Dragster - Speed Theory
Вставка
- Опубліковано 9 лют 2025
- This is a (not-so-quick) video discussing some of the theoretical limitations of Dragster. The main point of this is to help understanding of the possibilities and encourage further discussion.
I'm sorry if it's boring :(
CORRECTION: In the video I quote the speed ceiling for Gear 3 as 130. It is in fact 126. Either way, this error was in favor of a faster time, so it doesn't do anything to change the results.
CORRECTION 2: The distance addition actually occurs before the speed increment, so the correct formula would be prior speed + prior distance, not current speed. Just means that any of the frame numbers I discuss are in fact one frame slower, ie 5.51 to a 5.54.
whoa, coming from somebody who hasn't ever played Dragster, this was really well explained! It's so cool how much programming you can actually do with Excel cell commands
If Todd was smart he should say that his 5.51 was a typo of what should have been a 5.57 score. Mistaking the 1 for a 7 can happen, however he erroneously challenges the situation and sticks with the 5.51 story making him a complete fool.
But if his score only 5.57,people wont call him the king of video game
also he wont have the longest unbeaten record!!
fame > truth
Since 5.57 is the best possible score. It would still be unbeaten. It can only be tied,
Sry bae. it might be 1 girl. or 7, i dunno.
Very good! But don't give that goof any ideas ;-) Unrelated, I bet 25 push ups on command that he gave himself his ridiculous nick names. I don't know why, he just strikes as the type.
@@PerMejdal except he wasn't the first person to achieve the 5.57, so he tied the two previous record holders (if he did in fact even achieve a 5.57)
i just spent 20 mins looking at numbers from an 1980's game
IT was great
The music helped.
Poke711 *a
Lol
After witnessing how you handled that six month long ordeal/thread over at TG, you have made a lifelong follower of me. Your tact and patience in the face of unscientific, personal attacks from a supposed professional was inspiring. Your work is comprehensive and interesting. Please continue being interesting and thoughtful!
DichotomousRex Well said!
At least he can still hold the world record in longest ongoing video game related con.
When I realised what you were doing was emulating the game in a spreadsheet, well, I don’t think I’ve ever seen anything in my life like this before.
This is some next level stuff.
Whats funny is that Todd had a 5.57 listed in Activisions newsletter prior to the fake 5.51
Basically people like Billy Mitchell and Tood Rodgers have ruined it for anyone who did have genuine scores in the 1980s
Did you get sued yet
@@Jonatello Obviously not. How can they sue when they both cheated?? Their scores aren't even recognized any more officially
Judging by the single downtick
Todd Rodgers has been here.
It is Rogers No d
Actually, according to one of his new lawsuits, his name is Todd Todgers.
When you played the Back to the Future theme as you discussed the win condition, I felt great happiness.
I did a 4.96 back in 82.
For anybody that doesn’t believe me I have 3 words
Believe my words
Todd is a liar and Twin Galaxies is full of hot air now. TG of the 80's is dead.
Mondester Or maybe his calculations are wrong O_o
No, they're not. Unless you're a moron, the calculations are very clear and concise. 5.57 is the absolute lowest possible time, no ifs, ands, or buts.
If he's unsure and he created the program that leaves a lot of doubt. Seems like the "moron's" just haven't figured out how to get a faster pass.
He didn't just make all this up on a hunch. The actual game code (which is easy to figure out, since the game is ancient and very simple) makes it *impossible* to do better than 5.57. It's just not a thing that can be done. It's like saying Usain Bolt could get a 3 second 100m run if he just "figured out how to get faster". That's stupid. It is physically not possible to do (unless you're a big fat liar/cheater). Stop playing dumb.
Hi Todd! You suck at video games and twin galaxies is a joke but you're rocking the greasy porn-shop owner look.
8:23 The math behind Activision's original (flawed) "simulated" best time.
Thank you for this video! I'm hoping it clears up some things about the dispute.
Todd Todgers
Todd Roges
Tried this game online. It's not much of a looker, is it... The fact that some hang their entire existence on whether or not they are the BEST at playing this footnote of a game is really sad. I guess it ties in with the great tradition of desperately trying to make your 15 minutes of fame turn into half an hour or more. Americans in particular seem to desperately need fame as it their life depends on it. In a country where there is little middle ground between ridiculously overpaid and flipping burgers, it's quite understandable.
