Not even a week after the invalidation of a 35-year-old gaming record, another controversy enveloped a big-name game and top-flight performer: Billy Mitchell, the first player to score one million in Donkey Kong, home of arguably the most prestigious and longest-running competitive scene in video gaming for high scores.
Update (April 15, 2018): On April 12, Twin Galaxies returned a decision against Mitchell, finding that the three scores in question were not played on original printed circuit boards (PCB) of Donkey Kong and thus were not valid. Twin Galaxies vacated all of Mitchell’s scores and banished him from further participation in their leaderboards. This post has been slightly revised to account for the new developments.
Mitchell has had three million-point high scores thrown out by Donkey Kong’s largest online community. Another is under question at Twin Galaxies, the high score clearinghouse that, at the end of January, wiped out one of the most notable records in games, which was set in console gaming’s earliest days.
In extremely technical and detailed posts by his peers, Mitchell is accused of falsifying the platform on which he set several of his marks, if not also using prohibited methods of achieving them.
But the accusations have also led the gaming community to ask: What is going on? How did we come to this point? What is the sudden interest in high scores, and the challenges to their validity? And who is Billy Mitchell, and why is he important?
Let’s start there.
Mitchell, 52, was the rival to hero Steve Wiebe in the award-winning 2007 documentary The King of Kong. The film introduced thousands to Mitchell, a restaurateur and hot sauce impresario. He comes across as the black hat-antagonist to Wiebe, a schoolteacher from Washington state who sets out to break the all-time record in Donkey Kong but is continually thwarted.
Mitchell’s gaming career began as a schoolboy, becoming the first to reach Donkey Kong’s fabled “kill screen” (where the game’s programming exceeds the limits of its hardware) in 1982. He was later named one of Twin Galaxies’ video game players of the year when that organization was in its earliest days as an Iowa arcade.
Though Mitchell remained a major member of the arcade scene, his notoriety surged after the 2007 film. His behavior toward Wiebe, combined with an appearance that includes long hair, a beard and garish, often patriotic-themed neckties, has made him not well liked among many enthusiasts.
In the movie, Mitchell avoids playing Wiebe even after saying he would accept a challenge to play at a well known arcade. Mitchell was also accused, in passing, of submitting a doctored video tape to wipe out Wiebe’s high score mere moments after he had reached it.
Mitchell’s score of 1,062,800, set July 31, 2010, was the highest score in Donkey Kong for about two months. But more importantly, Mitchell was also the first person to score more than one million points in the game. While Mitchell’s questioned score is now thirteenth on the leaderboard (pending verification of a new score set Feb. 17), his others remain significant marks. Challenging his high scores would be akin to challenging to Babe Ruth’s 60 home runs in the 1927 season, a milestone still revered by baseball fans, even if it is only the eighth best record today.
But now, an extremely technical case has been made against the veracity of Mitchell’s score. A slowed-down video of his performance compares that to how Donkey Kong renders when it is played in emulation — that is, on something other than original cabinet hardware. The charge is that Mitchell used an emulated version of Donkey Kong and the replay footage to represent a continuous, authentic attempt.
A thread at Donkey Kong Forums points to other technical disparities — mainly in the way the emulator doesn’t natively account for the rotation of the CRT screen — where the screen is turned 90 degrees counterclockwise so that the right side is at the top. Additionally, “While many people have seen Billy play in public, there are no known independent, impartial, objective witnesses to any of the The Big 3 world record games,” said Jeremy Young, the administrator and moderator of Donkey Kong Forum, and the person who threw out Mitchell’s million-point scores from the forum’s leaderboard. “He has never scored over 1,000,000 in a live venue.”
Claiming an arcade cabinet record off an emulated playthrough would by itself be grounds to throw out a score, by whomever keeps a record book. A ROM of Donkey Kong run through an emulator may still use the same code as the original game, but an original motherboard with the code on it literally is that game.
“Determining whether a performance was done via emulation or original hardware is important, for a variety of reasons,” Young said. “Emulation can allow for cheating. Emulation is rarely 100 percent perfect. Emulation can allow for entirely different control schemes, display setups, and other environmental considerations. These considerations are different for every emulated game.
“Despite the bugs that are the crux of this whole issue, Donkey Kong is emulated very well,” Young added. “Well enough, in fact, that the Donkey Kong community largely supports the inclusion of MAME and arcade scores on the same leaderboard. But the potential advantages given by emulation (in the form of cheating) make it absolutely necessary to have clear requirements to determine the authenticity of any performance.”
ROMs in emulation can be manipulated by the software running them — like pausing the game, for example, or preserving an earlier state and replaying from there to undo mistakes or bad luck. It’s not by itself illegitimate play; it’s simply a different style, but it is not popularly viewed as the lead platform.
