and then start a new tunnel back towards where you started instead of coming back the same way you went in (now you are mining the whole time).
@@johnenright9859 That's one way to do it, but then you end up with way more tunnels covering way more area, which is fine if that works for your setup but I prefer space efficiency personally. using branches is certainly the best as far as space efficiency goes.
I honestly didn’t consider backtracking cause I always just mine a tunnel back, so I’d recommend that for 1x1s. In terms of branch mining, I’d say that’s less area and time efficient than just 2x1 tunnels, so if you’re wanting to backtrack, then 2x1s would be the way to go
water bucket makes it pretty fast in the crawl tunnels, as well as the fact that on long mining sessions it may be faster to go to the surface or nether and travel back that way (especially faster with elytra)
I think this might bring in too many variables that assume the player makes perfect plays- if we're going to bicker about where you end up at the end of the mine and travel methods through tunnels, we may as well start weighting and accounting for other complicated player actions or movements that aren't based just on movement, but interface interaction and your game state ahead of the mining session (efficiently dumping your resources, storage access, biomes that might hinder lighting or breathing, transportation of materials to the mining site such as food or tools, etc.)... seems like this experiment gets us some instructive numbers regardless of these other optimizations or mining style
My only question is, considering you'll be returning periodically to the surface from your mining journey (to empty your inventory, craft new tools, etc), how long can you mine a 1-block shaft for before the time it takes to go back to nullify the gains in relation to other methods?
That’s a good question, but I don’t think it’d be a problem. First, I’d only recommend the 1x1 for diamonds. In this case, the only resources I use/need are picks, maybe torches, food, and diamonds. I have plenty of slots to spread diamonds and food in. For the other two, if you need to replenish those, I’d recommend mining tunnels out and back in so that way when you return to your main starter tunnel, you can easily pick up/drop off resources without losing any mining time. Then any travel time to the surface is negated as all methods will be the same.
extremely underrated and very useful, thanks :)
this video started with a powerpoint transition thats how you KNOW this is gonna be baller
Since there are 4 layers with decent distribution, I would probably do a staggered offset 1x1 grid pattern (3 over, 1 up/down), as that would seem to be the most efficient way to expose most of the blocks in those 4 layers (since diamond veins not exposed to air almost never spawn as a single ore, you don't need to see every block, just most of them)
One issue I can see with this is that when you find a diamond block on a given layer, say the ceiling of your 2x1 tunnel, then you will mine the exposed block but also any connected diamond blocks . In other words, if you mine the diamond block above you, you usually mine the other diamond blocks in that vein. So the total per layer isn't quite right in application.
There's also the average vein dimensions. If your cross tunnels are too close, then even if you are exposing the ideal number of blocks, a previously discovered vein is very likely to bleed into the previous/next tunnel, whereas adding 1-2 more blocks of spacing would have all but guaranteed you found that vein while mining (potentially) half the blocks.
Good points, I'll make sure to double check and include these in the follow up video, thanks!
There should be a fairly clear relationship between the average diameter of diamond ore veins and the ideal gap between mining blocks.
@@andriypredmyrskyy7791 Agreed. I plan on doing a follow up video with a lot more data points so we’ll see if we can get that formula
More than the video itself, which actually go into some interesting statistical analysis, the comment section brings in very insightful details about travel time and other pertinent variables. For Minecraft. I just want to say that you guys would make for good professionals in many research fields, if you're not in them already.
Awesome video, cool to see such thorough measured testing to provide concrete results, my only legitimate counter to 1x1 tunnel 1 gap vs 2x1 tunnel 1 gap (2x1 being my preferred method) is return time. I don't mind setting up a trap door, but it's not uncommon for me to dig a tunnel for 30 min to an hour and return to my entrance to get back to the surface rather it's ladders or a soul sand elevator. So return time then, would be how long does it take to return to the start of your tunnel, with a 2x1 I can just sprint and spam jump so even after mining in a straight line for an hour I'm back in a minute, with a 1x1 this return time would be much longer. so if you're really grinding out tunnel mining sessions, you might dig in a line for 30 minutes, drop off the valuables and then start a new tunnel with a 1 gap. So my logic is a 2x1 tunnel with 1 gap is still more efficient over your total world life span as your not losing time on the return trip. That being said you could just 1x1 tunnel, and than 1x1 tunnel your way back with the 1 gap. I think I'm just missing something here as the data does clearly show a 1x1 is best at diamonds per time invested. If you have a clarification to my counter argument or if this is a legitimate argument that was not tested, either way I would love to know the answer.
