Post subject: Pokemon TPP Version (Anniversary Red hack)
Joined: 12/29/2007
Posts: 489
This is pretty must just to get the idea out there: Thanks to TwitchPlaysPokemon just having successfully finished their Anniversary Red run with a complete Pokedex, this can be considered a pretty popular/well-known hack. Unlike the Coin Case glitch Silver run and the Bad Egg glitch Emerald run, the Generation I movies here have no published "glitchless"/"no memory corruption" counterpart. Every one of our current RGBY movies exploits some fairly strong glitch to complete its goal much faster than intended, with the "catch 'em all" run warping everywhere to complete the Pokedex and the "regular"/"no save corruption" run using Brock Through Walls and experience underflow to beat the game even without glitch warps. In fact, none of our currently published Gen I moves defeat any Gym Leaders at all. A "no memory corruption" Gen I run could be an interesting category to bring back (the same way "no glitch warps" was brought back after item underflow was discovered only to have both routes become obsoleted by Brock Through Walls). Such a run could nowadays improve massively on Tilus's 1:51 record in 2005 by using the current RTA strategy of having Nidoking spam OHKO moves on everything, but I'd like to make the case of using TPP's Anniversary Red hack. In terms of the hack itself, it's considerably more difficult than any other Pokemon game published here or even any made by Game Freak, with Victory Road at and above level 80 and the League Champion/Rival topping out with a full team of level 100s. While this would usually lead to a grindslog in normal playthroughs, having a TAS crush the game despite this difficulty could certainly be entertaining. The game also, completely unlike the original Red, has a lengthy (but optional) postgame, starting with Dream Red immediately after the Hall of Fame, followed by rematches of all 8 Gym Leaders, an explicit goal to get the Diploma (which requires the Mew sidequest this time), and a true ending that is only unlocked by doing all three of the previous. (As with a game like Emerald though, it would be perfectly fine for a runthrough to stop at the first Hall of Fame screen or at Dream Red.) Any thoughts?
Skilled player (1741)
Joined: 9/17/2009
Posts: 4981
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
What relatively "weak" glitches are allowed? Even with that out of the way, how entertaining would it be for the battles to take a while since pretty much the only way to get 100's in a speedrun is to use the xp overflow?
Pokota
He/Him
Joined: 2/5/2014
Posts: 779
Call the category "All Badges", for one, since that was something you pointed out in your proposal. As far as no memory corruption and no glitch warps, I propose that a compromise could possibly be reached if Snorlax can be skipped going Vermillion -> Lavender since getting the Silph Scope and Pokeflute comprises half of the midgame fluff. Between luck manips and the inherent system, I don't think you'd need 100s to take down 100s - in fact I think that's part of the original proposal is to not grind if we don't have to.
Adventures in Lua When did I get a vest?
Joined: 12/29/2007
Posts: 489
jlun2 wrote:
What relatively "weak" glitches are allowed? Even with that out of the way, how entertaining would it be for the battles to take a while since pretty much the only way to get 100's in a speedrun is to use the xp overflow?
In terms of "no memory corruption", the Trainer-Fly glitch, the Old Man glitch, and Brock Through Walls all involve reading unexpected values from memory, which would classify as corruption. The Poke Doll skip and the Fight Safari Zone Pokemon trick (the grandfather of the Old Man glitch) arguably do not. Oh - that's another thing I should have pointed out. This hack fixes a few of the glitches present in the original Red, including both of the above non-corruption glitches (to require Twitch to navigate through the Rocket Hideout maze), and I think at least one of the AI bugs where Lance would spam the useless Agility against any Poison-type was also fixed. As mentioned below the point isn't to grind the player's team to level 100. It's to take advantage of tool-assistance to get through tough battles in a way that no real time speedrunner could.
Pokota wrote:
Call the category "All Badges", for one, since that was something you pointed out in your proposal. As far as no memory corruption and no glitch warps, I propose that a compromise could possibly be reached if Snorlax can be skipped going Vermillion -> Lavender since getting the Silph Scope and Pokeflute comprises half of the midgame fluff. Between luck manips and the inherent system, I don't think you'd need 100s to take down 100s - in fact I think that's part of the original proposal is to not grind if we don't have to.
"All badges" isn't the best name - incidentally that was what FractalFusion's "no memory corruption" Gold TAS was originally named when the map distortion Gold TAS was published, until someone pointed out that a real "all badges" run would have used at least the instant victory glitch everywhere. (Since then, the discovery of arbitrary code execution has rendered the name "all badges" even more obsolete.) "No memory corruption" as a category name should be fine I think. The compromise idea is interesting. Not sure if it would end up falling under TASVideos' "no arbitrary goals" policy though, since it amounts to picking one memory corruption skip glitch while avoiding several others. My original intent was to have an "effectively glitchless" run of a Gen I game to contrast with the numerous glitched ones. Your interpretation in the last paragraph is spot on - Gen I's battle system is simply inherently broken, to the point that even Twitch was able to beat 36 consecutive level 100 opponents in a row without healing, something that would never have been possible in a future game.
Former player
Joined: 6/30/2010
Posts: 1107
Location: Zurich, Switzerland
I don't really see the point of using a hack for this type of run. For a Pokémon hack to be published here, I would prefer it to be a completely new game, like most published hacks are. Also, there is a glitchless TAS on Nico, but I don't know how optimized it is. It's not too popular here because it abuses staying on low health to speed up the battles, which produces an obnoxious noise.
