I think one of the things boring about this is that minigames are done so fast, you really dont even get a chance to see (or hear) anything that happens, so it just seems like "oh well that would be easy to do in a speedrun", maybe a encode with the crosshair pointer would be more fun to see.
Aaaand the thrilling excitement that is navigating menus and doing mandatory busy work.
If it uses a glitch, then it can't be published as glitchless. And if it's just published as a regular run, then it's obviously suboptimal. Maybe if the glitch used couldn't also be used elsewhere, but the pause tricks are very easy to pull off and well-known, so using it in just one place is bizarre either way.
Second, you have a very poor case for even copyright infringement under fair use clauses, of which is satisfies nearly all of them.
The Creative Commons Attribution 2.0 license, which is what all TASes are published under on this site, clearly decrees:
"You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use."
Creative commons licenses are based on the holding and enforcement of copyright. It doesn't exist above and beyond copyright. If copyright cannot be enforced or doesn't apply, because fair use/parody/public domain/etc, then by definition, neither can a creative commons license.
First, plagiarism is an ethical violation only. Nobody, Youtube included, is under an onus to do anything about it. You probably mean copyright infringement, which is a legal matter.
Second, you have a very poor case for even copyright infringement under fair use clauses, of which is satisfies nearly all of them. Purpose is different. Different nature. Substantiality is less than 10% of your run. Effect on and competition with your 'market' is nigh nonexistent. The only one it violates is non-commercial, EXCEPT...
Third, 5,000 views is peanuts for YouTube. It's like... MAYBE $20-30, and that's assuming he/you get a fantastic CPM which is exclusively seen in the full-time video creators who get view counts in the millions.
The idea was that if the newest emulator is not required automatically, but only when it's possible to sync the movie on it
Again, who is going to be doing this work? You? It's a nice sentiment, but so is "everyone should have professionally landscaped lawns." What're you going to do when someone submits a run made on Gens or Visual Boy Advance (still acceptable right now, but not preferred)? Say "On top of the work you've already done, you need to do a bunch more work to try to sync it to another emulator even though the run you already have is acceptable." I can tell you now exactly how that'd go.
Are you volunteering to do this for every submission then?
Wouldn't it make more sense to simply edit the site to stop accepting them? I mean, there there's a continuance list, so doing that for other emulators seem to make more sense than to cause endless frustration on viewing and encoding.
Rejecting all movies but those using the single most accurate emulator for a particular system is very different from what was suggested and I'm pretty sure would make a lot of people upset.
Never played the game though. Thinking of tricks like a five hour movie in A to lure enemies away one by one (or put power-ups etc), so B gets really easy.
Nothing as silly as that is possible AFAIK.
I looked a little more into the differences, and it looks like while there are changes, they're nowhere near as numerous as Capcom advertised. The things I could find from a casual search were:
Triggering Brad (skip all weapons to a certain point in A, lets you kill him and loot his corpse for a costume in B)
Cord (Use at X, no zombies at X in A, but zombies at X in B. Two possible places to use it. If you don't use it at all, then zombies at both X and Y in B)
Sidepack/SMG (if taken, not available on B)
The gator (if killed via gas canister, not in B)
The gas/sprinklers (using them in A to weaken certain enemies makes them stronger in B)
Letting Marvin live (more zombies in his place at B, or maybe just him in addition to the regular zombies)
Locked door fingerprint (if skipped on A, can't open in B)
A number of those (maybe all?) are things that shouldn't have any effect at all on a TAS. Brad, obviously, and the fingerprint door just has an SMG/ammo.
Adding to this, some enemy and item placement was changed between the Japanese and international versions to make it harder (the later dual shock release added a "USA Difficulty" implementing that in Japan too), although you run past 99% of them anyway, so it's probably not a major difference. Also, the PC and DC versions add a harder difficulty, so there's also that.
Nicos you’re making no sense. Any submission of the B routes would need a save file to accompany them for sync anyway.
He means that the run would be one single thing, the A route that creates the save data, and then the B route that uses it. Creating saves from a clean file as you go is perfectly allowable (and often unavoidable) and there are certainly runs that use them in various ways. It's kind of like Kirby Super Star. Clearing one game plays an ending sequence and unlocks another. Except that the games don't affect each other while they do here... and RE2 boots you to the title screen to explicitly load the data instead of just merrily continuing on.
That runs into the potential issue that jlun mentioned though, each could be suboptimal for a more optimized paid. ie skipping a machine gun in A makes it take 10 seconds longer, but saves 15 seconds in B, or running past zombies in A saves 10 seconds, but killing them would save 20 seconds in B. Things like that.
In the regular speedrunville, the vast majority of the runs are just Leon A or Claire A. I suspect the logical compromise here is that an A run would be accepted as is, but could be obsoleted in the future by an A-B run if a good enough case can be made for any less optimal parts, and B runs on their own are unacceptable. Plus Hunk and Tofu, I guess.
