Apologies if it sounds a little echoey at the beginning, I’m working on solving this issue with my microphone. It’s weird because it gets better across the video.
I'm not particularly learning chess, but improving my English listening. You're such a perfect professional speaker like National Geographic or something similar. Nice to hear you!
woah. found your channel three hours ago. and my first video was the stockfish not actually stockfish video. I have no understanding of gambits or advanced chess talk but I'm able to follow along with these videos anyway, good stuff
Theory is that in 14:23 the bot cant find a move that could make his position better or ur position worse so it plays any random move to see what the opponent will play next (aka both positions cant make any progress)
Do it with pondering turned on (that means it continued calculating during your turn - the longer it takes for you to play your move, the stronger its reaction).
Am I missing something at 22:52 or does Qb3 just trap the bishop on b2? Either way, these videos are fun and you’ve got great quality for a smaller channel. Definitely subbed.
14:04 This is likely documented on ChessProgramming. Probably the same people who document CCRL. My best guess as to why SF specifically does it, boils down to the implementation of AI, after they lost to AI some years ago, against Leela (or AlphaZero iIrc?). A gross underexplanation would be saying attack maps, but dynamic. Basically, the AIs, in engines, dynamically consider squares for pieces in a priority list. All squares have a "value", that changes over time. The engine's goal is to maximize the value of all these squares, in its favor. It's also among the reasons that eval bars can look so vague. Sometimes, when everything is in the balance, the best way to approach the position is to advance the flank, because it maximizes the square values. All of them. In different ways, for different reasons. You can talk to a SF dev on their Discord for more detail. Most contributers are really nice people.
You said it yourself in the endgame of the first match. The engine has already considered the outcome of the endgame, long before you reached it. Especially if you had ponder on, like the previous SF video. It considered the likelihood of winning to be larger, when it pushed the flank.
@@iwantcheckmate Again, grossly underexplained. I've only ever made one engine; no AI. It's not on CCRL, and it definitely doesn't qualify for CCRL. Meanwhile, you have some SF contributors who have been a part of SF, Leela _and_ Torch, not to mention their own engines. Criminally underrated in their field, and a lovely rabbit hole to go down. (I mean, it's mostly pain, but we have fun)
interesting that for you the magnus bot was more difficult for you in terms of queens required. I would guess it plays more conservatively since it can see all the dangers of the two queens but doesn’t really make sense with what we saw during this game.
@iwantcheckmate “one of my old videos” sir its been a month😂 if you have made this much progress against bots in a month stockfish should be terrified of you
Great video! I don’t know if this is too much trouble for you, but could you put the pgn of the games in your videos in the description or a comment for those of us who want to analyse the games?
I really liked your ladder climbing video, will you go back to that style? As much as I enjoy these sorta meme games against bots, games against humans with analysis as you go is super informative as well as entertaining.
I really enjoy your videos, I recently started playing chess again (1-2 weeks ago). I think the audio of the video has a little echo, other than that as usual a good video. Edit: I just saw the pinned comment, nvm
17:53 Actually stockfish is "only" looking 35 half-moves in the future. Even that not for every possible variations. They using a pruning method (the real "magic" behind engines). So basically they evaluate about 8-10 half-moves and cut off the ineffective ones. As faster and better this preselection is, the more time the engine has to focus on the promissing lines. To get an even better engine you can give stockfish more time per half-move.
What's up with the audio?
Apologies if it sounds a little echoey at the beginning, I’m working on solving this issue with my microphone. It’s weird because it gets better across the video.
Lol, thought it was my AirPods
What I thought the audio was normal????
@@iwantcheckmate it returned 2 monke
Beat Martin without losing any piece
this exactly what i commented last video we need this bro😭 would be fire
Well scholars mate = easy
@@nfskull9900 i said something like it too on another video. We need him to do this
@@matveyokishev6504 ngl idk what that is i don’t play chess i just watch it
how many kings to beat stockfish 🧐
he mated the black king on the spot where the white king begins
I'm not particularly learning chess, but improving my English listening. You're such a perfect professional speaker like National Geographic or something similar. Nice to hear you!