Does anyone know the name of the song that starts playing at 08:40?
It sounds familiar to me but I can't remember where it is from.
Smartest Gamer Of All Time
Never ceases to amaze me what bored engineers do when they have a passion for something
we need a TAS to figure this one out. maybe a mem overflow somewhere
There are no overflows or other major technical bugs in this game.
Omnigamer for as much as common sense tells me “a 6502 game this simple is probably bulletproof”, I’m reticent to put an atari game in the same category of software as the firmware on a car’s ECU, being “on paper no software is infallible, but let’s pretend this stuff is, because if it’s not, we’re in trouble.”
To give an atari game such credit just doesn’t seem right.
Additionally, the hardware itself isn’t infallible, if it’s possible to hit one of these errors from within the scope of dragster is another question.
I want to believe.
Not that the 5.51 happened, but that it, or something similar, may be theoretically possible.
@@funposting8912 It's 100 lines of code. Come on now.
@@DrynSarcastic360 it’s 1,770 bytes of hand-assembled 6502 instruction, if you think there isn’t room for error in that I’d guess you’ve never written 6502 before.
The Furby was programmed in much the same way, and it doesn’t take long to find someone who’ll attest to how deranged they used to act.
It seems like the only way that it's possible is if there was a hardware malfunction that allowed a 5.51 to occur. It seems between you and Ben heck that it has been shown that 5.51 is not possible. It seems like the only way it would be possible would be with some form of consistent Hardware glitch that only existed on one or two older systems. Todd was pretty willing to come on the Ben heck show to prove or disprove the claim. Maybe a glitch put a one where a seven was or a hardware glitch allowed the 5.51 time to occur.
And if he was using a glitched system, it would explain his other disproven records. If you get some dust into the cartridge connectors, weird shit CAN occur. Kids these days, who never had to blow dust away from a cartridge don't understand that. It could also be a typo and memory can fail you.
If there was a hardware malfunction, then the record shoul also be invalidated as well, since the record was supposed to be in the, "no glitches", category.
Garry Guerra yep if this is the case which it probably wasnt his record would still be invalid
The chance that hardware malfunction results in a faster run is very very small.
Ok so here are the possibilities :
One, he didn't cheat and all his impossible records came from a _specific_ hardwear glitch that has never been replicated nor has the console in question ever been turned up. This also has to take into account that his other impossible records were affected as well by this (having a score that ends with an 8... in a game that only scores with multiples of 5). This hardwear glitch would have to be messing with the console's math, an error that big would also cause many more problems besides an impossible run. *Even then, the records should still be taken down because this was a no glitch record and it gives him an unfair advantage.*
Two, he lied. He photoshopped his recent 5.54 after all, and having so many runs be illegitimate raises many questions. All of his disputed records are physically impossible without glitching and even then, the glitches are entirely unknown and he cannot reproduce them. No video evidence exists nor photos (that arent clearly doctored).
He is either lying through his teeth or he never deserved his records anyways. Pick your poison.
Rodd Todgers is done for
But you forgot to start excel in 2nd gear
I just watched that interview, lol
3:49 is this song from Mischief Makers for the N64?
Holy shit. I thought, I was the only one to notice it.
My question is where those witnesses who saw him do the trick are. Theres even a news reporter from the 80s talking about how hes going to be at the electronics show to perform the trick i want to know what happened on that day
There were no witnesses because it never happened. It was just him embellishing the story.
One thumb down ... must be Todd
A second one now .. must be the Twin Galaxies CEO
Dark Tenka fourth one.....must be Twin Galaxies.
Ahahahah very underrated comment.
You're on Wikipedia!
The site anyone can edit!
Mile Noctis And the team that "decides" what is relevant is also often VERY biased when it comes to social issues like this. This *SHOULD* be attached to Rogers' Wikipedia page since it is a factual removal of credit to Rogers' of his claim to fame in the first place. This removes the reason that Rogers even has a Wikipedia page in the first place. Whatever editor at Wikipedia oversaw that removal, is doing the page a disservice.
Thankfully, they've since got a brain and added that back in again.
