I used the same trick ziplock used to skip the scolling to 'leve; 20' and setting the speed to 'high'. You can do this by inputting your settings, en then start you're movie. The game restarts then, but the settings are still the same (if you restart the whole game the settings are back to normal). It's being discussed if this is allowed, but I'll tell you why I did it:
After starting a level you can't manipulate luck. The only thing that can be manipulated it the level you enter. This level has it's own order of capsules that fall... so you have to pick a level that has the viruses placed in a good way, and capsules that fit.
You'll never get all the capsules you want... but you can try a lot of levels to look if it's good enough for a speedrun. Some time ago I wanted to start a speedrun of this game... and tried a lot of levels... but eventually all came to a point where the capsules didn't fit at all.
Ziplock had a pretty well chosen level, and so I tried a few levels using the same trick. The levels looked a lot better, and I recorded a few on 25% speed. One level was really good, so I took that one. The level I chose is an other level ziplock played.
The purpose of the game it to dispose of the viruses, by placing capsules. A horizontal or vertical line of the same color of either capsules or viruses, or both, that is 4 blocks or longer will disapear.
After I picked out my level, I played it again, and looked more carefull to find the fastest 'route'. Then I played it frame by frame to get the fastest result possible. The total length is 2 minutes and 2 seconds.
I chose to wait untill all capsules where placed, and pressed start to how the ending. If I would have stopped at the last capsule and let it drop by itself (like ziplock), the movie would been 2 frames shorter than 2 minutes.
Bisqwit: This is a difficult movie to judge.
My vote is a 'meh', slightly tilted to the yes-direction.
But under the current situation of apparent inflation
of yes-votes, I'm rejecting this submission.
Good luck next time!
feos: Baxter forced me to make this encode, sorry :p
This is indeed problematic - the movie is not poweron-anchored, but actually starts on a NES reset which is done after the settings have been set.
It is the same as if you started the movie from a savestate - the reset is there only to circumvent the word of the rules.
I don't think this is in concordance with the spirit of the rules.
This demands discussion. Comments?
As for the actual movie, it still has Baxter fighting against himself, but a lot less than dashziplockdash had. Which is good.
It's not against the actual rules, but that's because the rules assume that a reset will be enough for the game to assume starting position. It seems it's not.
We need either new rules, or a patch. Preferably a patch, because it's not obvious when this happens, neither for the recorder or the judges.
We need either new rules, or a patch. Preferably a patch, because it's not obvious when this happens, neither for the recorder or the judges.
Currently I have no idea how to even detect it. I don't know how does a poweron savestate look like and if I can trust it's the same for all games. And I don't know if anyone is actually able to produce such a save for their movies.
The motive for the rule is that
1. if the movie begins from a savestate, we can't know whether the savestate has been hacked
2. savestate-anchored movies are hard to compare.
Hmm, this explains why I had higher experience than normal in my 2nd Sword Master run...the game must save it as long as the power's on.
Good to know, a patch for this would be nice indeed.
-Josh
but then you take my 75 perchance chance of winning, if we was to go one-on-one, and then add 66 and two-thirds ch...percents...i got a 141 and two-thirds chance of winning at sacrifice
>Currently I have no idea how to even detect it. I don't know how does a poweron savestate look like and if I can trust it's the same for all games. And I don't know if anyone is actually able to produce such a save for their movies.
Until it's fixed, you could patch your own FCEU to force reload of the rom when the reset bit is set. If the results aren't the same as when playing the unpatched version, something is probably fishy.
Now that I think of it, how difficult would a patch be that switches console reset for rom reload when you record/play a reset-based movie? Doesn't sound too hard.
Well the question is whether or not a player could have done what he has done with a hard reset. If the settings for the emulator powering on are equal to the settings using the restart sequence, there isn't a problem. This is a minor abuse of what should be a new rule, but it can be abused to enter passwords and settings before the game is started, and in some games, set the continue to the final level or boss. But can anyone think of a game where this isn't detectable visually? Off hand, I can't. I would vote to simply automatically disqualify any movie that abuses this minor tweak until a patch is made for a hard reset. Just making it illegal will prevent anyone from trying and I don't think it would be a problem.
As for the video, it's way faster than me and it's a shame that I have to vote no ONLY because of this, especially since I accidentally discovered the abuse, but I hope you can hex edit it to keep it (not sure that's possible).
But another question: if a patched version of FCEU was released that fixed this problem, would it be possible to detect then? Could you tell from the fcm whether a person was playing on the most recent version?
>But can anyone think of a game where this isn't detectable visually?
Josh the FunkDOC mentioned that his Sword Master run had experience leftover from his last run, even though he reset the game. This went unnoticed by himself until now. The savestate could probably be hacked to set any quantity of experience, which would go undetected in all but very few cases.
