Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
I don't know where everyone is getting that the Coin Case glitch doesn't work on Silver.
The Coin Case glitch works on Silver, just as it does on Gold. In fact one of my test runs uses it on Pokemon Silver.
I've been busy lately, so I haven't started the final version of the run yet. There's a lot of luck-manipulation and planning to go around.
I heard though that the Coin Case glitch doesn't work on J versions. Can anyone confirm this?
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
thatguy wrote:
At the risk of repeating myself, I just don't understand this logic. Why should the TAS have been moved to Moons? Did it suddenly become more entertaining?
I agree that the routes are so different that they warrant different categories. However, the category of the slower run is unvaultable, and its (lack of) entertainment value makes it unmoonable (not my judgment, the community's). The only solution is for your run to be an obsolete any% run, which then allows it to take its place in the Vault.
Regarding this run's relationship to the fuller run, I've considered this as an obsoletion for it. However, that other run seems to actually play through the whole game instead of skipping most of it via various glitches. Therefore I do not think this should obsolete it. We'll need to come up with some name for that run to reflect its full state, and I'm currently accepting this run as the standard any% run.
Therefore, the glitched Pokemon Gold TAS is judged not to obsolete the old one.
Since the old one was originally in Vault, and the new one now exists, the old one must be moved to Moons to preserve Vault rules.
We have required some publications to be in Moons regardless of rating (including publications previously in Vault moving to Moons) because they cannot be placed in Vault without breaking the rules. With regard to these publications, lack of entertainment value does not make a publication "unmoonable".
Regarding retroactive obsoletion, I have thought about the issue for a while but I think that the use of retroactive obsoletion may open up a can of worms.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
CoolKirby wrote:
Also, we still have the Pokémon Blue run with a branch name that a couple of people had issue with. Is "no HoF warp" (meaning Hall of Fame) okay, or would that need to be spelled out too, which would make it too long of a branch name?
I changed it to "no glitch warps" for now.
jlun2 wrote:
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
YoungJ1997lol wrote:
So lemme guess, no one cares about gridiron football anymore?
Nope. Not even the teams themselves.
On a more serious note, there's still 10 days to go.
boct1584 wrote:
So you think it'll go octuple-overtime? That would be incredible.
Considering that only six NFL games have lasted more than a quarter-game (one period) of overtime and none have reached a half-game (two periods), we're going beyond reality here.
Regardless, I'd like a long overtime game.
Note: From the four major North American sports leagues, not a single game has ever occurred where both teams played two full games or more worth of overtime. Not even one.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
This sounds like a standard balls-and-bins problem in a probability course.
Monotonic means the sequence satisfies at least one of:
(1) ai≤ai+1 for all i,
(2) ai≥ai+1 for all i.
Let's take a weakly increasing sequence (ai≤ai+1 for all i) of length N on the labels 1,2,3,4,5,6. Let ni be the number of elements with the label i. Then ni≥0 for all i, and the sum of all ni is N.
(We assume from now on that all sequences are of length N and are on the labels 1,...,6.)
This is equivalent to placing N balls in 6 bins, which in turn is equivalent to arranging a sequence of N stars and 5 bars. There are C(N+5,5) ways of doing this.
So there are C(N+5,5) weakly increasing sequences. Likewise there are C(N+5,5) weakly decreasing sequences (ai≥ai+1 for all i), and a sequence is both weakly increasing and weakly decreasing if and only if it is the same number repeated N times, of which there are 6 of them.
Therefore the number of monotone sequences is 2*C(N+5,5)-6.
Now for a weakly increasing sequence to cover all the labels 1,...,6 at least once (where we call the sequence a covering sequence), that means ni>0 for all i, and the sum of all ni is N.
This is equivalent to placing N balls in 6 bins with at least one ball in each bin, which in turn is equivalent to arranging a sequence of N stars and 5 bars where the bars cannot be at the ends or beside each other. Considering the problem as inserting the bars in-between the stars, there are C(N-1,5) ways of doing this.
So there are C(N-1,5) weakly increasing covering sequences. Likewise there are C(N-1,5) weakly decreasing covering sequences, and a sequence can't be weakly increasing, weakly decreasing, and covering all at the same time.
Therefore the number of monotone covering sequences is 2*C(N-1,5).
The probability that a sequence is a covering sequence, given that it is monotone, is [2*C(N-1,5)] / [2*C(N+5,5)-6].
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
jlun2 wrote:
I'm just wondering why using the default NES emulator's RAM state is considered ok, despite also being arbitrary. :P
Because that's the way it is.
A particular RAM initialization (say, 00000000FFFFFFFF repeating) must be designated as the "official emulator power-on state" or else we would have endless pedantic arguments about what is valid and how to compare TASes (not that it stops us anyway). The result is that whatever the emulator declares the RAM should be at power-on is declared unquestionably "pure", and all deviations are "cheating". It is the emulator's edict, so just accept it.
(The fact that RTC can be set in VBA, and firmware settings can be set in DeSmuME, and not be considered cheating, are irrelevant.)
Previous threads with discussion on this issue: http://tasvideos.org/forum/t/14564 , http://tasvideos.org/forum/t/14531True wrote:
FractalFusion wrote:
True wrote:
I'm not so sure If want to verify this... :(
True, have you given any more consideration towards trying to verify this (either directly from power-on, which probably won't work unless you fudge the RAM using a custom cart beforehand as in my previous post, or using the reset method detailed in my submission text, which theoretically works on any NES power-on state)?