More videos like this please. :D
+1 for Blue Dragon music
I'm glad someone else noticed :)
Just a little over my head. I thought this was going to explain every possible outcome to disprove Todd "Mr. Activision" Rogers lol.
Do not violate the rules of Dragsterdynamics!
:0 blue dragon music in the background?? Awesome !!
The whole point is that people should never have believed him in the first place. It makes no sense that they did.
Great video!
You said the speed increase slightly more then double. 8:00 it goes from 64-130-252?? 252 is less then double 130. Plz explain
Please note the corrections in the video description! I was going off of memory for most of the video, so a few things weren't quite right. The end points are still valid though.
Way to go OmniGamer!
Maybe I don't fully understand, but does it make sense to say the first shifted gear happens in the 0th frame? Could it not only happen as of the first frame?
You can start the input for the gear press at any time before the race start, but you can't complete it (ie, release it) until the countdown hits 0. Frame 1 with time .0334 is the first frame that release won't cause an early flag. Frame 0 in the spreadsheet simply represents any activity before the countdown finishes.
Omnigamer so then you can actually shift into gear by releasing on the 0 frame, and continuing off from there. I know this was excel and beat scenarios, I was just confused as I thought the best scenario would be on the first frame, making the time even worse. Love the work you put into this though. Kudos and thanks for the reply explaining.
No
I almost want to debug this shit out for myself. I am a fairly competent hacker myself, having done quite a lot of research on Zelda64 (function documenting, figuring out how that pesky "wrong warp" glitch works, partial discovery of the GIM glitch). The hardest part for me to get starting with something like this is probably
1) motivation to stop hacking Ocarina of Time
2) setting up an emulator
3) learning "atari 2600" asm
If I had a solid list of links to tools and docs, and perhaps your disassembly notes, it would make it easy for me to get a sense of "context" and start poking into things.
My biggest worry is that there really is some way to get a 5.51, but only through some obscure hardware quirk.
It's actually pretty easy to get started with Atari reverse engineering - it's a MOS 6502 family instruction set, just like NES and SNES. It's very straightforward to keep track of things with an accumulator-based architecture like that.
The tool I used for this was IDA Pro for static analysis, and BizHawk for checking general function. IDA Pro has something of a learning curve, and isn't free, but there should be comparable alternatives out there. I don't know any off the top of my head though.
If there is any especially obtuse bug or quirk that allows a 5.51, it would almost certainly allow for significantly faster than that as well. The most glaring possibility for this is starting before the countdown ends, which I've examined pretty heavily. Everything else would require busted hardware or a corrupt ROM...
Darb and Omnigamer both have 5.51 on real hardware!
5.57 not 5.51
Paul-Stelian Olaru See above comment! :p
It somehow feels like this is a "how pixels should have ended" or a "pixels was bad" review. Then the actors tried to defend it and we dislike em even more. yet, Pacman is cool, some graphics in the movie are cool and games in the real world idea is kinda ok.
Ow wait, this is about Dragster, the game is cool, setting a record is cool... cheaters... those arent cool.
I only play 128 tick dragster.
music at 13:53?
Interesting video.
5.51 is a fake
Todd Rodgers? More like Fraud Rodgers amirite?
So, what specifically prevents the TAS from getting 5.54? Were you just underestimating how much the game pulls you back toward the speed limit in these calculations?
Also, its kind of ironic that Todd Rogers' faked score would actually be the the fastest possible time without that mechanic coming into play. If it weren't for all of the other clear evidence stacked against him and his outright lies (not to mention making up the whole "starting in 2nd gear" thing), I might be saying that perhaps he was playing a different build of the game that didn't require the extra shifts.
Please note the corrections in the video description! Everything I mention in the video is off by 1 frame, so all the times would in fact be slower than I discuss.
Omnigamer
Oh, got it! Surely that explains, then, how Activision came to the (wrong) conclusion that a 5.54 was the best possible time.
I love this.
digging the zelda OST
nice work
So playing this in Australia on a 50hz tv... I cant go as fast?
You can, in fact! The PAL versions of the game weren't edited to change the timing at all, so you can still achieve 5.57. It is also slightly easier to do so, since you'd have an extra 3.3 ms each frame to make correct inputs.
Omnigamer oh ok, hmm, so if one was to use a pal cart in an ntsc machine or vice versa (assuming they are different) would it be possible to run a faster time?
