Determine what the ending of the game is (and be prepared to present an argument of why that's a valid definition of "end"), and then end the input when no further input can make the ending be reached faster.
I'd prefer if you didn't state this as a matter of fact, as this is just opinion. Some people prefer to end input as soon as no further input is required to reach the ending, others think that it should be ended when no further input can prevent the ending.
If you end the input when no further input is required to reach the ending, you end up with ridiculous TASes where the game ends minutes after the input has ended, which calls into question how long the TAS actually is.
If you end the input when no further input can prevent the ending to be reached, then you would need to exclude certain buttons or their combinations from consideration on a game-by-game basis (because eg. pressing the pause button in many games would stop the ending from being reached; or more prominently the reset button. In some cases some key combination that effectively resets, etc.)
The form I advocate minimizes the such ambiguities, exceptions and cornercases. (I'm not claiming that there exists no game which would require an ambiguous interpretation, but I posit that there are significantly less of them than with those other two approaches.)
Would it be even possible? The game would run natively on the system, and the input would have to be somehow fed to it...
I suppose that depending on how the DS's OS is designed, this could be achieved with some device driver or custom OS (but it would, of course, require for the OS to be hacked. Depending on how well the DS is designed this can range all the way from trivial to almost impossible.)
Determine what the ending of the game is (and be prepared to present an argument of why that's a valid definition of "end"), and then end the input when no further input can make the ending be reached faster.
Typically in many games this happens when the final boss is delivered the final hit, for instance, even if the ending screen then requires keypresses to proceed. You simply argue that the ending is when the boss dies (after which the game proceeds on its own to the first ending screen).
Sometimes it can be subjective and up to opinion.
To be more specific, TAS optimization is reminiscent of the knapsack problem, and given the similarity I'd be inclined to say that TAS optimization in the general case is definitely NP-complete.
Be careful with your terminology. An NP problem is, essentially, a yes/no question that's very hard to answer, but the given answer is (relatively) easy to check.
This form of problem for a TAS would typically be "can this game be completed in at most N frames?" The answer may be very hard to find (perhaps requiring brute-forcing all possible combinations), but a given answer is easy to check (just run it, ie. you can check it in linear time).
However "what's the minimum amount of frames that this game can be completed in?" is a different category of problem (because it's not a yes/no question). If the problem is in NP, then it's a so-called NP optimization problem.
The difference is that checking that the answer is correct is not easy.
Is it better to accept them all, or have them obsolete each other due to similar content?
I'd go with 1 game per platform for each category. That way, any% might beat one version, and hardest difficulty would beat another version of the game due to the result looking more interesting or spending less time on an equally interesting/challenging run. Somewhat like polity for hacks.
IMO I don't see any problem in accepting several eg. chess TASes for the same platform if the submitter can argue why it's different enough than the existing one(s).
For example, one TAS could be of battle chess, just because it looks cooler. Another one could be of a notoriously strong chess game for the same platform, just because it would be interesting to see how many moves it takes to beat it.
I don't think the current tier system would exclude adding a fourth tier where things like board games and more esoteric notable TASes could be published.
Anyway, since the rationale behind the current Vault/Moon tier system is to help find the most liked TASes, I was curious to compare the lowest-rated Moon runs with the highest-rated Vault runs.
It turns out that a non-insignificant amount of Moon publications have a relatively low rating, the lowest combined ratings being 4.1, 4.2, 4.9, and 5.0. In contrast, five Vault TASes have a combined rating of 7+ (7.3, 7.2, 7.1, 7.0, and 7.0.) A quite large amount have a combined rating of 6+.
Does the current Vault/Moon tier division really reflect popularity that well?
1) Entertainment value will always be subjective and up to opinion. We can't do anything about this. I still feel such a division is necessary; otherwise we would be putting Super Metroid in the same category as You Have To Burn The Rope. The tier system exists to separate them so users can more easily find the higher quality runs. But naturally, there will always be borderline cases.
Doesn't the rating system take care of that?
Additionally, for me to consider it, I'd like to see a single organized post that details exactly what your idea is. (If there is such a post and I've missed it, please put a link to it.)
I suppose these two are like that:
http://tasvideos.org/forum/viewtopic.php?p=394886#394886http://tasvideos.org/forum/viewtopic.php?p=394961#394961
In short, the new "vault" would contain all any% and 100% TASes, not just the "boring" ones, the new "moons" (suggested "coins" in the thread) would contain all other game completions that are not any% nor 100%, stars would be like currently, and additionally there would be a fourth "demo" tier for anything else noteworthy or that doesn't fit into the other categories (this might include eg. board games with no clear game completion).
It's just a suggestion, and open to improvements.
I suppose this is very similar to what I have suggested in this thread, the difference being that "Vault" would contain all any% and 100% runs, not just the boring ones. (Also another significant difference is that the suggested tier system would remove subjectivity, since it's entirely goal-based. With the exception of stars; that's ok IMO.)
It was suggested, that Vault game-based criterion is lowered, Moons entertainment criterion is raised, and a middle tier is added, a to-be-defined tier.
I'm not sure I agree with that proposal (at least how I understand it from your description).
IMO the current tier system has its problems, but they are related to subjectiveness: A submission of any game is published if it gets a reception that's "positive enough", else it's published in the dumpsterVault, but only if it's an any% or 100% completion.
