Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Evaluate the integral of ln(1 + tan x) dx from x=0 to x=pi/4
EDIT: Confusing programming with math before, I wrote log instead of ln, not that it matters much, but...
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
It works fine for me on v22.
I've been using the temporary v23 to TAS it because of better Lua implementation. I don't know what version Mukki has been using or if it may be the cause of your problem. It's strange though, I thought latest releases changed only GBA emulation.
I'll try to keep up the good work, I'm currently running it very slowly, though. This semester surely isn't TAS-friendly.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
My arguments were aimed at this submission in particular, not at unusual categories in general. Better stylistic choices may make a "Tournament mode" run up to the standard of the other published movie, I don't know...
If you want my opinion on lots of categories, I am indeed against publishing a "new game+" run or other things like that just for the sake of "it's another mode of the game, has other things unlocked, blah blah blah" when the difference from the normal run or entertainment changes aren't big enough. Most of them are brought up every time someone attempts to pass a bad movie of a game that has already been TASed.
Anyway, I'm actually surprised that I was the only one to raise such a point here. If my memory doesn't fail me, when the first "playaround" run of UMK3 was submitted, although most feedback was positive, some people questioned stylistic choices and were against it, others even thought there was no need to publish a run of every MK game and suggested other games of the series.
Years later, we're publishing different categories of an MK game. It seems entertaining the audience is easier nowadays.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Nice things you did there, pirate_sephiroth. I only realized that mashing and moving down before the character was rolled altered the outcome after I did the first one, so I didn't bother to go back. Try the ones from the second player now.
(Seriously, you have no life if you do 1k rerecords just to show it can be done)
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
@Marcokarty
Good, now I think this has got to the relevant point.
I've played this game a lot. Like I said before, I've seen other TASes of it. IMO, this old run is the best tool assisted demonstration of this game. I really enjoy how he manipulates his opponents to do the moves to show the glitches and manages to keep an incredibly fast and destroying pace through the whole movie. I'd vote Yes if a Randper Kombat run was submitted here. With the glitches found by SDR, that movie can be a lot better!
Don't get me wrong, though. It's not my intention to troll this thread screaming that this run be rejected. More than 40 yes votes clearly indicate that I'm the minority here and I accept it.
I also think an ideal TAS should combine speed and entertainment. It's amazing when both goals converge. I just think that extremely optimized movies that are boring need to get more respect here, and that highly entertaining movies aren't free from "entertainment optimizations" like most people seem to think.
I apologize if I was too serious in this matter. I admit I have taken into account controversies that happened in other submissions, and took too harsh the misquoting and short replies to my arguments, which is normally used to say something is obviously wrong. No offense taken, sorry for misunderstanding you.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Marcokarty wrote:
p4wn3r wrote:
Perhaps it's because I'm a too speed-oriented TASer, but I don't enjoy playarounds because they have no well-defined objective to aim.
Objective = Max entertainment possible. No secrets here. ¬¬
Maybe I didn't make my point very clear. You're right when you say that it aims for max entertainment. However, what is highly entertaining for some people may not be for others. So it's not something you can easily decide on what's better or not, and not a precise goal. Time, on the other hand is much easier to decide. And, please, I'd like you to quote my full sentence because what's after it is important for understanding what I meant. In short, I said my entertainment rigor for playarounds/concept demo is a lot higher than for a fastest completion movie.
Marcokarty wrote:
p4wn3r wrote:
I take it you wanted to demonstrate a disputed fight
Yes, because the TAS could to finish so quickly, And it would have less fun. But It doesn't was SDR's primary goal. Just look at SDR's submission text: "I tried to show a good movie with fun moves/glitches and combos in that match".
Marcokarty wrote:
p4wn3r wrote:
The published run shows a lot more of the game, this one pales in comparison to it.
You do not have to compare this TAS to published run. Cause he submitted it for a Concept Demo. Like these movies: http://tasvideos.org/Movies-Hacks.html
So it's another category.
I'm aware of this. The problem is that the published run does EXACTLY the same thing as this submission, it also demonstrates cool combos and glitches and does it much better.
A disputed fight I'd like to see is: land a destroying combo in one character, let him recover and let him counter with another strong combo. It displays tool-assistance much better than having them miss projectiles, punch the air and guard hits.