Also, how about a voltage difference to the console?
The ecu in my car was logging a transmission fault because of the time it took to shift into torque converter lockup, i replaced the transmission and it still set the fault code, problem turned out to be my alternator was charging 18v most of the time and then occasionally dropping to 9.3v,
(faulty regulator and worn brushes) which then caused it to think the shift was lagging and set a fault code for transmission slip.
Wonder if a similar voltage change would cause a different possible time to be achieved?
Voltage differences and PAL on NTSC would not affect timing, since it counts loop rather than conforming to some real time reference. The electronics issue you mention with your car is a very different circumstance, and wouldn't happen the same way with an Atari.
Omnigamer oh ok, yeah i wasn't sure hence asking haha
Just trying to come up with a scenario where it is possible to exploit the system in some way to achieve a faster time, hmm, theres another thought, dont suppose a game cart could be modified in some way to achieve the 5.51 that is seemingly impossible? I have always speculated that Rogers was a marketing tool by Atari/Activision to boost sales or at least interest in their games...
If they gave him a special cart, with some sort of different calibration, he could have set the fast time, and naturally it cant be repeated on a regular cart...
Still as cheating as it gets, in no way "sticking up for him" but its a plausible possibility...
Why can’t i find any 5.54 runs again? Is it the lag?
What's demonstrated in this spreadsheet is a simplified theoretical test that removes some of the penalties that are part of the actual game. With the full mechanics, including penalties, 5.57 is the best possible time.
Omnigamer ahh, thanks a lot man.
Well getting a 5.51 is actually doable.
All you need to do is to start in 2nd gear.
No cheats, just a little human element envolved here.
Im wondering would it be possible to get such a score by modifying the games code? I know thats still cheating but im trying to figure out how he cheated infront of supposedly a bunch of witnesses. Also maybe activision was in on the whole thing too as a publicity stunt
Modifying the game code would allow you to do anything you want. Even relatively few changes would allow for very different times.
Again, there were no witnesses, that was a one-sided claim that has historical inconsistencies and a complete lack of substantiation. He never achieved his claimed time in front of others, and he certainly didn't know any special tricks that gave him better play.
So wait. 5.54 is humanly possible?
No - please note that all the calculations shown here are in the absence of critical game mechanics that will slow you down. Also note the corrections in the video description; I didn't have all of my notes available while making this video. 5.57 is the fastest possible time, human or otherwise.
Omnigamer wait what about Activision? How did they get 5.54
The easiest answer is they didn't. As best we can tell, the 5.54 claim only came from Todd. If they did any kind of simulations at all, the best we can guess is that they performed some basic math like is shown in this video and just used that, but there's no reference at all to corroborate.
Omnigamer if so, i dunno why they would lie about 5.54
I'm sure they said something about it at some point
This video kills Todd
Wow 5.17 ISN'T EVEN POSSIBLE!
5.51 is the minimum, lol. Oh well haha
Please note the corrections in the video description - any of the times noted are actually off by 1 frame, toward being slower. So any mention of a 5.51 is actually 5.54.
Anyway in the end is the 5.54 even possible? And the 5.51...
Also is the spreadsheet still publicly available?
it makes me think in tekken when ur on top in a tackle its left right punchs i would do it so fast it would do 2 lefts or rights sumwere nere the end all the time
I want to play dragster
based omni
my best time is 5.94
As far as I can see, this qualifies as forensic evidence that Rogers' claim is fraudulent.
There is a very easy way to obtain testimonial evidence, as well. Hook Rogers up to a polygraph, and ask him a few pointed questions.
That would put a swift end to this bullshit "debate".
Alas, polygraphs-not unlike video game results-can be manipulated. Hard to trust Rogers unless he can show definitive proof of his alleged record.
Given the ridiculous nature of the frauds he's pulled, I don't think Rogers is smart enough to fool a polygraph
AlmightyDreadlock I think the bigger issue is that he's convinced himself that it's true. If it is a lie that he's stuck to for decades it's likely become pathologically ingrained in his memory. Either way, no one takes his record seriously these days.
Any lie, given time, becomes truth in one's mind, and that is a confirmed way to be a polygraph.