Great point here, and if return time mattered I'd say a 1x1 wouldn't be very efficient. In the testing, I didn't consider return time at all because I assumed the player would be mining a return tunnel back. This is how I mine, but I probably should have mentioned it in the video lol
6:15 I don't think you should count it like that. the block you mine is a block you exposed previously, so it should count as an exposed block, and so should every block mined.
i didn't even realize that myself, LOL; the limit formulas would become:
(4x+1)/x for 2-block tall and (5x+1)/2 for 1 block tall
@@pancito3108 A few people have brought this up, so I posted a pinned comment explaining how I determined ‘exposed’ blocks. Basically I don’t count the blocks we mine as exposed, but I’d say if you want to it’s a totally valid way to do it! The trends of the results remain the same, so either way works
Cave mining was actually nerfed pretty significantly in Caves and cliffs update. When the first snapshots came out, they realized that there were way too many diamonds being exposed, so they implemented the code to delete most of the exposed diamonds in caves. Which means that number of exposed stone in caves doesn't correspond to diamonds anymore.
I believe exposed stone still correlates to diamond quantity, just that it's 1/4 of what it used to be
@seanweb9051 Someone needs to do a test of minig and cave diving to know for sure, but it would definitely need to be an extensive test since new caves have so much variety in size and depth.
I applaud the scientific diligence that went into this video
hey, i just noticed something right away that i was watching which maybe was overlooked. A simple sum of 3 or 4 layers might not properly capture the potential diamonds exposed for a 1x1 or 2x1 tunnel. The reason being, the middle layers of the tunnel will actually be 3 wide of exposed blocks, while the top and bottom layers will only be 1 wide of exposed diamond. To visualize what im trying to describe, think of the cross section of the tunnel as a "plus shape". So for the potential exposed diamonds you should actually use a formula that is something like bottom layer + (3 * middle layer) + top layer for a 1 wide tunnel, and a formula that is something like bottom layer + (3 * middle layer 1) + (3 * middle layer 2) + top layer for a 2 tall tunnel.
This has some implications, because the middle layer spawning rates are favored over the top and bottom layers at a 3:1 ratio (1:3:3:1) for 2 tall tunnels, and a 3:2 ratio (1:3:1) for 1 tall tunnels. This trend is also less true as the tunnel gets wider. For a 2 wide 1 tall tunnel, there is a 1:1 (2:4:2) ratio, and for a 2x2 tunnel there is a 2:1 (2:4:4:2) ratio.
My guess is that this wouldn't impact rates a whole ton anyway, as long as you're somewhere around y level -56 or -57, but if you're mining somewhere around y level -60 or y level -53 i imagine using a simple sum could begin to throw off numbers.
Yes. This is the second biggest error of the video. Didn't really affect the results much due to the relatively even distribution of diamonds at that level.
This really deserves more attention, i would love for this to become a series like "debunking minecraft with MATH" or smth
Keep it up my guy!
@@mrgomer64 Thanks!
Got any suggestions for other things to investigate?
@@SeaJayPlaysbest way to find certain structures
Wow! The best and most comprehensive video on this subject ive seen, by far! Im in bedrock so cannot use the 1x1 methods. But ive always mined on the layer or 2 above bedrock using 2x1 branch. Will certainly try Y57 2x1 tunnelling now.
@@killamackay Thank you! Based on the comments, there’s a lot more I could do and test so I want to do a follow up. I’ll make sure to include results for bedrock too!
Mining a 1x1 tunnel does expose five blocks per block mined. It's just that one of those blocks gets mined after you expose it. You get the diamonds whether or not it's in your mining path, so I see no reason not to count it.