I never voted No because I didn't want it to obsolete the current run. I did so because I don't see the need to publish another movie when there's a run with a much better entertainment value. It is in the Judge guidelines to keep categories to a minimum.
Marcokarty wrote:
p4wn3r wrote:
I'm a speed-oriented TASer
So you need to open your mind. Speed without entertainment = Nothing
Yet another incomplete quote... Maybe it's not something you think is ideal for TASing, but I do like frame wars and enjoy seeing improvements just for the sake of a record being broken.
Thanks for the advice, anyway. Still, votes are highly subjective and I'm free to vote what I want with any ideologies I may have. It's my policy to comment when I'm against it to make the author aware of what he did wrong in my opinion. However, since people have been so dramatic about popular runs getting No votes, I regret having talked about it...
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
I was the one who voted No, but waited a little to comment because of some controversy on an unrelated submission.
Perhaps it's because I'm a too speed-oriented TASer, but I don't enjoy playarounds because they have no well-defined objective to aim,... unless they completely blow my head, in this case I love them.
This movie, however, failed to do so. I think it's because I've seen other tool-assisted demos of UMK3 before, but I don't really see the entertainment in missing projectiles and defended hits. I take it you wanted to demonstrate a disputed fight, but I didn't like it. Besides, you picked repeated characters, your playing with Sub-Zero's freezing technique got old quickly, the white screen glitch needed to last less time and Sheeva was mildly entertaining for me.
While this would be tolerable in a speed-oriented movie, it's not on a playaround. The published run is so much better and shows a lot more of the game, this one pales in comparison to it. So, I don't think it warrants publication as another playaround. If you found new glitches, I'd rather see a new version of the published run.
Anyway, that's just my position. Judging from the votes and comments, this will likely be accepted. Congratulations, SDR!
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Apparently, I was wrong about a TAS of the second game being faster than the first. There's already a TAS of Pokemon Card GB2 on nicovideo by cstrakm and it's considerably longer: http://www.nicovideo.jp/watch/sm7256179
I never played this game and know nothing of Japanese, but from what I read, you need to have all legendary cards before entering the final stage. They can be gotten by challenging the Grand Masters once they're rescued, but he uses a faster alternate method to get them sooner.
This also allows to get Legendary Zapdos, whose pokepower can get rid of benched pokemon and deal damage to things that can't be OHKO'd. I could see that there's a trainer card able to deal damage to opponents too.
Also, it seems the RNG only rolls when the menu is open, so the randomness is probably different from the first.
The same author also made a TAS of the first game in the (J) version: http://www.nicovideo.jp/watch/sm6861809
Apparently, he didn't know the RNG manipulation techniques used in the published run and went with a different strategy. Instead of changing the deck before most duels, he made one with lots of cards like Bill, Prof. Oak, Recycle and Computer Search, which allow one to access more cards in the first turn.
That got him longer turns and made luck manipulation a lot harder though, he waited more frames and sometimes had to KO opponents with pluspowers or coin flips where it was possible to avoid it. He also couldn't manipulate Murray (it's a pain to get him with two Abras) and Mitch to be killed in two turns.
He also didn't use Defenders to stop the AI from attacking. Overall, despite the tutorial ending ~1:30 sooner in (J) version, his run is 4.2 seconds slower than the published (U) run.
Still, it's an amazing effort and his strategies could maybe be used to save time in a future revision.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Some weeks after the last movie posted, we've arrived at Cycling Road. We now have 29 pokémon registered, but this number is going to raise dramatically from now on, the next parts are where catching starts for real ^^
Cake. (which is a LOT different from pie)
Also, it's worth noting that this WIP arrives at Cycling Road only three minutes slower than the first published TAS of Pokémon RBY, that didn't use glitches.
Thanks for the encode, Critical Five! This time, the movie can finish the game ;)
Unless it crashes again in the next five minutes...
EDIT: An improvement has been found close to the end of this WIP, redoing...
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
This movie gave me the same feeling as the Skyroads TAS. It felt like remembering my childhood. I've been wanting to see a run of this game ever since DOS TASing became possible.
Corner boosting was very nice. Really cool to see that even in simple games, there are interesting tricks to be abused. Nice work Ilari, nostalgia Yes vote.
Also, I uploaded Ilari's encode to Dailymotion:
Link to video
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Thanks, I'm glad that you liked it ^^
Fractal and I discussed briefly about doing the sequel a little after we submitted the run. If the randomness works the same way it does in this one (and likely it does), it's just a matter of finding the new addresses, program a Lua script and it's possible to start right away.