Don't know where you people get the idea that Rogers has convinced himself that his "5.51" is legitimate. His efforts to convince everyone else are so flaccid, the logical conclusion is that his self belief is non existent. He would therefore fail a polygraph test, miserably
Maths are great.
What about a disabled sound chip? Would that run the game faster?
you know he sed he found a glitch to start in 2nd aye
There's no such glitch. Even if such a glitch existed, it still wouldn't be mathematically possible to hit 5.51 using it.
This is just a theory...
A GAME THEORY!
Repost without music.
Not possible. The music was present in the original recording.
Repost with music only
POSSIBLE it is.. Its just much work.. write the script what you said. mute soundline of original video. record a new line while you watching the video with the written script. ez as dat - (tbh i diddnt noticed before but the music is really annoying AF at some parts)
Schwanzus Longus hes not going to
without the human element, all of this is meaningless.
no one ever thought about that the arcade screen was a bit cracked and the ref thought its 5.51 instead of 5.57? todd just took it as it was cause he knew its unbeatable (he diddnt knew for sure but he was pretty sure about it i guess).. just a guessing
or maybe it was a little bit dirty and the referee misread it idk.
There was a coffee stain on the TV so they just added .1 to the score because they thought it might be posible
you see.. it wasnt rigged it just was at a time people diddnt thought this isnt getting to be a big topic like it is now ^^
The problem with that is
Todd always said that he got 5.51 3 times,2 times at live performance
and 1 times with his referee friend(Twin Galaxies referee)
And Todd is a arrogant person,he always brag about his record of 5.51
and when somebody dare to challenge him=he will belittle them and challenge them to beat his time(5.51)
how can you explain that then?
if he know from the beginning that his true record is 5.57
why would he do something like that?
for more than 30 years???
Everyone uses your evidence and you get so few views. How fascinating?
Have any interest in releasing this spreadsheet?
The spreadsheet from the video has actually been overwritten already, but you should be able to make your own following the same steps. Do note the corrections in the video description, though.
Omnigamer alright thank you.
So I may not be grasping something, but what is the blue area near the bottom and why is it set the way it is ( I found the spreadsheet from someone I think on TAS. assuming it's all right, I guess I am pretty brand new to the game, and well I'm just trying to understand.
The shaded cells were a conditional formatting. Most spreadsheet programs will have some way to do this; in my case, I set it up ahead of time to just highlight cells that had a distance greater than or equal to 97, which indicates the win condition.
Sorry I didn't specify ( assuming docs.google.com/spreadsheets/d/1m1JKUGQdqjRkgqWgY6j6Dp1dXqM7KKEuYwjw7fpnLSM/edit?usp=sharing was your final spreadsheet) The blue highlighted section, what was that? That is what I'm not fully understanding. Thank you.
Your maths are interesting but something needs to be refined i feel. Reason is i know a guy that got 5.51 on three occasions. His name is Todd Rogers, you can Google him
Googling "Todd Rogers" currently reveals plenty of pieces that are none to kind to his alleged achievements...
The Ben Heck show did this as well. Pretty sure he mentioned you as well. The 5.54 was mentioned. And one frame equaled .03 seconds. Could a hardware glitch, namly the atari system used have any bearing on time achieved?
Nope. No authentic Atari system would behave differently with regard to timing.
How do you feel that you alone with your research took down this scammers record on this game and all his others?
Are you calling everyone who witnessed the 5.51 liars?
The other possibility of how a 5.51 is possible is by shifting BETWEEN frames if you time it perfectly. This would stop the speed from freezing between frames and a 5.51 would theoretically be possible.
There's no such thing. The game only reads shift inputs once in any given frame; the state of the shifter for any other timeframe has no effect on calculations.
But what about games like Super Mario Bros. where speedrunners have been continuously finding ways to effectively break the game's theoretical limits (aka rules) in ways they had no way of knowing being possible, through actually fooling the game. They are breaking the game's logic, by fooling it. They had no idea it was possible until it was found, and then they had to reverse engineer how it happened and why, but the difference is that they were able to see what was done that changed things. Unfortunately, we just wouldn't be able to know with the supposed fake score. It could have been a 1 in a billion chance.This is one way his supposedly impossible time could be possible, I guess.