Twin Galaxies’ leaderboard recognizes scores set on the Donkey Kong cabinet. “I think the general public, rightly or wrongly, don’t give as much weight to emulated scores,” Young said. He also pointed out that high scores in emulation have frequently been higher than on an arcade cabinet, although that’s not the case now.
Donkey Kong Forum keeps a combined list of 435 verified scores, noting the platform on which the marks were achieved. (The site has a new all-time best record that apparently has not yet been submitted to Twin Galaxies for verification.)
“I obviously can’t speak for everyone, but to say that [emulation players’] accomplishments on emulator are somehow less than that of their arcade competitors is ridiculous,” said Young. “But they achieved those scores according to the rules governing emulated scores. Billy did not.”
In its April 12 finding, Twin Galaxies made no determination as to whether Mitchell had used emulation in the videotaped submissions that were originally used as proof of three million-point efforts between 2006 and 2014. “From a Twin Galaxies viewpoint, the only important thing to know is whether or not the score performances are from an unmodified original DK arcade PCB as per the competitive rules,” the organization said in its ruling. “We now believe that they are not from an original unmodified DK arcade PCB, and so our investigation of the tape content ends with that conclusion and assertion.”
Twin Galaxies, which collaborates with Guinness World Records in maintaining what most consider the authoritative all-time high score list, has not yet rendered any judgment. A discussion thread on Twin Galaxies’ forum, now stretching more than 1,400 comments long, narrates the case against Mitchell and the mitigating evidence for him. Twin Galaxies, in July, developed a process to resolve disputed scores, seeing disagreement among its longstanding members about the veracity of some high profile marks.
“Twin Galaxies is 30 years old, and a lot of people have been around for a long time,” said Jace Hall, the games developer and esports promoter who bought Twin Galaxies in 2014. “This has finally been a pressure release valve.”
In a conversation with Polygon, Hall defended the challenge process introduced. The site has a reputation- and tenure-based system that only allows Twin Galaxies members who are consistent, good-faith participants to bring a challenge or vote to express their opinion of the claim. A poll tallying the community’s feelings is not even exposed to users who are not of sufficient standing.
The final judgment on whether a score stands or goes, Hall said, rests with Twin Galaxies; the challenge is not determined by the poll alone. That’s just a way for Twin Galaxies management to understand community sentiment as it considers the evidence. But Hall denied that Twin Galaxies’ users were now settling personal grudges, thanks to the opening of the challenge process.
”I don’t see how you can have a credible system that refuses to error correct,” Hall said. “Nothing is infallible, and it would be hubris to think something is perfect.”
According to Hall, Twin Galaxies doesn’t seem anywhere close to rendering a judgment. “We have to make sure that we look at all of the angles and consider all of the arguments formally presented in the dispute thread,” he said.
To Polygon, Mitchell declined to comment for the record about his case, though he did say he had an explanation for what was going on. It appears that he is waiting for Twin Galaxies to come to a judgment. In a podcast earlier this month, Mitchell suggested that the footage of his gameplay may have been fabricated. Young found that entirely implausible.
Why does the Donkey Kong record matter?
Donkey Kong and Pac-Man are the two biggest names associated with the early 1980s craze of arcade video gaming. They’ve also defined several of moments in which all of gaming today has its roots. The difference between the two games, however, is that Pac-Man has a known maximum score: 3,333,360, which Mitchell incidentally was the first to achieve back in 1999. In records acknowledged by Twin Galaxies, seven others have since hit that score too. But Pac-Man has a fatality in its programming that renders the game effectively inoperable after 255 boards. Every board to that point has a fixed number of dots, ghosts that can be eaten for points and special fruit.
Donkey Kong, though it does have a “kill screen” where the game crashes, is different. On the girder stage, one of four that repeat through the game, a Donkey Kong player faces a constant stream of barrels that may either be smashed with one of two hammers; it can be jumped for a score as well. A scored timer forces the player to complete the level or lose a life, but as an elite player accrues extra lives, this penalty becomes meaningless. Indeed, there’s a tactic called “barrel farming” in which an elite player, on the final girder stage before the kill screen, trades all but one of their lives for the opportunity of scoring points from smashing and jumping barrels.
So Donkey Kong still tantalizes observers because, while it cannot be played infinitely, its maximum high score is still impossible to fathom. That’s why the possibility that the game’s best known high score was faked is one that holds members of the record-keeping community rapt. For now, the jury remains out.
Corrections: An earlier version of this story incorrectly described the nature of the allegation against Mitchell; the wording has been revised. In either case, Mitchell is accused of using Donkey Kong in emulation to represent a high score achieved on the arcade version. Additionally, an earlier version of this story misstated which score was the first million-point game in Donkey Kong history. That mark was set in 2006. This post has been revised to reflect that.