Posts for Whelkman

1 2 3
6 7
Posted: 9/18/2010 1:06 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
No way. That's insanity, particularly when it comes to party stat generation. It takes 8792 frames, including 40 optimal button presses, just to see the results of a new trial. Given that you'd probably need to run hundreds of trails to get the desired stats, time would be better spent actually cracking the RNG. Add to that the hundreds of critical hits with particular damage ranges you'd need to manipulate plus manipulating misses against your party... On top of all that you need to manipulate enemy signs for most battles... I'd rather wait until pcsx-rr matures.
Posted: 9/16/2010 1:24 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
I can't get Lua to work, either.
Posted: 9/6/2010 10:29 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
I haven't touched pcsx-rr in a while, but, for me, every release since the August 2009 version crashes upon attempting to save a state via Lua. Literally all I need to do is run: state = savestate.create(); savestate.save(state); And it crashes. I don't even need a game loaded. If I give savestate.create() an argument, savestate.save() will create an empty state file before dying. I got it working once by putting '10' in as an argument, but that only worked once. Besides trying all recent versions, I have also tried the following to no avail:
  • Making new seed save state to start from in case the old one wasn't compatible with the new version
  • Clearing registry settings
  • Formatting memory cards
Posted: 8/22/2010 8:05 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Enjoyment of this run likely depends on one's remembrance of these hacks, which were especially prevalent in "corner stores", though some arcades carried them as well. That includes me, so I vote yes, though I admit I had forgotten just how insane Rainbow and friends were. I think "Accelerator" was the version I played most. Incidentally, Street Fighter II Turbo/Hyper mentioned by JXQ was Capcom's official response to these hacks. It incorporated the increased speed and many of the moves, some of which were seemingly programmatically identical, leading one to wonder whether Capcom stole from the pirates which in turn stole from Capcom. I know TASVideos has a policy against hacks and bootlegs, but Rainbow is better known than some of Capcom's official titles, like the entire EX series.
Posted: 8/2/2010 5:09 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
I once knew this guy, who could touch his eyeball with his tongue! It was totally gross, ya know?
You mean this guy? http://www.youtube.com/watch?v=avGVcZpPHCg
Posted: 7/31/2010 2:35 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Ah, shoulda checked NewMovies.html. That brings up another point: besides monitoring this page and visually picking out or searching for "Updated", is there a way to discover that a recently published torrent has been replaced? Old torrents don't seem to consistently expire; though, looking through my archive, the recently updated ones are expiring. It'd still be nice to get notifications of updates to recently published movies (7 days? 30?) through the RSS feed. Also, unless I'm missing something, there's nothing to identify the "updated" status from either the RSS snippet or the movie page itself. Maybe something like "{UPDATED} [95] Genesis Quackshot (any,r0) by nifboy in 25:00.05 (Published by Bisqwit)" is in order. Otherwise an update could easily be passed off as a duplicate entry. EDIT: I can't speak for other feed readers, but I don't mind receiving notifications of every update, especially now that I know they're updates and not bugs. It's not like TASVideos is high volume. I still think there there should be something to distinguish the "updated" status, however.
Posted: 7/31/2010 2:14 AM
Post subject: What's up with the RSS feed?
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Over the past week (very) old entries have been appearing sporadically. Most recent examples:
  • [95] Genesis Quackshot (any,r0) by nifboy in 25:00.05 (Published by Bisqwit)
  • [86] NES Blaster Master (USA) by Devindotcom in 43:57.95 (Published by Bisqwit)
  • [84] NES Rockin' Kats (USA) by Teowind in 23:38.65 (Published by Bisqwit)
  • [81] NES Super Pitfall (USA) by Arc in 11:52.22 (Published by Bisqwit)
It's not always stuff this old. It seems totally random, but all of it has been published by Bisquit.
Posted: 7/20/2010 8:57 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
In before happylee counters with a 12 frame improvement. EDIT: Flag glitch in 8-3 was cool.
Posted: 7/16/2010 6:26 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Hope you fellas brought your glowsticks: http://www.youtube.com/watch?v=nRGWdqLUlME
Posted: 7/16/2010 6:21 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Man, Rocksteady is moronic in this game, even more than usual even. The jumping sound effects were like a timewarp back to 1979 Atari 2600 games. I love how sound stops whenever Projectile Kitty fires his weapon. Would have voted yes were it Pumaman!
Posted: 6/12/2010 6:14 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Played the DOS version to death. Seeing this destroyed was a treat. The SNES ending gyps you; the DOS version would show all 12 civilizations destroyed. Playing on Emperor is nigh academic given the rate you destroyed the opposition. It's not often I cry "unfair!" during a TAS, but the way this game was brutally destroyed I actually felt sorry for the CPU, all 3,58 MHz of it.
