Player Crushes Atari Record Using A Glitch Thought Impossible

A speedrunner using a glitch never performed on consoles has set a new record time in the Atari 2600 game Barnstorming. The glitch creates a new speedrun category for the game, whose scores were once held by the controversial video game player Todd Rogers.

The 1982 game Barnstorming has various mode to play. The goal of Barnstorming's "Hedge Hopper" mode (also known as Game 1) is to see how quickly you can fly a propeller plane through 10 barns. The barns are located close to the ground and the player must swing down under their roofs without crashing into the barn.

The current world record time is a 32s 720ms time by Clay Karczewski. A speedrunner named ersatz_cats has beaten this time thanks to a glitch thought to be limited to tool-assisted gameplay, clocking in at 29.97, outlining the history of the run in a detailed Reddit post.

In the 1980s, the top score was held by someone else: Todd Rogers. Rogers is an infamous high score setter who made a name for himself in the 1980s. Last year, his record in the Atari 2600 game Dragster - considered the longest held video game world record - came under scrutiny after a speedrunner used emulation tools to show that the score was not possible under normal conditions.

All of Rogers' video game scores were later removed from the leaderboard of Twin Galaxies, a top record-keeping group, after an arbitration process concluded his Dragster time was impossible. The Guinness Book of World Records voided Rogers' world record for longest-standing video game record.

In the spring 1983 edition of Activision's newsletter, which was called Activisions, Rogers is credited with a "Hedge Hopper" time of 32.74 seconds. Even today, this score would be considered exceptional, but rumours of a lower time of 32.04 also exist on message boards.

That score was supposedly submitted to Twin Galaxies and logged as the best time but later removed. For some, it has become known as the the "coffee stain" score because of stories that the submitted time had been misread due to a coffee stain on documentation.

"I've only heard about the supposed stain," Rogers told Kotaku via Facebook Messenger. "When I submitted any of my scores, they were good and clear."

The time from the coffee stain run would surpass Karczewski's and seemed impossible. However, it is possible to hit the roof of a barn in Barnstorming in a precise location such that the game counts it as though the player has successfully flown under the barn. The impact knocks the player back before the barn.

With the proper timing, they can crash into the barn over and over to get an incredibly low time. Using this trick, a tool-assisted speedrun achieved a time of 14.74. This was done on an emulator using special tools to go frame by frame through the speedrun. It was unclear if this trick could be executed on console. ersatz_cats' speedrun is the first documented use of this technique on an actual console.

In their run, they hit the barn in that exact spot twice at the start of their run.

"I've been aware of the roof bang glitch since the beginning," Rogers said. "If you bang or nick the very front of the barn with the top of your wing, you'll be credited with passing through a barn. It's a cheesy way of manipulating the way the game is played."

Roger denies ever using the glitch. "When I played for Activision they watched my entire performance, just as in Dragster. There were no flaws in my game play. I performed without glitches. If I played any other way do you think they would have hired me on to showcase their products."

Ersatz_cats' run bests both Karczewski's record and the supposed "coffee stain" score. The glitch completely changes how the game is played and has led to the creation of a "Glitched" category for Hedge Hopper on leaderboards.

"This discovery on Barnstorming should be taken for what it is: A cool new skip on an Atari classic," ersatz_cats said on Reddit. "Not only did we break new ground on a very old game, not only did we beat [Todd Rogers' time,] but we beat the original coffee stain, 26 years later."


Comments

    It's a cheesy way of manipulating the way the game is played.
    A good description of speedrunning generally it seems. Sometimes you break the rules, you get a new category. Sometimes you get wiped from the record books. Depends on the kind of rule you break I suppose?...

    There is a difference. Rogers manually fudged his dragster screenshots to show a time he never achieved, because it cannot be achieved, as has been proven definitively.

    I mean if you want to create a new category of obvious falsehoods in speed running, I can tell you that I have a 0 star run of Mario 64 that took 3 seconds and a twin galaxies mad can confirm this because they, like, legit saw it happen.

      I suppose. I was just pointing out that the 'cheesy manipulation' quote kind of applies just as well to the meta game of speedrunning times as it does to the actual computer games themselves.

      Besides, wasn't the whole point of this article that the glitch was previously considered impossible, but that turned out to be wrong?

      because it cannot be achieved, as has been proven definitively.

      This is kind of proving a negative. One might be extremely sure, but "definitively"? Nah.

      I mean it's similar to when someone says some code is bug-free. Any experienced developer will tell you there ain't no such thing.

      Speedrunning is actually the perfect example of where new ways can be found to trick very old code/hardware. This happens regularly.