I don’t think treating them all as a single thing would make sense. If you’re going for a speed record, each individual route will be its individual record, and you would try to optimize them individually. Why would you want to slow yourself down so that you can get some sort of feeling of “continuity”? Who cares? It’s a speed record, go for the fastest setup.
That leads to the situation then where there's:
Leon A (clean save data)
Claire A (clean save data)
Leon B (loaded from a suboptimal Claire A clear save, optimized for Leon B)
Claire B (loaded from a suboptimal Leon A clear save, optimized for Claire B)
4th Survivor (save data must include B route save data with A rank)
Tofu Survivor (save data must include 6 routes completed plus unlocked 4th Survivor)
You can obviously put all of those on a single memory card, but that's pretty far from what "loading from save data" usually means. Also, he hasn't provided his save, so who knows if he even did use that other (suboptimal) run to create this one. Or if it's optimized for this one because he hasn't written much of anything.
Speaking of which, it's a moot point anyway because this is not optimal for Claire B so should be rejected for that. An easy improvement is to shoot the last incendiary at the ground in the elevator Birkin fight to both save a shot and start moving again sooner. I also highly suspect he can be stunlocked sooner. See here from a Claire A TAS:
http://www.nicovideo.jp/watch/sm25138606
RE's entire plotline is a trainwreck, but Leon A/Claire B has a lot more holes and things that don't make sense. Also, Sherry shows up in RE6 and references things that only take place in Claire A/Leon B.
I think you are correct that they later stitched together some parts of Leon A re: Ada though, so maybe consider it a melange, but most likely, Capcom's writers just have no idea what they're doing and/or got confused.
Claire B is an unlockable after completing Leon A, which requires save data, helpfully not provided, so this is already unverifiable at the moment and in violation of the rules.
Yes, that was his master plan all along. Create a very suboptimal TAS for one route in order to set up an optimal TAS for the second. Genius.
Also, I guess I assumed people would know, but probably not, so I should also mention that the B scenarios are directly affected by what you do in the A scenario that creates the save data for them. They're a continuation created from specific save data, not just "have a Leon A clear and you unlock Claire B." Two different Leon A saves will create two different Claire B scenarios.
For example. In Claire A, there's two ways to beat the alligator boss; shoot the hell out of it, or grab a gas canister, toss it in its mouth, and shoot that for an instant kill. If you take the first approach, the alligator is a boss again in Leon B. If you take the second approach, it never appears. Less dramatic, if you collect a machine gun in the A run chronologically before the B run visits that area, it'll no longer be there. If you skip it, it will be.
Enemies, items, etc. All are affected by what's done in the A route that unlocks that B run. This makes the B scenarios completely unsuitable for TASing on their own, and still kind of silly for TASing attached to an A run, although there is a case to be made for a Claire A/Leon B (canon) or Leon A/Claire B run. Absolutely not subdivided like this though.
Where do I begin? Let's start with this. Claire B is an unlockable after completing Leon A, which requires save data, helpfully not provided (yet obviously loaded from at the start), so this is already unverifiable at the moment and in violation of the rules.
Actually, let's just stop there until that's addressed.
Are there any ingame-time runs that are not (at least) one of a) for a game with a speedrunning community that relies on the ingame clock, or b) for a game with very long time-based score tallies?
My inclination would also be to go on a case-by-case basis, but on the other hand my impression was that the rules for the Vault were rather strict and simple (viz. fastest realtime any% or 100% runs only). I wouldn't want someone to spend a lot of effort on a "Vault-only" ingame-time-oriented run only for it to get rejected...not that I have any evidence that someone is considering making such a run.
Kind of repeating what I said in the other thread, but...
In a lot of the currently published movies, sacrificing real time I'd say already falls under the "sacrifices time for entertainment" categories. Otherwise, they'd need to get hit or waste weapons/shots by shooting into the air just to make the end-of-level countdowns go faster. Similar goes for fighting games. Many have a special message and bonus for taking no damage which takes far longer than taking a sliver of damage would, so most fighting games that aim for absolute speed will have to work in getting hit which seems contrary to 'perfect play.' Of course, I'm not a great fan of pure speed records for fighting games anyway, so...
I hadn't really thought about it the way Feos says, but I kind of agree. Is the goal getting to the endstate of each individual stage/fight as quickly as possible, or to get to the next stage/overall be as quickly as possible? Obviously there's tradeoff when it comes to things like powerups and advantages carrying over between stages, but if someone was able to shave two ticks off SMW 8-3 (so the timer ends on 246), they'd end up losing significant time. I believe (though could be wrong) that there's a similar situation with the SMW fadeout lag and that a number of stages could be completed marginally faster, but the added lag would wipe that out.
