Forensic Analysts Accuse Billy Mitchell of Cheating for Donkey Kong Record (vice.com) 54
A new forensic analysis of controversial Donkey Kong world records claims those records were scored on an emulator and not on original hardware, essentially accusing the record holder of cheating. From a report: The controversy revolves around Billy Mitchell, a well-known player who holds several records on classic arcade games such as Donkey Kong and Pac-Man, and the main character in the documentary King of Kong. For years, some people in the retro arcade game community have accused Mitchell of lying about his Donkey Kong records, prompting Twin Galaxies, an arcade game community that keeps track of high scores (among other things) and the Guinness World Records to strip Mitchell of its recognition, though the organization later reversed its decision.
The new technical analysis focuses on Mitchell's Donkey Kong records of 1,047,200 and 1,050,200 points. The author of the analysis is Tanner Fokkens, a hardware engineer and a competitive Donkey Kong player. His report was backed by five other experts. The crux of the controversy and accusations against Mitchell is that he claimed to have scored those records on original Donkey Kong arcade hardware, while his critics accused him of using MAME, an emulator that is recognized as a legitimate way to play the game, but records scored on these two different platforms are recognized as two different categories of records. "MAME scores which are passed off as coming from original arcade are disqualified," Fokkens wrote in his report.
The new technical analysis focuses on Mitchell's Donkey Kong records of 1,047,200 and 1,050,200 points. The author of the analysis is Tanner Fokkens, a hardware engineer and a competitive Donkey Kong player. His report was backed by five other experts. The crux of the controversy and accusations against Mitchell is that he claimed to have scored those records on original Donkey Kong arcade hardware, while his critics accused him of using MAME, an emulator that is recognized as a legitimate way to play the game, but records scored on these two different platforms are recognized as two different categories of records. "MAME scores which are passed off as coming from original arcade are disqualified," Fokkens wrote in his report.
Oh god (Score:3, Insightful)
Not this shit again, Donkey Kong sux anyway (and yes I played it back when it was still new at the arcade)
Re: (Score:2)
This makes me want to replace real DK boards with raspis running MAME in the wild.
If the emulator behaves like the real thing, who cares? If it doesn't, why not?
Re: (Score:3)
A lot of the arcade hardware had.. Really odd things going on that are very hard to emulate.
I don't know about DK specifically, but I do know a good number of games are not actually playable in Mame because they contain things that need to be specifically emulated, and the original hardware isn't available to actually test what happens in specific cases.
Probably the most (in)famous example of that is Space Invaders speeding up when you're clearing the field. That's 100% hardware related, and they never 'fix
Re: (Score:2)
What gave the game away here is that Donkey Kong draws into video RAM as the display hardware is generating the video signal. Because of that the displayed image momentarily contains half drawn platforms at the start of each level. They look different on a real system with real video hardware, compared to an emulator.
MAME, the emulator he used to cheat, doesn't simulate the video hardware with that level of accuracy. It just generates an image from the contents of video RAM at the end of each frame, rather
Re: (Score:2)
MAME, the emulator he used to cheat
I don't think he's been accused of cheating in the game yet. He's been accused of cheating by lying about the platform.
Re: Oh god (Score:2)
Not quite: it is about the timing of drawing the graphics buffers as they're still being written.
The analysis compared the real first level, the MAME first level, and a photo of a recording of a latter level of one of the record runs. Since the photo looked more like the MAME first level, and they never considered the idea that timings could vary with time, they concluded that the run must have been done via MAME.
That's not to say that the evidence for the other side is any better...
Re: (Score:2)
Probably the most (in)famous example of that is Space Invaders speeding up when you're clearing the field. That's 100% hardware related
No it isn't, it's 100% software.
Re: (Score:2)
It was the render time reducing when less was drawn. If you want to call that "100% software", then sure.
Re: (Score:3)
It was the render time reducing when less was drawn. If you want to call that "100% software", then sure.
Nope.
It's programmed to move one invader per frame because that's all the CPU could do. There's no "render time", No attempt was ever made to render all 55 invaders per frame. The amount of work done per frame is always the same (ie. move the player, move one invader, move the player's bullet, move the invader's bombs, etc.).
When all 55 invaders are on screen at 60Hz then it takes nearly a second for each one to move. When there's only one invader left it moves 60 times per second.