@@DanielLCarrier yeah I think it’s fair to count it as 5 instead of 4. I’ve got a pinned comment explaining why I didn’t count the mined block if you want to check that out
ocupas estudiar límites
I am so glad there are people whose minds work like this. If everyone was like me we would all be head butting trees and eating whatever fell out.
majority of poluation is capable of this, most of them just aren't interested
Hi, hello I'm the gal who is interested in the standard deviation; I just wanted at add a little bit more analysis, because I believe that your conclusion is wrong due to your low sample size, and the difference in gap size is insignificant. For my analysis i compared 0, 2, 3-block gaps with the 1-block gap and used a confidence requirement of 95% to determine weather your results were accurate or not. I chose the null hypothesis H0: x_1 = x_{0,2,3}, and alternate hypothesis Ha: x_1 ~= x_{0,2,3}. I did three tests, comparing 1-Block gap, 2-Block gap, and 3-Block gap. The averages and standard deviation were used from 21:28. Since n = 10, the degrees of freedom = 10 - 1 = 9. Calculating the T test values yields critical values for a one tail test yields +-2.26 (either plus or minus, depending on the test). The two sample test scores were as follows:
T_0 = (71.767 - 88.179 ) / sqrt(17.443^2/10 +43.700^2/10) = -1.1
T_2 = (71.767 - 95.158) / sqrt(17.443^2/10 + 48.227^2/10) = -1.44
T_3 = (71.767 - 108.76) / sqrt(17.443^2/10 + 75.516^2/10) = -1.509
since the values fall between +-2.26,6 we must fail to reject the null hypothesis, thus these methods are not measurably better or worse than each other
You basically did a 10fold cross validation. You should have created some boxplots with the 10 layers of each. That would have been cool to see!
Amazing video. Time flyed and I didn’t realize how long it was because I was so engrossed
I appreciate that, thanks for watching it!
Well done, I honestly want to appreciate your work. It must have taken so much time and nerves, but you got it.
@@greatdanchik Thanks! Yeah it was kinda a lot and i definitely won’t be doing any more manual mining tests lol but I will be doing a follow up (code based) to hopefully get more and better data
The limits are incorrect, you should divide the expressions by x to fix them, but apart from that detail, it’s 100% fine! Great content
@@miguelcerne1150 yep, that’s what I forgot to do lol thanks!
I love how you brought in a calculus limit. I was thinking the same thing!
Thanks, I’m hoping to show more next time. But i did forget something, the formulas in both limits need to be over x otherwise they go to infinity lol
Yeah I was going to say something lol @SeaJayPlays
Amazing work. Hard to argue with your conclusion.
great video! whole different vibe watching this at 3pm. btw first time outside school seeing lim to be actually used
@@grey521.mp4 I appreciate it!
I did make a mistake on the limits though lol, I forgot to put the equations over x. Otherwise they’ll balloon to infinity
Magnificent analysis!
Now do it again with water and lava diving (diamonds generate more frequently when exposed on all side, and that includes water and lava) >:D
That and running around caves with or without a mob switch
EDIT: You just mentioned that you wanted to test that, woops
i really loved this video, hope you will reach more people with you accurancy. And i would love to see a tunnel mining vs cave mining video, I've been thinking about this for a while now and I can't find an answer.
Keep going!
@@danielepaschetta4742 thanks so much!
I’ve been thinking about how to test cave mining for a while now, but can’t think of a good way to do it. And I really don’t want to actually cave mine in-game for a hundred hours 😂
I’ve got a couple ideas that I need to look into more, so hopefully I can make a video on it
If you are making a part 2 to this, you should mention moss mining too. I'm curious about how it compares to strip mining
@@TraitorousHomeworlder based on all the comments recently, I’ll definitely have to do a follow up lol
I’ve never heard of moss mining, but I’ll make sure to check it out!
Fascinating video ! I too always thought that the 2×1 tunnel with branches was more efficient, I know I won't be doing that anymore.
I wonder how feasible it would be to run that whole process programmatically in order to get the numbers for an even higher sample size, as well as test with efficiency 5 pickaxes.
@@witherflower42 yeah, I was a bit sad it wasn’t very efficient.
I’ve been thinking about how to do this recently, especially with cave mining. I don’t know how Minecraft generates the diamonds so I’m not sure I could replicate it in python. I haven’t looked into mods yet or even making mods, so that could be a solution.
If I do an updated video, I’ll be sure to test out efficiencies!