The major problems I see with that are: 1) the division is highly subjective and up to opinion, 2) it completely leaves out many games and even entire genres if they just aren't "entertaining" enough, and 3) it unfairly gives the impression of Vault publications as somehow "inferior".
I don't see how the change you mention would correct any of those problems, except perhaps that more games would be accepted (which admittedly would be a positive change).
The proposal of a complete rehaul of the tiers was discussed in length in that other thread some time ago, so there's no need to go through that again. (Although I still don't know exactly why it wasn't accepted.)
If that proposed rehaul is not feasible for whatever reason, then why not simply do as I have suggested, ie. just create a fourth tier for those submissions that do not fit anywhere else? I think it might be the easiest solution. (Although I would prefer the rehaul. Many people object to the bad reputation that Vault has, and I agree with them.)
Add http://tasvideos.org/forum/viewtopic.php?p=405806 to the list (with its rejection reason "As Jenga is a board game, this game also does not qualify for the Vault.")
I still think we desperately need a category for these games. What's in essence "(entire game genre) does not qualify for publication" feels completely contrary to the spirit of the site, IMO. If it does not qualify for vault, then create a category where it does.
It seems to me that the only pet that's shown on a somewhat realistically regular basis is Angel, possibly because Fluttershy's home is always full of animals, so he fits there nicely.
Perhaps the second most regular pet that gets any kind of screen time is Opal, and even she appears quite rarely.
All the other pets, however... only if the episode is somehow about them. Otherwise they are mysteriously out of sight and aren't even mentioned, like they didn't even exist.
Even joke submissions have a chance to be published, if they comply with the rules and requirements. And a run like this once again raises the question of what should be considered the true length of a TAS.
(This is why I advocate the principle of "the input should end when no further input can make the ending of the game be reached faster". IMO it's the most unambiguous and clearest way to measure the length of the TAS. Of course with some games it may become an issue what exactly is considered its "ending", but that's another discussion.)
In my opinion, hardware and OS/execution environment (if any) should (as a rule) be included in the definition of a "platform": even when a game is cross-platform and built from the same source code, its behavior can change. The simplest example is rand(): it returns a 16-bit number in Windows, 32- or 64-bit in Linux (depending on if a 32- or 64-bit distro), uses different PRGN sequence, and so on. But there is a lot of other little things that the OS/execution environment can do, such as emulate in software a functionality that is done by hardware in systems that support it, or add some lag because it processes events from other processes differently. Older platforms, such as NES/SNES/SMS/Genesis, are generally exempt from this rule because they didn't have an OS running alongside the game; so the hardware is the platform.
This has one problem: There is no one single "MS Windows" OS. There are several. Even if we use extremely inclusive criteria to classify different versions of Windows, and even if we drop out outliers from consideration (ie. NT and versions of Windows prior to 3), we are still left with two rather radically different versions of Windows by your own criteria: Windows 3.x, and Windows 95 & its successors. (And even then you could draw a line between the Windows 95 family and the Windows XP and its successors, because the latter was a merge of Windows NT and the mainline consumer version of the OS, and was quite radically different from the Windows 95 line.)
One of the major differences between Windows 3.x and Windows 95 is that the former used cooperative multitasking while the latter used pre-emptive multitasking (or at least attempted to; Win95 was notoriously bad at being a competent OS). This ties closely to what you are describing above. In fact, what you are describing does not describe Windows 3.x very well (because of its cooperative multitasking design, which means that any program could hijack the entire computer for itself and the OS could do nothing about it.)
Moreover, Windows 3.x ran on top of DOS. It never bypassed DOS or replaced its system functions. It simply ran on top of it. You could in some ways call it a GUI program for DOS. (Windows 95 was the first Windows to try to replace DOS rather than running on top of it, but still did a poor job at it. Windows XP is probably the first version of Windows that replaced DOS completely, relegating it to a compatibility mode box run inside the OS on demand. This is probably because XP was basically the next version of NT, which in itself was an almost completely independent OS, developed separately from DOS or the Win3 or Win95 lines.)
The line between DOS and Win3 is really blurry in this regard. Are they different platforms, or are they simply a bunch of libraries running on top of the same platform (ie. DOS running on a PC)?
If your argument is that they work so differently that they have to be considered different platforms, then you can use the same argument to argue for Win95 (and its successors) having to be considered a separate platform, and perhaps even WinXP (and its successors).
It all becomes much simpler if we just look at the hardware all of this is running on. The OS could be some kind of platform subcategory, if needed.
I think this idea has been pitched before, but how about a short segment showing how a TAS is made?
It doesn't have to be a long explanation; something like 5 minutes, just briefly explaining the basics and showing the making of a very short segment (show how you make a short segment of a game with frame advance and rerecords, explaining both features, and when you are somewhat happy, show the result in real time.)
Viewers could get bored if all this takes more than 5 minutes, but I think a decent presentation could be made in that time.
Btw, is the Megaman TAS out of question?
We may be sick of seeing it over and over, but how many viewers of SGDQ have seen it? I think it would be impressive, especially when run on the real console.
Yes, I prefer using highest gamma level because there's too much sunlight in my room and I can barely see things on the monitor + there are no curtains in my room.
I think you should turn the gamma to normal when capturing the video. It makes it look horrible.
