Posts for FractalFusion


Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Hm, I never knew about the Synchronize thing. (I assume you are talking about the Synchronize ability, and its effect outside of battle in Emerald.) I do not know if that is the root cause, but the developers seem to like fiddling with wild Pokemon generation that appears to be beyond just making one ability work. In Emerald, I discovered some time ago that it's even worse, and I'd rather not go into it now. I might eventually figure out the "how does it work" of it, but the "why" part of it I will never know. By the way, the new run will be on Pokemon Ruby. I know you might think it is slower, but the main reason is that we're using Taillow/Swellow for this one, like in cstrakm's run. I'll leave you to fill in the details. We hope to save another minute or so from the test run.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Ryuto, could you explain why this submission was cancelled?
mklip2001 wrote:
I think one question is appropriate here: what was the challenge in making this run? Presumably, if the stylus can select pieces each frame, and the puzzles have a fixed number of pieces, then this run is already optimal.
Optimization is not easy at all. It's actually very difficult. It is not just stylus-stylus-stylus-... . Using the stylus alone only accounts for removing 2 tiles every 4 frames. The biggest challenge to optimizing it is finding long sequences of stylus-A-stylus-A-stylus-A-... With this, you can remove 2 tiles every 2 frames. However, you can only do this as far as there is a chain of selectable tiles, since after using the stylus, direction+A can only select a tile nearby. The optimization challenge is to minimize the number of chains. Personally, I'd prefer a Hard mode run. I don't care that it's 12 stages long; it's no longer than 3 minutes. I find the Easy mode too short and rather trivial (for the fact that it's Easy mode and it's only 4 stages, not because of optimization). I was also disappointed at the lack of stylus-cursor playing during downtime. It would be cool for a TAS to do some tricks.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
I've only worked disassembly of any kind with two games, NES Monopoly and FDS Backgammon. Both times were to understand the RNG and how the game uses it. It wasn't particularly difficult (making 40-minute-long TASes generally takes up a lot more time) but it wasn't easy either. I would only ever consider disassembly for RNG purposes. This is mostly because RNGs are one of the least intuitive things during normal TASing, and are highly procedural in code. Tricks and glitches, on the other hand, are the opposite: highly intuitive during normal TASing, effectively non-procedural in code. Computers don't understand tricks, glitches, or intuition. They just do what they are told to do. Looking at code is much the same way. Even for RNGs, it is sometimes possible to determine behavior just from trial and observation of its memory address, in which case disassembly is not needed. In Pokemon Sapphire, for which I have created a page on RNG mechanics, I only used existing knowledge (fortunately there is a lot), RNG observation, and pattern-finding to obtain the results. They may not be perfect information (only full RNG disassembly can do that) but they are good enough for my purposes. Another example is the infamous cannon launch event in Mega Man X5, which at the time, no one knew when the launch without parts would succeed or fail, or even its exact probability (rumors of about 1/100 were frequent). It seemed like something that would require disassembly, but through testing in a more ideal setting (a couple parts rather than no parts, which raises the probability), I noticed that on a couple instances of success, one byte of the RNG memory address, normally some number from 0-255, was 6 and 7 at the time of decision. I cracked the code a few hours later. As an aside, the reason for the 1/100 rumors (the actual probability is 1/16) is because players on the PSX version would keep resetting for the event, but it always used the same RNG value.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Patryk1023: If you put in some work, you'll be able to produce something good. Please don't feel bad about yourself and don't listen to those people who would rather spend so much time and effort to lower their reputation. Most people here were helpful, so listen to them. It takes some time to have a good end product. It depends on the game; some games can be done sooner than others. But the choice of game is also important. This type of game hasn't been known to show much TAS skill (see #2672: Lucky Bacon's DS WarioWare: Touched! in 54:00.02, #1989: DonamerDragon's GBA WarioWare Twisted in 53:31.55, #2682: MUGG's GBA WarioWare, Inc.: Mega Microgames! in 29:19.45). There are other games that I believe are better for TASes.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
The game just continues as if he caught it. As far as I know, there aren't any other effects. The Ralts is not tied to any Ralts or Gardevoir that Wally would have in later battles. Trainer Pokemon that you battle appear to have set stats and flat IV distribution (all one number). I should check again though.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Something I discovered yesterday. It is possible to force Wally to 2HKO Ralts during the tutorial for catching a Pokemon. This saves a lot of time (about half a minute). To do this, first Zigzagoon must have 11 or 12 Atk (Zigzagoon's stats are set when walking out of the gym). This isn't hard. The hard part is that Ralts must have 17 HP and 6 Def. This means Ralts must have a very low HP IV and a very low Def IV and a Def- nature. If this is satisfied, then Zigzagoon's Tackle can do a max of 10 damage on the first Tackle and a max of 7 on the second Tackle, just enough to KO Ralts. By the way, p4wn3r told me the ROM that allows cstrakm's run to sync, and has offered to encode the run for better quality. The ROM required is "Pokemon - Sapphire (J) (V1.1)", or alternatively, "2126 - Pocket Monsters Sapphire (J) (V1.1)". Edit: Wow. What a stupid mistake. No one saw that. It never happened. <_<
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