Haha thank you! :)
woah. found your channel three hours ago. and my first video was the stockfish not actually stockfish video. I have no understanding of gambits or advanced chess talk but I'm able to follow along with these videos anyway, good stuff
Thanks! Glad to hear it and thanks for the kind comment :)
Stockfish: You can't defeat me
iwantcheckmate: I know, but he can
iwantcheckmate with extra queen emerges
Someone LITERALLY someone(or something ig): *Wins*
Someone LITERALLY someone(or something ig): Ya-
The website: Fanfare!
The website: *Plays an extremely unneccessary trumpet fanfare*
Theory is that in 14:23 the bot cant find a move that could make his position better or ur position worse so it plays any random move to see what the opponent will play next (aka both positions cant make any progress)
That’s interesting and makes sense, thanks!
I don't think humans will ever win against robots
Video idea: remove one of Stockfish’s pieces until you win
Excellent idea
Do it with pondering turned on (that means it continued calculating during your turn - the longer it takes for you to play your move, the stronger its reaction).
Am I missing something at 22:52 or does Qb3 just trap the bishop on b2?
Either way, these videos are fun and you’ve got great quality for a smaller channel. Definitely subbed.
That’s a great find, I hadn’t noticed that! And thank you :)
I am really impressed with this. You barely needed any help!
Thank you!
can't wait for part 4 XD
I like your videos. U should start some recap videos of some old games. That would be great to watch.
Hi! Recap videos of older games on this channel?
@@iwantcheckmate why not ?😁
This and outright ignite are my fav chess channels
Let’s goooo
I love these new mini series
0:00 already a good video
Thank you haha
@ I thought it was relatable to say that so that's why 0:00 was already the best
This guy is always posting new vids. Keep the grind up man
Thanks! 🤝
Playing with the big bots now huh IWC!? I absolutely love watching people or bots play Stockfish 17, its alway so interesting.
It wasn’t fun to set up but it was so cool when I finally did it. I’ll admit I wasted some time just testing it out for fun hahaha
Algorithm showed me this and it's great, keep up the good work!
Thanks, will do :D
Are you gonna do streaming content in near future aswell? Like yk grinding elo on stream.
Also Congrats on 4k subsss 🎉✨
Potentially! Looking for a good time to work it in. And thanks :D
Looking forward to it ❤
i like your voice, very soothing :)
Thank you :)
14:04 This is likely documented on ChessProgramming. Probably the same people who document CCRL. My best guess as to why SF specifically does it, boils down to the implementation of AI, after they lost to AI some years ago, against Leela (or AlphaZero iIrc?). A gross underexplanation would be saying attack maps, but dynamic. Basically, the AIs, in engines, dynamically consider squares for pieces in a priority list. All squares have a "value", that changes over time. The engine's goal is to maximize the value of all these squares, in its favor. It's also among the reasons that eval bars can look so vague. Sometimes, when everything is in the balance, the best way to approach the position is to advance the flank, because it maximizes the square values. All of them. In different ways, for different reasons. You can talk to a SF dev on their Discord for more detail. Most contributers are really nice people.
You said it yourself in the endgame of the first match. The engine has already considered the outcome of the endgame, long before you reached it. Especially if you had ponder on, like the previous SF video. It considered the likelihood of winning to be larger, when it pushed the flank.
Thank you for the insight! That’s really interesting and it makes sense that it would look for those little incremental improvements.
@@iwantcheckmate Again, grossly underexplained. I've only ever made one engine; no AI. It's not on CCRL, and it definitely doesn't qualify for CCRL. Meanwhile, you have some SF contributors who have been a part of SF, Leela _and_ Torch, not to mention their own engines. Criminally underrated in their field, and a lovely rabbit hole to go down. (I mean, it's mostly pain, but we have fun)
interesting that for you the magnus bot was more difficult for you in terms of queens required. I would guess it plays more conservatively since it can see all the dangers of the two queens but doesn’t really make sense with what we saw during this game.