It could easily have b
Re: (Score:2)
In order to 'behave like the real thing', you'd need to reduce latency for input, video, and audio far lower than what can typically be achieved on that kind of hardware. Audio is especially hard to get low latency on, as the buffers want a certain number of samples before they can play.
If you manage to get latency down to 4-8ms, then you'd have something that could possibly compete.
Something like a Mister could probably work already as a replacement.
Re: (Score:2)
In order to 'behave like the real thing', you'd need to reduce latency for input, video, and audio far lower than what can typically be achieved on that kind of hardware.
So what I'm reading into this is that playing on an emulator is actually harder, because of variable input lag...
Re: Oh god (Score:2)
Re: (Score:2)
There's a neat trick for dealing with input lag. When the controller state changes, you restore the emulator state to some number of frames in the past, running forward an unrestrained speed until you get back to 'now'. (Ideally, you'll be able to do all of this in the time you have remaining in the current frame.) Think of it like sending you inputs back in time.
The only real problem with this is that you sometimes get jarring visual "jumps" when the new 'now' is sufficiently different from the old 'now
Re: (Score:2)
I mean ya maybe you have to do this on a emulator running on a old display while also using wireless controllers or a really piss poor system (aka pi 100% before v3 arguably pi3b)
my personal emulation box is a old iCade iPad stand which I boxed in and is using a fairly stout for a netbook celeron and intel graphics that was deshelled and mounted in the box... even with the USB 1.0 lag (I used the cheapest wired usb controller on the planet to solder the arcade buttons to) the LVDS connection from the outpu
Re: (Score:2)
I just want to say thank you for saying
Something like a Mister could probably work already as a replacement.
instead of saying
"insert FPGA platform here" cause it fukin replicates the hardware perfectly down to the analog level
the amount of times I have to explain "emulates" means "imitates to the level or improves on" vs "replicates" means "to make an exact copy down to the finest detail" is maddening
but yea I know, FPGA's can replicate the C64's 1080p digital format HDMI port like a boss... just like 1982
Re: (Score:2)
Would you not be able to catch the stitching with automated forensics tools that consider artifacting?
Could you not implement state on the real hardware, by adding more hardware?
Re: (Score:2)
You can check them out at TAS Videos [tasvideos.org].
With an accurate enough emulator you can render the replays perfectly.
Re:Oh god (Score:5, Insightful)
"Could you not implement state on the real hardware, by adding more hardware?"
That's kind of the point. On an emulator you can restore the entire machine state to _exactly_ what it was before you made a mistake, and then resume from there. From the game's point of view there was never an interruption at all, right down to the game's internal clock showing a continuous uninterrupted game.
"Would you not be able to catch the stitching with automated forensics tools that consider artifacting?"
Catching Artifacting? What source are you thinking they are analysing here? It's not some mp4 that they suspect mitchell stiched together from video taken on his new iphone pro and then submitted. I expect it was originally recorded on VHS 20+ years ago, likely copied from tape to tape to tape, and then likely digitized from that copy of a copy of a copy at some point, and that's what they've got to work with now. (I admit im speculating, and would love to know for sure, but i expect im in the right ballpark.)
And furthermore, although MAME does allow things like pause/rewind/save/restore of state, or playing with the clock timing to simply slow it down a bit, as far as i know nobody accuses mitchell of actually having done any of THAT in this case. As far as I know, he legit played the game start-to-finish in one pass without any cheats.
So the contention here is not really that he used MAME to actually cheat... but merely that he used MAME and then claimed it was the original hardware. Which given that records for the two are kept separate would still be dishonest, even if he didn't use MAME to actually cheat.
Hell, it's at least theoretically possible (in general at least) that if the unit wasn't his, and he wasn't looking for it, that he might not even have known it wasn't real hardware. I have no idea if that's plausible here. But even if that happened, and there was no intentional deceipt at all it should still be stricken from the record if it could be proven it wasn't the original hardware.
Re: (Score:2)
that he might not even have known it wasn't real hardware. I have no idea if that's plausible here. But even if that happened, and there was no intentional deceipt
That would really suck for all people involved.
Re: (Score:2)
With emulation, it's possible to produce a perfect 'stitched' run.
Could you not implement state on the real hardware, by adding more hardware?
