Dispute

The following is a formal dispute, proposing that the verification of the Vash and Graystripe performances on the leaderboard as of 2-13-2019 should not have occurred.

Introduction

Historically, the verification standards for this leaderboard have not been explicit nor enforced to ensure a reasonable level of certainty concerning its sub 1:03 performances. In both of these cases there  exists a lack of reasonable supporting evidence coupled together with questionable content. If there has ever existed any suspicions from any of the moderators that these performances are not accompanied by reasonable supporting evidence then they should not have been placed upon the leaderboard during the verification process. It is extremely important that all of the acting moderators feel reasonably confident that the runs which they verify, especially sub 1:03 times, will be able to stand the test of time and basic scrutiny.

General Criteria

In recent years it has become increasingly important to ensure that emulator performances are accompanied by basic supporting evidences that lead us all to be reasonably certain of their authenticity. The following list are general criteria which would grant credibility to emulator performances which are absent from these submissions.

1) Include Emulator and version used in submission notes.
2) It is preferred that the entire emulator window is captured, to help prove the emulator being used.
3) Recommended to use input display, frame counter and have “No Movie” showing to help prove legitimacy.
4) Recommended to always include the reset/power on from the previous run in your submission. This helps to show that the run was not started from a save state or a movie file.
5) While not required for every run, it is highly recommended that runs are livestreamed. This lends another amount of credibility to a run if it is done on stream. Certain games require streaming of attempts below a certain time, again this is on a game by game basis.

For sub 1:03 runs I think it is highly vital that people state exactly what emulator and version they used, show the emulator itself so we can clearly see exactly what is being used, that basic information such as input display, frame counter, and “no movie” visible, contain a reset or power on from a previous run as is the rule for Super Mario Bros, and have attempts live streamed so that we can see the work that was put into the performances which also displays a basic skills test.

General Rules for Speedrunning NES Games: https://docs.google.com/document/d/1apP69VDIzUx-iVO_sj5YjTxBAE2ukHPa6kl33AOW9hA

Graystripe 3-17-17 1:02.86 https://www.youtube.com/watch?v=4Sc4_AF7_4g

This submission is not located in the list of obsoleted submissions on the leaderboard, either because it was simply deleted or never submitted. This performance can be seen on the users Youtube channel. This performance was done on the Donkey Kong Classics rom. The video starts off with the game showing Mario get the frame perfect ladder followed by visible confirmation that favorable barrel RNG was achieved then the game is reset for some reason. During this time we can see the mouse cursor moving during game play. The game appears to restart for us only to see the exact same movements, frame by frame, which we had just observed before the reset. Same barrel movement is observed. It is suspicious that this performance contains a highly improbably duplication of outcomes and the extreme unlikelihood that a person would reset upon getting a favorable setup. At best this might be a prerecorded game and what we are seeing is perhaps a recording of a movie file playing. This is not allowed since there is no evidence to indicate whether or not this was a performance done by the player in real time or was a movie file created by a tas editor. In this case a player can simply set the tas editor to record mode to copy inputs and boards can be completed over and over to get both a favorable outcome while still giving the appearance that it is not a traditional TAS created frame by frame. There isn't any evidence that demonstrates that this is not a TAS movie.

On such a short run higher standards are required to ensure that viewers of these performances can readily differentiate between a live performance, a replay of a live performance via movie file, or a movie file created by input recording in the tas editor. Sub 1:03 runs need to be clear in its supporting evidence as to the very nature of what it is.

Vash 5-23-17 1:02.630 https://www.youtube.com/watch?v=r2yQqUtq6Zc

Likewise in the Vash performance we see a video which contains only the start up of the rom and the performance itself. You will notice that this video contains several frames of gray at the start. A movie file created by a TAS would only contain exactly what you see here. A reset from a precious attempt in the video submission itself would clear up whether we were watching a performance done in real time or if the TAS was involved. Following the general criteria above, which is both easy and free, would have given more credibility to this performance. There are no past broadcasts of streamed attempts, there are no obsoleted runs in the leaderboard archive. There isn't enough evidence that demonstrates that this is not a TAS movie.

Graystripe 5-26-17 1:02:38 https://www.youtube.com/watch?v=S2TE4C5KLvQ

Three days after the Vash submission we see another submission from Graystripe. There is no reset to help demonstrate that we are not watching a TAS movie. There are no past broadcasts that I could find containing these attempts. This video contains a precarious 10 frame loss on the barrel board by coming to a complete stop. If they were trying to execute the TAS ladder trick on the first ladder and then failed its seems incredibly odd that the game was not reset to try again. A 10 frame loss on the barrel board at this level of game play is not an acceptable loss especially since a frame loss of 3 or 4 frames may be acceptable since a) there is a 3 frame window for getting off the “glitched ladder” and therefore one could lose 2 frames if they go on the 3rd frame, and b) however many frames a person may take if they walked 1 or 2 frames beyond the earliest time to ascend the final ladder. The Spring board is frame perfect, perfect movement to the final ladder and perfect ascension. Likewise, the Rivet board is TAS-like in that I haved TASed it at 32.596 seconds while the Graystripe submission took 32.666 seconds. This would amount to perhaps a loss of 2 frames on the entire board that would require 12 frame perfect input transitions on ladders, not including rivet turnarounds, in order to match the TAS. While it is not impossible that this can be accomplished, this submission comes only 3 days after the Vash submission and is executed to such a high degree more than the previous submission. Despite the TAS-like Spring and Rivets this submission does not contain any evidence that we are not watching a TAS movie.

Chambers' Theory TAS for reference: https://www.youtube.com/watch?v=71ze8LAZX8M

Chambers' Theory TAS compared to Graystripe run synced after stutter step at start of barrel board: https://www.youtube.com/watch?v=2Kpj4_7Vutc

Conclusion
It is my proposal that both of these submissions be removed from the leaderboard and all future sub 1:03 submissions should be verified based upon higher standards that can offer reasonable certainty that we are not watching a TAS movie. Any and all emulator criteria should be required in order for it to be considered, all which I stated in the General Criteria section. The general lack of evidence in these submissions lead to suspicion and their credibility can be too easily questioned. Moderation never should have approved them and should have had clear rules stated and enforced to ensure strong submissions to the leaderboard.