Posted: 5/9/2010 5:33 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Chalk me up as another who thought he was the only one who ever played this game. I agree with the sentiment that seeing this game destroyed isn't nearly as satisfying as it could be as the playthrough doesn't give nearly the "Nintendo hard" impression that this game really is. The player needs to manage a complex web of killing points while battling odd (and changing) physics. Still, I enjoyed watching it for the nostalgia value. Another thing not captured is the the intro tune, which is one of the best Adlib tunes ever composed: http://www.youtube.com/watch?v=b7yCVHni7uI For some reason there's no video. After a while the game goes into demo mode, which is why you hear jumping.
Posted: 2/13/2010 5:41 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Sir VG wrote:
It's weird that the plugin allows me to set the controls, but when the emulator is run, no input is recognized.
I wasn't aware this was a global issue. Dr. Hell's WinMM works for me if anyone's looking for data points. I use a PlayStation 1 controller connected through a Smart Joy USB adapter.
Posted: 2/11/2010 3:20 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
That Shotaro, he's everywhere.
Posted: 2/10/2010 6:55 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Open the file in a text editor.
Posted: 2/9/2010 2:47 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Having the DLL in the same folder as the lua script should be sufficient. Beyond that I'm out of ideas. Guess you're waiting for BadPotato.
Posted: 2/9/2010 2:38 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
The script requires luasocket. Did you download it? http://luaforge.net/frs/?group_id=23&release_id=837
Posted: 2/9/2010 2:29 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
You need to run it via File -> Lua -> Run Script.
Posted: 2/4/2010 1:15 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Replying to PM by DoctorJohn, who asked for Lua assistance, here in case anyone else is interested: Here are some scripts from my bug reporting posts about pcsx-rr: http://tasvideos.org/forum/viewtopic.php?p=221795#221795 Additionally, here's the proof-of-concept script previously mentioned. You'll need libfft.lua linked in the above thread. As you'll see, the routine is quite rough and doesn't offer much more than a way to crudely inflict damage. The final version of this would overload the function call with '<' and '>' so that it could be used for evasion as well. http://dl.dropbox.com/u/2275704/TAS/Final%20Fantasy%20Tactics/libfft.lua If you're planning on engaging Lua in pcsx-rr be forewarned that any pcsx.pause() statements used cannot be removed due to them affecting program flow. Personally, I'm waiting for the next version before I head back into this with any zeal.
Posted: 2/3/2010 12:59 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
I replied, but it's still sitting in my outbox for whatever reason. Here's the text:
I don't actually know the details of which values generate successes. What my proof-of-concept script does is monitor 0x00962D4 which, I believe, is the damage to be displayed on screen; this address is compared to a provided value. The script loops back to a save state, increments a frame, attacks, and compares again. It does this until the damage equals at least the number provided. Given FFT's crit system of awarding bonus damage from 0 to XA-1, resulting damage may be greater than intended, which has happened in testing.
Posted: 2/1/2010 4:14 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
I tested it a bit, and I don't think there's much to figure out. Input has limited effect on the RNG in the short term. You can influence it by dismissing dialogs then bringing them back, but that's unlikely to save time over just waiting. Since the RNG has only 8 bit precision, your average wait time is 128 frames assuming only one value in the bunch generates a success. However, some events may require certain consecutive values where more detailed manipulation may be warranted. Thus far, I have not had much difficulty in manipulating constant critical hits, even at maximum damage values. I haven't kept track of average delays, but it's probably less than two seconds a pop. Note: I have not yet tested how moving the cursor around the battlefield affects the RNG. It would make for an interesting bot if it does.
Posted: 1/29/2010 12:20 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Rats. I thought maybe someone had some some more work on it.
Posted: 1/29/2010 12:13 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Which page is that? I am curious as well.
Posted: 1/24/2010 1:09 AM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
List of "battle blockables": * Navigation message ("Are you really sure you want to do this?") * Exp & JP earned * Ability name * Randomly triggered battle chants associated with spells and abilities Most of these are pretty annoying, and most people probably shut them off. The navigation message is especially horrendous and I'm always surprised when watching videos with it turned on. Exp & JP earned seems innocuous until you realize how much it delays the next action. Besides, seeing exp gained isn't too useful because the math is trivially simple and the stat is listed along with all the other basics by the character portrait. JP earned is even more useless due to JP sharing and high targets clouding perceived progress. It's really no different than not knowing the exp totals in any other TAS featuring level gaining. The characters still jump up and down when gaining levels, and the tone indicates the type of level gained. Ability names would be skipped anyway due to dismissing dialogs on the first possible frame, and the battle chants are only mildly amusing, mostly for the typos.
Posted: 1/23/2010 6:30 PM
Quote
Whelkman
Experienced Forum User
Joined: 7/7/2007
Posts: 161
Like I said, it's highly unlikely I'm going to wind up with anything productive. The best I'll probably end up doing is creating some scripts potentially useful to others. If anyone's going to produce something targeted toward a general audience, it won't be me.
1 2 3
6 7