Not for everything you'd need to be able to save and restore state for. Think: the internal state of custom chips.
save states / playback an move file and fake play (Score:2)
save states / playback an move file and fake play can be done with an emulator
Re:Oh god (Score:4, Insightful)
It's definitely news for nerds. Check out the technical analysis linked to from the article:
https://perfectpacman.com/2022... [perfectpacman.com]
It answers all your questions:
* the emulator doesn't behave like the real thing because it's written thru a framebuffer and the MAME developers don't care about 100% analog accuracy.
* emulator scores are ranked separately from original scores because emulators provide more opportunities for cheating and more forensics evidence to combat cheating (and just in general humanity has a fetish for "genuine")
* if the real thing behaves like the emulator, it's a sign the contestant cheated.
* your raspberry PI would be caught by the officiant before the attempt.
You may not care about Donkey Kong, and I get it... I don't care about DK either. Or Chess. Or Tour de France. Or the latest crypto-scam. But fraud/cheating and how it gets exposed are awesome fascinating topics ("true crime" for those of us who'd rather not see blood and guts). And this is a particularly good one because the details hinge on subtleties about how our computers and monitors work.
Comment removed (Score:4, Insightful)
Re: Oh god (Score:1)
Re: Oh god (Score:1)
Maybe society should impose a higher penalty. (Score:3)
So many cheaters among mankind. For the curious, here's a study of wasp cheaters and the price other wasps make them pay: "They have less time to feed and to take care of their offspring." https://news.arizona.edu/story... [arizona.edu]
Re: (Score:1)
Or maybe none of this matters in the first place. The "cheating" is stupid, just move his score over to the "appropriate" category. The people who care this much about it in the first place are also stupid. They've spent over 100+ times the manhours arguing and analyzing than Billy even played for the run, all that for what amounts to a 3000 point difference in his score - assuming these so-called forensic analysts are even right. Either way, this is all pointless bullshit.
Wasn't there like a movie about this plotline (Score:2)
Yeah, a movie that starred some fading celebs that involved somebody who cheated and had to then save the world? Donkey Kong was there too.
Re: (Score:2)
Re: (Score:2)
One of the better Otto Preminger movies but that's before my time.
Re: (Score:1)
King of Kong, and it's really great
Re: Wasn't there like a movie about this plotline (Score:2)
Pixels?
Re: (Score:2)
Yeah, that's it. Honestly, I couldn't remember it because it was a forgettable movie. after re-reading the IMDB plotline I do recall the Peter Dinkledge character cheated at something but it was Pacman, not Donkey Kong.
Oh well, a forgettable flick, the DVD of it would make a great table leveler.
Re: (Score:2)
Re: (Score:2)
Naw, but that was a cool episode unlike anything from the final season where you could tell they phoned it in.
Re: (Score:2)
I mean couln't we leverage that claim against every single record on Twin Galaxies or Speedrun.com?
Hobbies are just that, hobbies.
Re:Is that really the best use of their time? (Score:5, Insightful)
Maybe it was a hobby? Some people spend months playing games online and don't really accomplish anything.
Mitchell makes a lot of money off of his "Fame". He charges for signatures, guest appearances, advertising, etc.. Mitchel has a lot to lose, which is why these drag on and why there are lawsuits involved.
And how would you like to be the guy who dropped down from first place due to a cheater, who gets none of those monetary benefits?
Re: (Score:2)
Re: (Score:2)
Such as creating an open forum where people can post their ideas on how to solve the world's problems? Brilliant! Let's get right on it after we're done bitching on an open forum how some people use their time differently than others.
Re: (Score:2)
what about JFK! mario moves back and to right (Score:2)
what about JFK! mario moves back and to right
No! (Score:5, Funny)
You're telling me that a guy who looks like this https://en.wikipedia.org/wiki/... [wikipedia.org] cheated for a high score?
Re: (Score:1)
trying to give a sh!t - (Score:1)
It's bad enough that we may have cheating at the highest levels of chess. Bad enough supposing anyone cares about chess any more. But Dongkey Kong? Unforgivable!
(For the slow crowd: This is a sarcastic comment. It compares the ancient respected game of chess with a frivolous modern children's game. It suggests that the latter is somehow more important and so cheating there is more offensive.)
Not even the most vapid entertainment news. (Score:2)
Re: (Score:2)
Good point.
I'll just post Billy's Evidence Package pdf (Score:3)
and let everyone enjoy all 156 pages of it (at least the first image):
https://drive.google.com/file/... [google.com]
Link from https://games.slashdot.org/sto... [slashdot.org]