        Well, the simpler the code, the easier it is to claim that its bug free. From what I understand of Dragster it's a single input game and many years of software testing and use of perfect tool-assisted runs has shown that a perfect run can only achieve 5.57 seconds with Activision stating a theoretical 5.54 was possible way back when. 5.51 is theoretically impossible based on an overwhelming body of evidence and the image he used as of 5.54s proof is viewed as likely photoshopped.

        All of this is about as definitive as you can get.

          Well, the simpler the code, the easier it is to claim that its bug free.

          Absolutely, but there could always be something. And even if a program is simple it is running on top of some kind of operating system or environment which could have its own flaws, plus there could be hardware issues. With arcade machines there could be ones that have different characteristics.

          Like, purely for arguments sake, imagine that a game is written to work on specific hardware. Maybe the game animation timing is tick based not real time based while the clock timing is based on a real time clock. This all works fine when built to spec. But maybe it turns out that a few of the arcade machines were built with an incorrect bit of hardware which runs faster. Maybe the clock is faulty.

          Bugs are very imaginative. They pop up in all sorts of places. They take all sorts of forms. They can be functions of extremely complex interactions between multiple systems. You can never be 100% sure they aren't there, just that you can;t find any more.

          I mean programmers say "this shouldn't happen" while looking at something happening way too often.

            I mean, sure all of what you say is true. The question here is what is more likely:

            1) He had the only buggy machine that enabled him to post a time that was, and still is, considered by pretty much everyone to be impossible.
            2) He lied about his time, just like he has before any number of times and scores, as an example his 15 million high score on Donkey Kong. Or for example his barnstorming time, where to prove their point, critics stripped the game of obstacles and proved that you can't even fly fast enough in a straight line to beat the time that he claimed.

              Critics proved his barnstorming time was too fast.......it is possible to hit the roof of a barn in Barnstorming in a precise location such that the game counts it as though the player has successfully flown under the barn. The impact knocks the player back before the barn. With the proper timing, they can crash into the barn over and over to get an incredibly low time.
              ...but it turned out they were wrong.

              To be fair though, I agree with you, it's probably more likely that he's just been lying for 40 years or so. But I don't know for sure (occum's razor, confirmation bias blah blah blah).

                ...but it turned out they were wrong.

                To be fair though, I agree with you, it's probably more likely that he's just been lying for 40 years or so. But I don't know for sure (occum's razor, confirmation bias blah blah blah).

                Yet, from the article:

                Roger denies ever using the glitch. "When I played for Activision they watched my entire performance, just as in Dragster. There were no flaws in my game play. I performed without glitches. If I played any other way do you think they would have hired me on to showcase their products."

                So either he is lying about his time or he is lying about not using the glitch. Either way, he is lying.

                  Whether or not HE used the glitch, it is still possible. If it is possible then the critics who said that it's impossible are wrong. If they are wrong about this - something that was 'definitive,' at least before it was proven to be false - then they may be wrong about other things. Do you know EITHER WAY whether he had a faulty cabinet or not? Almost certainly not. If so I have no idea why you're going in to bat against this dude and his times, which weren't really the topic of conversation in the first place.

    edit: oops wrong place...

    Last edited 11/05/18 7:08 pm

    You are using a classic fallacious argument.

    1) It is not possible to achieve x time without glitches > 2) glitches exist in the game that can reduce the time below x > 3)therefore people who claim he cannot do this without glitches are false.

    Historical context matters. Todd Rogers comments on this article are akin to Lance Armstrong(a fellow cheat) complaining that riders these days use lighter bike frames (our glitch equivalent) because back when he won 7 tour de France(which have been stripped, just like all of Todd's records) , he did it on a 6.5kg bike(glitchless). Except they are still both cheaters and liars.

      That's not what I was trying to say. Sorry I haven't been clearer. To use your example -
      1) It is not possible to achieve X time without glitches
      2) glitches exist in the game that can reduce the time below X
      *3) therefore people who previously claimed, or proved, that no-one could achieve a time below X, glitches or otherwise, were wrong.

      I'm not concerned with Todd Rogers specifically at all - and like I've said, I agree with you. I'm of the opinion that he probably cheated - just pointing out that we can't ever really get a definitive conclusion here. Without actual proof that his cabinet was not faulty at the time (which understandably, would be quite impossible to test now), you can't say that he definitely cheated. Conversely he also can't prove that he didn't cheat without actual proof that his cabinet was faulty at the time. Or maybe I just have an unusually high bar for proof, I dunno...

Join the discussion!

Trending Stories Right Now