@@SeaJayPlays From my experience with datapack I'd say this would be a problem that could be fixed with those. In them you could iterate through your wool blocks and check each adjacent block whether it's a diamond ore
@@kofidrng That’s really exciting to hear! I’ll have to look into making a data pack then!
Also another point -- one of the important things you didn't really consider is space efficiency of each of these methods and how long it takes to reach untouched rock. If you can get super fast diamonds while you are mining then that's great, but if you're crawling through an infinitely long tunnel to get there then that's no real use. I'll probably start using 1x1 mining for this video, but I might try a layered version, and have major routes that I normally would go down every time as 2x1 as that's faster (without elytra/dolphin tunnels/etc).
I’m not sure how this would play out but if I end up making a follow up video I’ll make sure to look into this!
Doing gods work for the community ❤
bro when you showed that stats screen i fell in love with you thanks you for beautiful statistics :D
@@lejgaming7031 thanks! If you like seeing data like this, I’ve got some farm optimization videos you might find interesting!
thank you for this! I realize how much work this was. very cool and helpful
glad research is being done on this subject
Nice work! I know I'm late to the party but I'd like to ask about your choice of metric regarding blocks exposed per block mined vs blocks exposed per second. Considering that to go from one to the other you only have to multiply or divide by the average blocks mined per second, I think that blocks exposed per block mined is a better metric, since the other choice introduces a lot of subjectiveness/uncertainty. For example:
* What is the time spent moving your cursor between mining blocks?
* What gear do you have? Iron pickaxe? Diamond/netherite? Enchanted?
* Do you have a haste 2 beacon or not?
* How consistent you are between measurements
All of these factors affect the metric, and seeing as your final standard deviations were quite large I'd wager that these uncertainties can't really be ignored, but I may be wrong
Edit: Nevermind about the standard deviations, I missed that they were between layers and not measurement runs. But I think the point still stands regarding blocks exposed per block mined being the more objective metric
@@tomwaligorski9882 Thanks for this comment, you bring up a good point! I’d say I chose the diamonds/sec metric because it does try to account for ‘real world’ game play. I think some of the biggest differences between the mining techniques come from how much time is spent using them. For example, I wanted to include time spent placing torches (you technically don’t need to place torches with the 1x1 so that could increase its speed), and time spent turning for the branch mining methods.
You’re right that it introduces a lot of inconsistency on my part in measuring that, and after getting feedback on this video, I’m convinced I need to do a follow up video/better version that measures things like that more consistently.
I might have to get some community help in collecting time data for various tooling/effects lol
i'd love to see a test for other ways of getting diamonds. maybe use some code to ingest one thousand hours of someone caving and one thousand hours of someone end busting, and also i wanna see some calculation for the 3 most common designs of tunnel bores, with 1 module, 4 chunks, 8 chunks, and 12 chunks, and also checking how many times you should fire the tunnel bore before checking for diamonds.
personally i went with the newest 2 block high tunnel bore, just high enough to not see bedrock, 8 chunks long, and 2 fires per scout. i've had some tests to see what's faster, caving in an ancient city with a warden switch, or using the tunnel bore, and got to the conclusion that if you include time between each ancient city, its roughly the same amount of diamonds per hour, but my data is only roughly 5 hours on each method, and isn't really scientific in nature.
That’s a good idea and I do want to make a follow up video (code-based) going over more variables. Stay tuned!
Would love to see these stats with other popular mining patterns.
Windmill especially, which attempts to optimize walking time, compared to straight branches, which typically involves some degree of backtracking. (I guess you could zig-zag a branch mine, but windmill also aims to optimize the walk-back to start by mining in all 4 directions, and creating a diagonal path back from any of the corners.)
I do plan on doing a follow up to test a lot more methods, and I’ll make sure to add this to the list
Amazing and entertaining video, enjoy the algorithm boost this is criminally underrated
Thanks!
Fantastic! Thank you! Been looking for a deeper analysis and this hit the spot hahaha
Glad to hear it lol
very thorough, clean presentation!
Thanks! Hopefully the next one is better!
This notation was really confusing to me at first, because I kept expecting tunnel height to be mentioned. Some of my mines have 1x2x1 and others are 1x3x1, because having more headroom lets you run faster. But it also means mining an extra block, so I usually skip it and make the tunnel only 2 blocks tall. So the 1x1 versus 2x1 confused me, because my tunnels never end up being double-wide. I can see how it would help, but 2x2x1 and 1x2x2 tunnels are only used to connect my mines to each other.