>But another question: if a patched version of FCEU was released that fixed this problem, would it be possible to detect then? Could you tell from the fcm whether a person was playing on the most recent version?
Yeah, you could. If someone records a movie from savestate+reset with a "bad" emulator, that movie would desynch when replayed by a "good" patched emulator, because the starting conditions are different.
Joined: 8/31/2004
Posts: 298
Location: Falun, Sweden
Not if it's started corectly, then it wouldn't make any differance.
Edit: By the way, the movie was nice but I think some more future planing could have been made to execute some series.
Bein' away for like five years, and not a single new post in the ZSNES forum... :'-(
Good point.
Hopefully only if they used something which relied on different starting conditions. I don't think many runs did in the past. If the power-on state is the same as the savestate+reset state, which I can imagine that it usually is, there should be no problem. I can't say for sure though. Testing needs to be done on this.
If it turns out that most or some movies doesn't work, it's always possible to save the emulator version in the header of the movie file. Have the emulator being able to play both old savestate movies and new poweron movies, but only accept the new version this site henceforth.
It will be tricky to figure out if any of the old movies are not legit in that case, unfortunately.
Gotcha - so what you're saying is a movie won't desynch unless the restart save-state affected something in the movie that the power-on sequence doesn't. In that case, I think that's all the more reason for the patch to be a good idea. Might be good to test the movies here to see if they abuse this tactic.
I'm probably making a mountain out of a mole-hill, but it's interesting.
Do the other emulators have this concern?
Virtuanes has a similar but more severe issue with this - it also always loads a savestate, but there is no reset, so you have to trust the user that the savestate is really from the first frame of power-on, and hasn't been doctored. Because of this virtuanes movies were judged unacceptable until this is fixed.
The others I think are truly power-on based. The page keeping updates on this is
http://tasvideos.org/EmulatorRecordingStatus.html
Until it's fixed, you could patch your own FCEU to force reload of the rom when the reset bit is set. If the results aren't the same as when playing the unpatched version, something is probably fishy.
And I won't be accepting a solution that can't be automated (i.e. implemented in the webserver part that analyzes if the file can be even submitted).
Joined: 8/31/2004
Posts: 298
Location: Falun, Sweden
Does that mean that it's no idea to do anything in FCEU until this is fixed? (Sorry if it's a dumb question but I'm having severe problems understanding English sometimes)
Bein' away for like five years, and not a single new post in the ZSNES forum... :'-(
Does that mean that it's no idea to do anything in FCEU until this is fixed?
I don't like to revert rules, so FCEU is currently still accepted, but there exists this grave possibility that a submitted movie could have a doctored savestate that does something nasty to the game even when followed by a reset.
The actual bit that should be changed in FCEU source code is here:
// trigger a reset
if(flags & MOVIE_FLAG_FROM_RESET)
ResetNES(); // NOTE: this will write an FCEUNPCMD_RESET into the movie file
if(!fname)
FCEUI_SelectMovie(CurrentMovie,1); /* Quick hack to display status. */
else
FCEU_DispMessage("Movie recording started.");
This code is from src/movie.c, and instead of ResetNES(), it should call PowerNES().
I recommend doing this change to FCEU as soon as possible. It does not affect the playback of existing movies.
Btw (Quietust, Nach, TNSe), FCEU fills the NES RAM to "FF FF FF 00" instead of "00".
I have voted no for this reason. Be logical guys when you see the settings is the same than before resetting the game, it means that is a soft reset. I encounter this problem in MTPO when I started the timeattack and quickly find the solution to avoid that kind of incident.
Joined: 3/13/2004
Posts: 1118
Location: Kansai, JAPAN
I had to vote no because I didn't enjoy watching it. I don't know if that's a condemnation of puzzle games, but I wasn't entertained, despite the fact that I've played this game a lot.
Just to make clear, I was referring to my second version of the Sword Master movie. The one currently in the submission queue (v3) is clean.
-Josh
but then you take my 75 perchance chance of winning, if we was to go one-on-one, and then add 66 and two-thirds ch...percents...i got a 141 and two-thirds chance of winning at sacrifice
I agree with Bisqwit that this type of exploit is againt the spirit of the rules. I believe that all movies should demonstrate the limits of what could actually be played on the original system and hardware, which means that I'm also against the use of Left+Right and Up+Down, since they couldn't be performed on the original controller without modifying the hardware (i.e. removing the D-pad). We're supposed to be demonstrating what could be done if you eliminate human barriers like reaction time and "skill", not hardware barriers like the inability to press opposite directions at the same time. Perhaps the rules in general are due for a rewrite, with that italicized passage as the overriding "mission statement".
TASing or playing back a DOS game? Make sure your files match the archive at RGB Classic Games.