IIRC, the game is a bit longer than this one, and some battles have strange requirements, like having Jigglypuffs or banning some type of energy from your deck. However, it'll probably be faster because there's no tutorial.
Both of us are pretty busy with other projects though... I intend to give it a go, but not too soon. If anyone wants to try in the meantime, I'm here to help ^^
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Oh my God, and I thought Breath of Fire was long.
I played it on 5x speed, the max my computer could handle. It synced fine for me until the end. I used VBA v23 svn217. Try deleting your vba settings file and it may sync.
Given the run's incredible length and the site's tradition to accept well-made runs of long RPG games, we certainly need different considerations to vote on this.
At first, I looked away from this submission because I didn't think 12k rerecords were enough to optimize a 6-hour movie. However, I must admit that, at least on fast-forward, it seemed well done. I liked how you used the last-sanctum trick to save time, and the walk-over-lava glitch near the end. Boss battles looked long, especially Doom Dragon, but I don't know how they could be shortened.
In general, the run resembled a good Let's Play, and the game is pretty well known and a good addition to the site.
I'm inclined to vote Yes, but I'll marathon it in real time in the weekend to give more constructive comments.
Anyway, congratulations for the longest TAS submitted (apart from Desert Bus...). I suggest that you clean the run tag and put the attribute "Forgoes time-saving glitches".
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
reflect wrote:
That run is still glitched. Like I've posted before, this is the most recent true TAS of this game
http://tasvideos.org/567M.html
Why?
According to this, a TAS is a speedrun made with the use of tools. The 30-minute run and the less glitched one satisfy this criteria.
Why can't they be qualified as TASes? What makes a TAS "true" or "false", anyway?
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
I'm aware that there are other versions of this game which would be more worthy entertainment-wise. I did try TASing the SNES version, but gave up.
On these versions, bubbles are a lot faster, there are more items, animals drop on the screen sooner and shots must be a lot more precise.
Doing them would require a much greater commitment to optimize, which I'm unable to offer at the moment. Honestly, I believe that, even in this case, people would have issues with the game's repetition. I'll let someone else do that.
Feel free to vote No if you think it was done in the wrong version.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Bad news for anyone accompanying this. Today, I've encountered a horrible bug which made it impossible to continue with the current route. Apparently, the game crashes whenever you try to take a Lv1 pokemon from a box, includeing our Mew.
http://dehacked.2y.net/microstorage.php/info/487820238/POKEBLUE.vbm
Now, we're gonna have to redo everything from route 25's grass.
(And after I spent the whole day hex editing the movie to manipulate that annoying L15 Drowzee...)
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
On the entertainment side, this TAS is really bad. Level design is not very interesting, almost no enemies are killed (the reason is obvious, it takes a lot of time to do so) and the game is pretty slow.
Technically, the points where luck is manipulated are pretty obvious to the viewer and look like sloppy play (particularly when the character turns back and shots are missed). Sure, this is not the author's fault, and the movie is well done. However, I feel it's because of the game's relative simplicity, when there isn't much to optimize.
So, I vote No, because of game choice. However, it's a great effort from the TASer. I'm sure the author can apply his TASing skills to a much better game.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
That's an amazing improvement, Kriole! I'm amazed by how fast you TAS. You were worried about not finishing it this year and you beat your deadline by three months!
I think the use of glitches would add a lot to the run, but I can't really complain in view of this awesomeness. For me, this is the most entertaining TAS of the site. Yes vote for epic pie!
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
jlun2 wrote:
So the run's going to be warpless/glitchless?
Probably, it's going to be like that. There's no concrete proof that a glitched TAS can be done yet, only a probability of a 15-minute tweak to the end of the game. Since there's no .dsm/video proof that it's possible, nothing can be planned upon it yet and, thus, should be completely ignored.
Still, I'd really like to see the tweaking run, if it's confirmed doable. Pokémon TASes shine most for their glitch abuse. Without it, I think entertainment goes down significantly as it seems that there won't be much interesting luck manipulation as there was on Sapphire.
Anyway, glad that you're working on this game, FractalFusion. I'm amazed at the downhill RNGs took, going from RBY's insane hardware register entropy to a crappy LCG.