I watched the run a couple times. Interestingly, the first time I watched the run, I completely missed the part where it used the continue. I guess it's because I watched Metal Slug X first that I don't find this one as appealing, but it's work, and it's work well done. I see now what the deaths and continue look like, but I don't see at all the reason why so many people are making such a big deal about it as if one wants to put down the run just for a single action alone. This run does not deserve that treatment. On the other hand, I don't like the deaths either, and I feel they detract from the run. The continue isn't worse than a death; in fact, I think that each one of the five deaths during the last two bosses is worse than the continue. But there is so much good stuff in this run that it deserves publishing, and I vote yes.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
andypanther wrote:
Why do the graphics look so glitchy? I don't remember that from the published run.
Sorry for the poor quality. I got this from nicovideo ([1][2]) and uploaded it directly to Youtube. There's nothing I can really do about the quality. cstrakm left a movie file, but I have been unable to find the correct ROM to play it back. For those who aren't aware: I'm building a page that lists how the RNG works in Ruby and Sapphire (and Emerald). I worked out most of this from trial and error, and observation. Things currently covered include formulas, trainer ID, critical hits, starting Pokemon stats, Pickup ability, and catching a Pokemon. I'm trying to find out how the Pokemon encounter mechanic works, but the way it is implemented is needlessly confusing, and is proof of how much Nintendo hates game modification devices. There are also subtle changes in Emerald as compared to Ruby and Sapphire, so I do not know for sure how much of the page applies to Emerald. Also, if you see kirbymuncher, thank him for getting a King's Rock via pickup, and catching Kyogre with a Great Ball with almost no delay. It's in the first published run of Sapphire. Stuff like this doesn't happen every day.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
For those who are interested, here is a run on the J version of Pokemon Sapphire by cstrakm: http://www.youtube.com/watch?v=F2ZgwcYi8M8 http://www.youtube.com/watch?v=m9d4ABPXugc The final time is 1:26, which I am unable to compare to the U version. I have left annotations explaining some stuff.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
5.4 seconds for the first level (Defection), of course. The other levels would have significant though not-as-drastic savings as well.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Personman wrote:
I want to know what he said right after the corner move though!
In the movie, this is all you see, for one frame: The full message (not in the movie) is: ギリギリまで踏み込む! (girigiri made fumikomu!) "I will break in as far as possible!"
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Here's a quick catch mode TAS from some TASer, for those who wonder what it looks like completed: http://www.youtube.com/watch?v=doZxE5C3_wg
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
neo_omegon wrote:
I don’t totally trust TASes on nicovideo which were made by authors who didn’t provide input files. They must be cheating. Even though their videos are highly rated, I regard them just as references. It is a pity that such videos are ranked in.
For the most part, I don't have a reason to suspect that a particular run is cheated or played by cheaters, just because they don't give a movie file. Most games do not have a high-profile for competition anyway. If nico rankings required movie files, there would be very few entries. Mostly because no one cares enough to post movie files. Who else would use them anyway? It is both good and bad, though. Good, because no one cares too much, bad because we can't reference the movie files!
Slowking wrote:
I retimed abeshis MM TAS and it turns out, he timed till the last hit to Majora, not the end of the movie file. His TAS is more like 1:33:12 (maybe even a few seconds longer, there seem to be a few black seconds missing in the beginning and where he switches videos). So you might want to change that.
Noted. I do not have control over that blog, though.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Good job, sparky! I remember spending a lot of effort in making the published run. Good to see that someone else is even more dedicated than I am. The only thing I didn't like was the ending. But otherwise, it is all good and deserves to obsolete the published run. I vote yes.
NrgSpoon wrote:
Comparison videos are up on nico. Registered Nicovideo users: Part 1, Part 2 Public Nicoviewer: Part 1, Part 2
That has to be one of the best videos I've ever seen.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
The latest FBA-RR version is here: http://code.google.com/p/fbarr/
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Try hex-editing the 'X' in 'PXM' to 'J' so it says 'PJM'. It's byte 1 of the movie file. Also, change the '.pxm' extension to '.pjm'. Whether it syncs or not or it loads the savestate properly is another issue. Edit: It's byte 1. Computer people start at 0. Also, there's no \ at the beginning.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
There isn't any problem with the time limitation. There are 2^31 such uids and only 2^16 possible initial seeds, and I believe every one of the initial seeds is possible. My method of manipulation back then was to delay frames until I got a critical hit or such. I would just press frame search over and over; that is a simplistic way to manipulate. However, I think there should be a more advanced way of manipulation now. Just to let you know, I heard that GoddessMaria15 is also working on this run.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Here's some news. I figured out how the game determines its initial random seed in Sapphire. It is set on the 6th frame after reset. It depends on the uid (Unix time) of the movie recording, or your system clock if there is no movie. Note: Due to VBA limitations, the uid (Unix time) must be between 0x00000000 and 0x7FFFFFFF. In other words, it must be between 1970 Jan. 1, 00:00:00 and 2038 Jan. 19, 03:14:07.
Let d be:

- the number of days after Dec 31, 1900, if the year is 1999 or earlier
- the number of days after Dec 31, 1999, if the year is 2000
- the number of days after Dec 31, 2000, if the year is 2001 or later

Let h be the number of hours after midnight that day. Let th be the tens digit and uh be the ones digit.
Let m be the number of minutes after the hour. Let tm be the tens digit and um be the ones digit.

Calculate v = 1440d + 960th + 60uh + 16tm + um. Let x=floor(v/65536) and y=v mod 65536. Convert x,y to hexadecimal.

Then the seed is:

0x0000????

where ????= x xor y.

Ex. 
2011 Feb 6 14:16:18

Number of days after Dec 31, 2000: d=365*10+2+31+6=3689
h=14, th=1, uh=4
m=16, tm=1, um=6
Seconds does not matter.

v=1440*3689 + 960*1 + 60*4 + 16*1 + 1*6 = 5313382
x=floor(5313382/65536) = 81 = 0x51
y=5313382 mod 65536 = 4966 = 0x1366

x xor y = 0x51^0x1366 = 0x1337

Therefore the initial seed is 0x00001337.
RNG information can be found at this page: http://www.smogon.com/ingame/rng/pid_iv_creation#what_is_a_pid Finally, here's an updated Lua script that displays Pokemon IV, natures, and hidden power for the lead Pokemon (starter when you select it). For the enemy, replace 0x3004360 with 0x30045C0. http://pastebin.com/v7DAv0md
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
What's interesting is that in the latest nicovideo "monthly TAS ranking" (linked again below) that Kles posted, the Hikaru no Go video that pirohiko referenced earlier (not the same as this one) ranked #1. Account: http://www.nicovideo.jp/watch/sm13441264 Free: http://nicoplayer.blog60.fc2.com/?id=sm13441264&w=640&h=480 Indication of a cultural divide? Perhaps. In any case, it seems that a person would need to be very familiar with Go (which not many people around here are) to appreciate this movie. I've watched the movie, and the only thing I seem to get is that: - playing in the corner in real life is in general a bad move, and - playing in the corner messes up the AI, which, while an interesting concept, isn't compelling enough, because I don't understand anything else. I was hoping that the game would at least allow moves at superspeed. The thing that seems to stand out the most for me is the cursor playing when selecting map locations, as unfortunate that would be to have that as a highlight for this run. The scrolling Japanese text isn't supposed to be an issue here. No one can read it anyway since it scrolls so fast, and understanding it won't improve anyone's opinion of the run. The fact is that it's Go, and it's not very popular around here. If it was chess, with strange and counterintuitive moves from both player and AI, the reception would be much different. I'll refrain from voting. For now.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
I suppose there is something special about the Tengu shutter, presumably that because the room is long, you regain control. In rooms that are normal small boss rooms, doing that would only result in getting stuck in the bottom-right corner while walking. However, Burner Man's room is long like Tengu's room, and it seems not to work either. I can't explain it. Only other thing is, it can't be done in Wily 3 Stage, except on the final boss, or else Rockman will be stuck. I can't say for sure how it works, but I'm sure parrot14green and sparky have tried every way to do it, so they should know.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
mmarks wrote:
http://dehacked.2y.net/microstorage.php/get/266078324/Rockman%20%26%20Forte%20%28J%29%20%5B%21%5D%20-%20Cold%20SnowMan%20-%20sparky%20%26%20mmarks.smv Cold MAN - SnowMan: finished at frame 10196 and in the previous run finished in the frame 10280, improving 84 frames faster than in the publication of parrot14green
You have 80 bolts. Parrot14green has 225 bolts. Please stop and think.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
Derakon wrote:
What are the requirements to perform the quick kill? It seems strange that it's only used twice.
To perform the quick kill, the boss must be damaged just when he has started his charge-up sequence from 0 to 1 HP. However, it is necessary to have a fired weapon active in that room, and lasting long enough to be able to hit the boss. Remote Mine does the trick. The other issue is to be able to fire the weapon in the room. The glitch at Tengu Man allows control of Rockman to fire the Remote Mine. The other glitch at Wily 1 is due to an exploit that allows both a normal shot and Remote Mine to be on screen at the same time, so that the Remote Mine is still active when Wily 1 is defeated. There are also quick kills in the Mega Man 7 TAS.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
http://saisokugame.blog75.fc2.com/ This Japanese-language blog (which I call Game Saisoku) contains a compilation of video game records, mostly TASes. It has some TASes from this site, and some TAS videos from nicovideo, mostly of Japan-exclusive games that are obscure to English-speaking audiences. For me, it is easy to navigate but for ease I will list the links to the listings given on the blog. Listings are by game and branch (in Japanese, using Japanese names, of course). In each one is a streaming video from Youtube, Dailymotion, Nicovideo, or other. All of them are embedded and Nicovideo ones do not require an account. For some reason, some Nicovideo ones are blank. It depends on how the server feels at that moment, so maybe it is temporary. NES/FDS: http://saisokugame.blog75.fc2.com/blog-entry-616.html SNES: http://saisokugame.blog75.fc2.com/blog-entry-617.html GBx/GBA: http://saisokugame.blog75.fc2.com/blog-entry-618.html DS and N64: http://saisokugame.blog75.fc2.com/blog-entry-620.html Genesis (Megadrive) and PCE: http://saisokugame.blog75.fc2.com/blog-entry-619.html PS/PS2: http://saisokugame.blog75.fc2.com/blog-entry-1071.html Other (PC, Arcade, VirtualBoy, WSC, Gamecube, Wii, everything else): http://saisokugame.blog75.fc2.com/blog-entry-621.html If I have some time, I'll dig up the ones I like and list them here. I'll focus mainly on Japan-exclusive games but I have other interests too (such as Pokemon TASes).
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
sameasusual wrote:
Why was Rush Search summoned in the Wily Machine fight? I saw it dig up a small weapon refill (2 units) Was this for Remote Mine? [which at last visibility, had two (2) bars remaining from the Groundman fight] How many bars are needed to use R. Mine? 3 or 4?
Just before he used Rush Search, he used a Remote Mine, so now he has 0. The point of using Rush Search is to dig up a small weapon refill. When there are 0 bars, Rockman fires a buster shot. Then he grabs the refill (0->2), then fires a Remote Mine. This is the only way a buster shot and Remote Mine can appear on screen at the same time.
Editor, Experienced Forum User, Published Author, Skilled player (1941)
Joined: 6/15/2005
Posts: 3247
This game reminds me of Azure Dreams. Except without all the plot and stuff, and a whole lot of running down stairs that becomes repetitively funny after a while, and actual battles. And a lot shorter.