Btw, you should mine with a beacon, use 1x1 tunnels, BUT mine in a "horse circle", meaning you make a
⬛⬛⬜⬛
⬜⬛⬛⬛
⬛⬛⬛⬜
⬛⬜⬛⬛
pattern. This makes your tunnels get a wider area.
And you mine faster with the beacon, AND you get everything in those 4.5 layers mined.
The only thing more efficient is a tunnel bore that is in a world without lava.
@@DevidCipher I do want to test how status effects affect the results. But I’d say that pattern probably isn’t the most efficient. You can expose all the blocks in a volume by leaving two blocks in between tunnels on the horizontals, then shifting that pattern for the vertical layers
How is mining with a beacon even done? You'd quickly leave the range of the beacon.
I feel like since villager trades, diamonds are far less needed, instead i would really like a in depth netherite video, since there is no other way to obtain it rather than mining. Still a great video btw good job.
@@daviddansD I’ve never mined for netherite before so that would be a good chance to check it out. I’ll add that to the list, thanks!
6:16 The blocks in fornt also count as exposed blocks because you dont know them. Also when you counted the blue glass and the blue wool and divided them, you didnt count the wool blocks them selves as exposed blocks.
So the exposed blocks/ blocks mined ratio for 1x1 tunels is 5:1. The 2x2 ratio is 8:2, same as 4:1.
I would appreciate you answering to confirm what i am saying.
I also want to thank you for doing this types of videos that are actually interesting.
@@lucasbarroso2997 Hey, this is a good question, and I can definitely understand your logic.
My answer would be that it depends on how you want to define exposed blocks. In my opinion, what matters is the blocks revealed as a result of your mining (so blocks that aren’t required to be mined). I’m gonna have to mine my tunnel no matter what, so to me that doesn’t count. Maybe a better way of phrasing it would be the blocks exposed after each block mined.
Luckily, either way, the trends of the results stay the same!
Yes the results are the same. Now I can understand better the way you thought about the exposed blocks. But I keep thinking Its 1 more exposed than you say
No matter what your way is valid in this context because as you said the results are still the same, not proporcionally but in the same order.
And also, keep the amazing content.
@@lucasbarroso2997 I appreciate your explanation, it makes a lot of sense and I think either way is a good one to have
@SeaJayPlays Hi. I like to find someone as nerdy as me when it comes to games. I see your point. By "exposed" you mean revealed for free, without mining time used. That's fair. When it comes to the "exposed/s" metric, it's quite misleading thought because people might think it means the (exposed + mined)/s metric, which it is not.
@@luminas-d9w Yeah, I’ll just have to clarify more if I do a follow up video
Blocks exposed per second was what u wanted to hear thank god.
I found people stupid who crawl very slow just to get a better ratio on blocks exposed:mined blocks.
Love u bro.
I think the #1 thing videos like this overlook is taking advantage of world gen behavior - we also need to look at each method when doing specific things with knowledge about world gen. Ex: as previously mentioned being between chunks, or the method of mining a chunk until you find the vein (and for this one, the optimal amount you should search before giving up & moving to the next), or how each method stacks up when leaving your current chunk as soon as you find a vein there ect... I think it makes a big impact overall
You absolutely should not being mining a chunk until it's drained, that'd be crazy slow. And mining on the borders of chunks is slim to none difference. The average is effectively the same, but there's a higher potential that 2 or even 4 diamonds veins spawn against eachother
@@Nate-bd8fg show the math, then :p
How did you come up with "effectively the same" other than 'it seems right'?
@@kirasmith1147 This is a good point that I want to test further in a follow up video. I just haven’t been able to find any answer as to how diamonds are generated (specifically) so it’ll take some work to get that data
@@Nate-bd8fg You could use a method where you mine on the edge of one chunk, and then mine into the adjacent chunk five blocks in. You would then skip the rest of that chunk if you find a big vein of diamonds.
I'd consider this too mind-intensive, keeping track of chunk numbers and whatnot. And if you're going to use mods to help with that stuff, might as well just use X-ray. Or creative mode.