Also, another possible answer is like in Man vs. Snake, where one player had his score removed, because it was found that the motherboard he was playing on in Nibbler had a tiny electrical mechanic issue that made the game run slightly faster. It's not outside the realms of possibility that something similar may have fluked for the supposed 'worst cheater of all time' (wow, what an overblown label for a god damn high score on an Atari game, right?).
I feel like these are questions that should have been asked before demonizing a very skilled video game player. I mean, Atari had come to him and questioned him about his play method, and he seemed to make the designer think "without a shadow of a doubt" that his score was correct.
We did ask those things - many times, in fact - and many times found that there's no such way to impact the timing or get around the game's rules.
Also, there's no evidence that Todd ever actually spoke with Activision regarding that particular score.Other information has come up that completely contradicts his version of events.
"We did ask those things - many times, in fact - and many times found that there's no such way to impact the timing or get around the game's rules."
But that was part of my point. The Super Mario Bros. speedrunners thought there was no way to get a better time, but through years of constantly running tool assisted speedruns, new things slowly and sporadically popped up. Of course, they are going to believe that there are no other ways to improve it, until they got slapped in the face with a new way shown, that they could then reverse engineer and understand how it happened. The hypothetical problem in reverse engineering and figuring out how a 5.51 might have happened is it was allegedly one time, and he doesn't know how, as opposed to now, where it's all recorded and able to be looked over.
"Also, there's no evidence that Todd ever actually spoke with Activision regarding that particular score.Other information has come up that completely contradicts his version of events."
That couldn't be further from the truth. Activision met up with him immediately to question him and his playing style, and went on to form a friendship with him, turning him into a form of spokesman for the company. Here's the developer of Dragster: www.twingalaxies.com/feed_details.php/87/dragster-designer-without-a-shadow-of-a-doubt-about-todd-rogers-record/1
With this simple and easy fact you seem to not understand, it makes me question how many other things you might have failed to understand, or even see, when you claim how much you looked over "many times".
SMB, as a piece of software, is orders of magnitude more complex than Dragster. It's impractical to try and read through the code of SMB as is easily possible with Dragster. That said, the improvements you're referring to are just easier ways for humans to achieve TAS-level optimizations - there haven't been improvements to the TAS in many years.
There is no more room for anything strange to happen in Dragster. Everything has been examined. If you don't want to trust that, you can read through Thomas J's annotated disassembly of Dragster and decide for yourself whether or not anything is missing.
As for your other line, please re-read what I said and what the article said. David never claims to have seen or questioned the alleged score. You are citing Todd, not David, when you mention that his playing style was questioned. By all accounts, Todd had already started a relationship with Activision prior to any reasonable estimate of when he claims to have accomplished the score. The timing has been contradicted by the Activisions newsletters themselves, and Todd's story has been contradicted by both Crane and other alleged 5.51 players.
Before you try to trudge further down the rabbit hole, everything you are trying to bring up has been discussed to death in more than 2700 posts on TG. If you are honestly interested in the truth, I recommend going through that thread to fill the gaps in your understanding.
Sorry, I had somehow molded together the 2 of their separate accounts in my memory, forming them into all being said by Crane. I have been going down the rabbit hole all night, and now my brain is basically mush. I diverged into this subject, starting from Billy Mitchell's current situation, and am still curious in seeing the source of this new Dragster 5.51, even if a hacked version.
I have no idea who the individual was that Billy is referring to with Dragster, but there are numerous "AutoShift" hacks for Dragster available that are messy, to say the least. You should eventually find some links to videos of the hacks within the first 50 pages of the thread, but they are tricky to find on UA-cam due to overloaded terms. Many may also have been removed or are unlisted. The hacked ROMs should be very easy to locate, however.
So this still doesn’t disprove Roger’s claim.
Secret !
No but if a tas can't match the record I can't imagine the record being legitimate.
I mean for crying out loud, you're talking about the community that turned super Mario brothers into a sub 5 minute run.
Yes it does. It proves 5.57 is the fastest possible time
Viewtiful Z some people refuse to accept facts when presented to them
Secret ! Todd has being proven to be a cheater
Doesn't fucking matter anymore. Mr. Fraudivision has been killed. Guiness cucked him just like Omnigamer did. Anyone who defends the self-proclaimed king of video games is just as much a piece of cockroach dick as he is. Put me 1v1 against him and I'll fuck him up in any game.