I have to disagree with your position on this, feos. We should accept new movies as an improvement based on better accuracy alone, this motivates people to improve this aspect of TASing. The idea that "we don't know if this better accuracy is perfect accuracy, so it can't be counted" just seems plain wrong, for the same reasons why I disagree with the rule for vaultable highscore runs having to have 100% proof that they can't be improved. You can never proof something is 100% perfect, such proof should never be required. We don't follow this mindset for normal speed records, so why in these instances? And of course, there's also the problem with TASes that are optimized to SMB1-levels, we should be able to replace those movies if necessary, even if no gameplay improvement is found.
Allow resyncs when a movie is to replace another movie from the same author(s). This sounds quite reasonable in any event; what's to prevent an author from retracting their old runs before submitting the new one to circumvent the resync designation? Unless you can't retract a run from the site.
Allow resyncs when a known improvement is made to a core as opposed to resyncs across cores. This could tie resyncs to improvements in hardware knowledge better.
What we see here is an example of a movie that can't really be called a meaningful tool-assisted record.
http://tasvideos.org/MovieRules.html#MustBeClearlyDefinableAsAGameWhichHasAchievableGoals
http://tasvideos.org/Vault.html#TechnicalRequirements
CASE 5: "The Big Sweep"
Touche would take a new approach. Rather than gather the evidence case by
case and prosecute Le Fiend four times, he decided to take everything on
at once. The whole city, the entire gang, all the cases and, of course,
Le Fiend. He rose early that day and took a deep, invigorating breath.
"Room service," he said into the phone, "please send up breakfast. What
is today's special?" The voice on the other end replied, "French toast."
PROSECUTION REQUIREMENTS: All of the stolen goods and evidence in cases
1-4 must be found and returned for verification
at their place of origin. Le Fiend may then be
rounded up and booked at Police Headquarters
once -- and for all.
STATUTE OF LIMITATIONS: 20 minutes
LOCALE: Entire city, 248 blocksHey. i want to TAS with TAStudio but as soon as I want to open any .tasproj files or save them, TAStudio crash. I have BizHawk 2.3. The error is in french and I will try to translate it. An exeption happened in a composant of your application. If you click on "Continue', your application will ignore this error and try to continue. '-10' is not a valid number for 'LargeChange'. 'Large Change' needs to be superior or *this stop here* And after this there is a huge text which I can translate if you want. Help me please, I want to tas again but everytime I try, nothing works x)
The current unassisted word record for The Legend of Zelda: Ocarina of Time any% category is 17m 07s, made on the Wii VC version of the game. There's an unofficial TAS of the same category that's 15m 43s. This TAS wouldn't be accepted at tasvideos because it's done on a N64 emulator and uses a glitch (the so-called GIM glitch) that cannot be used in an actual N64 console without the game crashing (and the sole reason why it does not crash with the emulator is because of faulty emulation due to a faulty graphics plugin). This TAS, using the current any% route (ie. the same one used in that unofficial TAS), could be done on the Wii VC version of the game using the Dolphin emulator and, as I understand, this would be accepted at tasvideos. My question is: What if I were to create a TAS on the Wii VC version of the game that's faster than the unassisted WR, but slower than that unofficial TAS? Let's say, for example, that the TAS would be 17 minutes flat. Would it be accepted? It would break all existing official records (both unassisted and tool-assisted), but it would have known inefficiencies. What are the rules regarding this?
I wonder that if I made a TAS of the Wii VC version of the game that's not optimal but still beats the real-time record, if it would be accepted. After all, one of the requirements for publication is that it must break all existing records (it would, since no better run has been officially published here or in the real-time community), but it's highly unclear to me that if there are known improvements whether that would constitute grounds for rejection. Assuming nobody else would be willing to make an any% TAS of the Wii VC version (which at this point appears to be the case), we would be in a strange situation where the fastest completion time that anybody has made would be rejected, but nobody would be willing to make a better one.
In rare cases, there will be a sloppier movie that is faster than a more optimized movie due to the use of a major skip discovery. In this case, the faster of the two movies is preferred.