With the RAM addresses, yes, I might do the cart method, but I need a cart to use. I am unfamiliar with how mappers work at power on and don't care to do anything with it which is why I would like an NROM board to do these with.
What do you mean by the "reset method?" If it involves playback of the movies, I really don't have the time for that... though testing if those work would also be a good thing to do...
Thanks for your response True. Don't worry about it too much. This submission isn't going to be accepted anyway unless rules are changed.
Since this submission has pretty much run its course, I will cancel it. Anyone who wishes to look at console verification is still welcome.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
CoolKirby wrote:
Along with the problems with the run itself, TASes on snes9x v1.43 aren't accepted here anymore unless they've been granted a continuance (which this run hasn't).
Strangely the "Emulator" field in the submission says "Emulator: snes9x-1.51-rerecording-v18-win32", even though the submission text says "snes9x-1.43-rerecording-v18-win32". Don't ask me what that means.
Also, Snes9x 1.43 files are unsubmittable (as in, you can't even submit them).
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
Ilari wrote:
Hmm... If one did the following:
- Take the movie file, and remove savestate and reset from it.
- Load the modified movie, pausing immediately (frame 0).
- Run a lua script to poke the memory locations listed.
- Continue execution.
Would it sync then?
True, have you given any more consideration towards trying to verify this (either directly from power-on, which probably won't work unless you fudge the RAM using a custom cart beforehand as in my previous post, or using the reset method detailed in my submission text, which theoretically works on any NES power-on state)?
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
You can try using
Lua. Check first that you can record Lua drawings in AVI.
You will need to record twice since the screen is not big enough for Lua drawings in a separate space. You can draw a black background over the game screen if you need a black background.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
Radiant wrote:
That's a good point actually. Could someone please explain why there is a rule that "a board game, sports game, or game show game" is not eligible for the vault?
I think the reasoning a long time ago for putting that rule is because of the fear of TASVideos being flooded with these types of submissions.
I'm all for rewording the rule to allow for more games, but I'm not sure how to do that right now.
True wrote:
Maybe I should finally make that boot memory cleaner I was talking about? Anyone have a spare game genie I can rip apart?
How long does the most recent RAM state persist in memory after the NES is shut off? Is it feasible to use a custom NES cart to write what you want in memory, and then turn it off, quickly switch to the cart of the game whose TAS you want to verify, and verify it?
I don't know if this has been answered before.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
I have that problem as well. Both "Mirror archive.org (MKV Modern HQ)" and "Mirror archive.org (MP4 Compatibility)" encodes start with the audio in the middle of the movie (after the logo, it starts with a Yoshi sound and is partway through the BGM at the very beginning, when there should be no sound).
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
I think that the effects of Trainer-Fly (and Old Man glitch/Missingno.) are rather minimal to be referred to as memory corruption.
Also, "warpless" by itself is ambiguous. Something like "no warping to Hall of Fame" would be better.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
Forum username has been changed.
Your Wiki profile may need updating; someone else is responsible for that. Tell us if anything else is broken from the username change.
Edit: For reference, previous username was rasengan0.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
I watched the TAS, and it just doesn't seem optimized (the pause trick seems to kill a lot of momentum and is used in places where simple evasion would have probably been better). I also think that pause tricks and pulling Guts blocks out of nowhere shouldn't be abused in a "low glitch" version.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
henke37 wrote:
The verification movie seems to doing too much work. It doesn't just unlock all the questions, it relocks them as desired.
Some questions require the locking of other questions to occur. For example, question 818 (TOOTSIE) in the bonus round requires locking all questions 800-817 (and a few others). Without that, question 818 cannot be called up. The routines that the game uses are at http://tasvideos.org/GameResources/NES/WheelOfFortune.html .
Anyway, this TAS was intended as a demonstration of how fast Wheel of Fortune can be completed from a reset (allowing for whatever you want beforehand as a set-up; I don't know if this tactic has been used by speedrunners in other games, since they are obviously so used to pressing the reset button). I cannot prove that this arrangement of RAM can occur immediately upon power-on on a real NES, but I don't really care.
I recall that adelikat has programmed a way in BizHawk to start from any RAM and register state upon power-on. If anyone likes, I can try to convert the TAS to BizHawk format using this feature to make it seem more authentic, though personally, I don't see what the point is (see the last sentence of the previous paragraph).
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
feos wrote:
Looking at the recent edit in the pub description I want to ask FractalFusion: did you really listen to that solo? :D
I never thought much about that ending theme.
It is better than the noise in the level, but that's not saying much.
Actually, the whole thing makes me wonder if the sound is really just an emulator glitch.
Editor, Experienced Forum User, Published Author, Skilled player
(1941)
Joined: 6/15/2005
Posts: 3247
saomoore94 wrote:
Warp wrote:
AKheon wrote:
Based on a quick test, a single flawless victory screen wastes 107 frames. If you could take damage in under 107 frames, you would save time.
I wonder if the time wasted getting hit could be avoided by having both characters hit each other simultaneously, with the computer opponent losing on that hit. (It might not save time at all, or it might even lose time, but just an idea to try.)
I don't even know what your talking about.
Warp is wondering:
1) if both player and computer can be hit at the same time,
2) if the computer can be KOed at the same time,
3) if it saves time compared to the player getting hit only.