Questions:
Have you found any relevant memory addresses yet? I get the feeling a deeper look at the battle RNGs could be useful. For obvious reasons, only the pseudo-randomality behind encounter/IV/ability generation is well known.
Is the seed for encounter generation affected by the battle RNG or the two are completely separated? If the former is true, it should be possible to manipulate encounters away by letting the RNG change by waiting more frames during the last battle faced.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Warp wrote:
I commented that in my opinion if the debug menu can be used for entering cheat codes in the same way as you can do with a cheating device such as GameGenie, it would make the technique invalid and against the rules.
TASing can be viewed in a computational way as trying to find the smallest input file that takes the game to a certain state, normally the ending. The input file consists only of the buttons you pressed, it doesn't include any cheat code you entered on certain frames. Therefore, anything that uses Game Genie to complete the game will desync for everyone on playback. This run, however, doesn't use any third-party device to finish the game, it gets to the credits by entering a possible input from the controller. Your comparison doesn't make sense.
The reason in-game cheat codes are usually forbidden is only one: entertainment, which is what the site stands for. Codes that make the run trivial, like Metroid passwords, are rejected. Passwords that take you to the last levels of a puzzle game can be accepted, if finishing all the levels would be repetitive and uninteresting. No line was ever drawn, what determines if something is acceptable is not a generalized concept, but a search for entertaining movies. If you didn't watch it before giving an opinion, you can't tell anything of its value if published.
About special categories, we've also had slower movies obsolete faster ones that weren't as entertaining. There are also movies that obsoleted others on different platforms. New categories are also heavily influenced by entertainment, before arguing on non-obsoletion, it's necessary to know if the old movie offers a decent amount of entertainment when a faster possibility is known.
Warp wrote:
(And before someone argues that that one pokemon run also jumps straight to the end of the game from almost the beginning, two points: 1) There's a difference: It's not using a cheat code to do that. 2) Personally I find the entire concept of that run a bit dubious, especially since it uses save corruption which is something I don't really like as it feels almost like using a cheating device, but that's just my opinion.)
Knowing Pokémon better, I should say that there are some items that were programmed in the game for debug purposes, like the WTW item and the surfboard, that can be accessed only by glitching. My run doesn't abuse their effect, but it's exactly one of the IDs of these items that allows warping to the end of the game. You should clarify what you mean by "cheat".
Saying that save corruption is "almost" like cheating devices is imprecise and unclear. If you mean that it destroys the normal gameplay and completes the game in an entirely different way, I agree with you. However, if you want to say that it's like Gameshark, I strongly disagree because hard resetting is a completely valid input and the effects of save corruption are a lot more limited than having a Gameshark, corrupting the save doesn't mean you can do "anything" to the game.
Again, Yellow is in its own category, while the glitched CT run obsoleted the older one. Save corruption is also difficult to fit in.
If you think my work of searching the game's memory for a faster warp was just a rebuscated form of cheating, there's nothing I can do except telling my strong opposition to this statement. Fortunately, the opinion of only one person isn't enough to change the rules of the site.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
thundrio wrote:
also quick memory question for anybody who can answer it. do pokemon games typically keep every enemies pokemon's hp in the same address (e.x. im facing pidgey, his hp value is 20 stored in 0690. i kill him, go to a gym fight brock, his onyx's hp value is 50 stored in the same address of 0690.) and in general what addresses would be useful? thanks.
1st question: yes, pokemon games do store the enemy hp value on the same byte, not depending on the species.
2nd question: IVs of the first pokemon in your party, your opponent's stats and HP, and, if you're taking luck manipulation to a high standard, RNG seed.
I think you'd rather wait for the american release to start a serious TAS. Seeing that there's a lot of interest in Pokémon games in general, someone is probably going to crack the RNG or find a glitch soon. Aside from that, (U) releases normally change the RAM map completely and any addresses you find now will become useless in the future.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Chamale wrote:
Strong Yes vote from me. I love the RNG abuse, and I love how the game quickly process attacks and damage - unlike the regular Pokemon games.
One question, though: Can you manipulate coin flips by simply timing the A press? I recall being able to get much better than 50% of the coin flips to go my way, by simply pressing A when the Heads side of the coin was visible.
It's impossible. During battles, the game only calls the RNG when it needs to. I tried what you suggested, the RNG doesn't change at all when I mash. Perhaps the effect you had was because of a non-deterministic emulator?