That was also my first video ever so I didn’t really know what to expect or look for, but I’ve made many similar videos since
@iwantcheckmate “one of my old videos” sir its been a month😂 if you have made this much progress against bots in a month stockfish should be terrified of you
no game analysis anymore?🙁
Good point forgot to add it :(
1:37 B5
How much time did you give stockfish 17 to think? I usually put 90 seconds
I did it by search depth instead of time, so depth 35
You are very quick to respond when someone comments on your videos.
Haha I try my best
Remember me when ur famous!!
Ok
You got it hahaha
Great video! I don’t know if this is too much trouble for you, but could you put the pgn of the games in your videos in the description or a comment for those of us who want to analyse the games?
That’s not a bad idea at all, I’ll see into this
The first game wasn't actually bad! as a 1500, surviving for that long against stockfish 17 is impossible.
Hahaha
I really liked your ladder climbing video, will you go back to that style? As much as I enjoy these sorta meme games against bots, games against humans with analysis as you go is super informative as well as entertaining.
I’ll have to go back to those then, I’ve received a few requests for it :)
How many bishops to beat stockfish video
I really enjoy your videos, I recently started playing chess again (1-2 weeks ago).
I think the audio of the video has a little echo, other than that as usual a good video.
Edit: I just saw the pinned comment, nvm
No worries, and thank you for the kind comment :)
Another great video let’s go, bro eventually you’re not gonna be able to look at all these comments you’re gonna be too famous
Haha I’m checking regularly to keep up 😅
My guess: 0 😊
great youtuber and you seem like a great guy. whats your FIDE ranking, btw?
Thanks! I have a uscf rating of mid 1800
Play against any strong bot, but bot is forced to play polish gambit (it is poor opening)
Definitely enjoy the content but I think for viewers sake and better understanding you should have an evaluation bar
Video idea: speedrun every intermediate bot.
Keep up the great work❤️
Will do! And thanks for the idea!
Bread🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞
Bread🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞
Bread🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞
Bread🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞
Bread🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞🍞
That’s a lot of bread
W audio quality. Is bro talking into a bucket? All jokes b4 some1 gets offended
Lmao sorry about that!
@@iwantcheckmate It's okay. We all know you wanted to put stockfish in that BUCKET instead
Under one hour
Let’s goooo
Man is playing chess in a cave 😂
I am never coming out
Can I get my 12th hello?
Also, I see youre growing really fast, do you think theres gonna be a time when you wont give me my hellos anymore?
Not any time soon! I’m too quick hahaha so here’s hello number 12!
How old are you and where are you from? Face reveal when?
Haha not sure if I would do a face reveal!
Queens get more clicks!
Haha exactly!
So you can beat stockfish17, but you will never beat stockfish18 (me) even with 50 extra queens
One can dream 🙏
Bro I see you in comments every vid let’s gooo this is proof that’s the fans are true fans
Yaay 36 minute ago I’ve been wondering if u can help me get good at chess cus u have way more ELO than me I have 428
Dude I love your videos, but please change your theme back 😂❤
My theme! What do you mean haha
@iwantcheckmate I mean your game theme. The pieces dude 😂
Oh! Haha yes I am using ChessX so I can play Stockfish 17. Most future videos will use chess.com again :)
17:53 Actually stockfish is "only" looking 35 half-moves in the future. Even that not for every possible variations. They using a pruning method (the real "magic" behind engines). So basically they evaluate about 8-10 half-moves and cut off the ineffective ones. As faster and better this preselection is, the more time the engine has to focus on the promissing lines. To get an even better engine you can give stockfish more time per half-move.
Thank you! That’s really good to know. I’ll keep this in mind for future videos
👍
Worst part of the video 29:52
I agree 🤝🤝