Either way, pausing eats up both those times so I'm a little leery on that counting.
The generation of random numbers is too important to be left to chance.
But yes, I was oversimplifying. Still, there really aren't any TRULY random number generators. The best we can do is assume something is random (or random enough) and base generation on that.
Besides, if RNGs were truly random, then the same input from the same starting point would produce different results, and duplicating/verifying anything that relies on one would be functionally impossible, so in the context of any kind of duplicatable results, you can't ever be talking about a true RNG.
If you didn't know, drops and layouts in the game are controlled by random seed (which you can manually select, but internally counts as cheating), not an RNG. Aside from meaning a finite number of possible layouts (ballparking at a mere 37^8), that doesn't mean there exists a seed where all those drops can possibly actually happen. What that run does is modify the item and layout files in the game, which is very different from luck manipulation.
I admit to not knowing a lot about the item generation, but it's quite possible that there are normally limitations or pools that certain items are drawn from, so X dropping at the very start may preclude Y from dropping until much later, or X may be tied (intentionally or not) to certain layouts normally. Or the algorithm that the seed is input into guarantees a certain number of short and long levels. Etc etc.
Anyway, the point is that an Isaac run would likely not have anywhere near as much RNG manipulation as there seems since everything critically important is controlled by a single variable which is set at the very start and doesn't change for the entire run. In essence, and to answer the OP, the important variables already are fixed, not truly random. If you know the seed, you know everything.
Is in game time even vaultable? I know of at least one game where focusing in game time would allow an infinitely long real time movie, but 0:00:00 in game completion.
Its smash brothers.
https://www.youtube.com/watch?v=cfX2KVXlAF4
I think the rule's for things like games that count down score, 'flawless victory' announcement, or do things like Super Mario's fireworks that would otherwise force a player into taking unneeded hits, wasting weapons, or wasting time just to avoid them, but that said, a cursory glance through Spikestuff's link, I don't know that I saw anything that paused to manipulate luck. Also I'm absolutely sure things like the Super Smash thing you linked or another example, Super Meat Boy's 0.00 timer glitch, would get flat out rejected as a valid goal. A Punch Out run using a timer glitch to go for in game time was also rejected. It seems like a tag in need of clarification.
looks good and efficient to me, would have preferred a bit more description / notes on the submission for any decisions taken during it / things to keep an eye on. But as it stands it was enjoyable to watch and seems like a great finish time.
Looking at the list of submissions there doesn't seem to be a RE2 run which is actually up since previous ones just got rejected, so this definitely deserves the place, if people want to complain about things they think aren't optimised enough then go and improve them and beat the time like normal. We need to at least have the base movie to improve on rather than just rejecting whatever comes up.
I don't think "remembers to skip cutscenes at the earliest possible time" is too much optimization to ask for in a run that's over 50% loading screns and 40% "walk from point A to point B with no obstacles in the way." It's not an isolated problem either. He's a couple seconds slow to skip the one at 27:30 too (jumped to a random spot in the middle to spot check, 29:50 in the comparison version I posted above).
It needs another pass and to pay a lot more attention to those kinds of simple things.
Edit:
Here's another section that's noticeably slower than the real time run. It doesn't look like luck there was manipulated at all. This is what it should look like: https://www.youtube.com/watch?v=EWAZI-zogN8&#t=46m10s
To be fair, if you remove the loading screen door transitions, that cuts off around 30 minutes, if not slightly more.
...Which is still way too few rerecords, not to mention no excuse for being behind the real time run at barely just over a minute in. I'm not even sure the time spent to set auto aiming was necessary. I know it's generally frowned upon (or at least was a while ago) by regular speedrunners and there are generally very few shots fired in an RE speedrun anyway, so sacrificing a few seconds to set it instead of using the manual aiming seems very questionable unless I'm forgetting something about the game.
Here is an album of just a few places that air sliding is meant to/being used http://imgur.com/a/8InpT
This is convincing evidence. Yes vote.
Oh hey, the gallery's been edited yet again.
================
Instead of denigrating other games/runs, you'd be better off explaining what this TAS has that make it better than the previously rejected Burst Chaser TAS besides using a different game as a base and having air sliding. http://tasvideos.org/4105S.html
Do it using only things found in the TAS. Nothing ineffable. Not how much effort went into the hack. Not who the hack's author was. Not things in the hack that aren't present in the TAS.
It's easy to spot the differences and how they change the whole games from the bottom up with the already published Megaman hacks (and pretty much all the hacks published so far).
http://tasvideos.org/1534M.htmlhttp://tasvideos.org/2583M.html
By trying to mimick the original game and be just like it, though, this ends up... showing the exact same things as the original, just on fastforward, plus an air dash.