Current project: Gex 3 any% Paused: Gex 64 any% There are no N64 emulators. Just SM64 emulators with hacky support for all the other games.
Experienced player (704)
Joined: 2/5/2012
Posts: 1797
Location: Brasil
i think it's a bad idea,because the category removes absolutely all optimizations conceived in the speedrunning community.Or can we accept the arbitrary category of "nido family ohkoing with horn drill x accuracy the whole game or just luck manipulate the ohko because it's tas"? why not just simply 100%/catch'em all GSC or RSE now that we have glitches to hatch the most different eggs possible in some of those games? Your suggestion just seems like those people who want a OoT glitchless intended route speedrun,all tricks removed just run right for great justice all over again.
I want all good TAS inside TASvideos, it's my motto. TAS i'm interested: Megaman series, specially the RPGs! Where is the mmbn1 all chips TAS we deserve? Where is the Command Mission TAS? i'm slowly moving away from TASing fighting games for speed, maybe it's time to start finding some entertainment value in TASing.
Editor, Expert player (2073)
Joined: 6/15/2005
Posts: 3282
I have been briefly thinking about a glitchless 100% TAS of this hack since a few days before the run was finished (not making it personally, just thinking about it). My feeling is that, although it would definitely make a good TAS in my mind, I'm not so sure that it would make a good submission to this site. I am aware that hacks in general have been received poorly on this site and I think it is the second-most likely reason for the TASVideos collective to completely tear down someone's submission (the most likely being alternate goals or "goals"); I myself have been guilty of such behavior in the past. I have become rather cynical in regards to submitting stuff for the site so I just want to warn about this. That being said, I think that hacks should be TASed on their own merit, and not solely to get around some limitation in the original. Since the idea of TASing a hack is to show off the hack, "glitchless 100%" is the first thing that came to mind for this hack. Of course "glitchless" will always be subject to debate; personally I think that the specifics of what constitutes "glitchless" should be left simply to the author's discretion and let the TAS speak for itself; otherwise there is too much red tape. I think a good 100% would be completing all important events exclusive to this hack (gym leader rematches, Dream Red, Mew event, Battle Tent, catch all the Pokemon, and E4/Champion rematches). Since I don't think many people have played this hack (outside of TPP) so far, I'd hold off on making judgments of what a TAS would be like without studying this hack a bit more. The patch for the hack is currently available at Reddit as download; however there are no guarantees that the version is final (note that bugs are still being reported on that thread). By the way, even though this hack seems a bit too similar to the original game, there are ROM hacks with movies that are accepted on this site that are hardly anything more than replacing the main character, such as: [2521] Genesis Knuckles in Sonic the Hedgehog by marzojr & WST in 13:05.29 [2752] Genesis Sonic 3 & Amy Rose by WST & marzojr in 29:27.58 [2514] Genesis Tails in Sonic The Hedgehog by marzojr in 13:21.87
andypanther wrote:
Also, there is a glitchless TAS on Nico, but I don't know how optimized it is. It's not too popular here because it abuses staying on low health to speed up the battles, which produces an obnoxious noise.
It is definitely optimized (not MrWint level, but more than any other Pokemon TAS; over 280k rerecords were used). I think the lack of popularity here has more to do with the fact that it is in Japanese than any other factor.
grassini wrote:
Or can we accept the arbitrary category of "nido family ohkoing with horn drill x accuracy the whole game or just luck manipulate the ohko because it's tas"?
Nobody was advocating for such a thing. Zowayix's comment of Nidoking using OHKO moves does not refer to a category in itself, but merely a means to achieve optimization (which we tend to call "tricks").
grassini wrote:
i think it's a bad idea,because the category removes absolutely all optimizations conceived in the speedrunning community.
grassini wrote:
Your suggestion just seems like those people who want a OoT glitchless intended route speedrun,all tricks removed just run right for great justice all over again.
http://tasvideos.org/Movies-C3056Y.html There are 68 movies published on this site that do not use some time-saving glitches (although they might use others). 61 of them are Moons or better. Clearly the idea that introducing new categories that eliminate some or all glitches "removes absolutely all optimizations conceived in the speedrunning community" and results in a TAS with "all tricks removed just run right for great justice" is a bit absurd.
Pokota
He/Him
Joined: 2/5/2014
Posts: 779
To play Devil's Advocate for a moment, the three examples you listed don't just replace sonic graphically, they also replace his playstyle and mechanics with the new character in question - fundamentally it's a different game, just with scenery we've already seen
Adventures in Lua When did I get a vest?
Joined: 12/29/2007
Posts: 489
FractalFusion wrote:
I have been briefly thinking about a glitchless 100% TAS of this hack since a few days before the run was finished (not making it personally, just thinking about it). My feeling is that, although it would definitely make a good TAS in my mind, I'm not so sure that it would make a good submission to this site. I am aware that hacks in general have been received poorly on this site and I think it is the second-most likely reason for the TASVideos collective to completely tear down someone's submission (the most likely being alternate goals or "goals"); I myself have been guilty of such behavior in the past. I have become rather cynical in regards to submitting stuff for the site so I just want to warn about this.
I may be missing something, but I'm not aware of specifically hacks and alternate goals in general being looked down upon on this site, nor even specifically hacks and alternate goals that haven't been seen before. I'm guessing you're referring to recent submissions such as #4563: IgorOliveira666's SNES Super Mario Bros. Enhanced "warps" in 11:02.71, #4618: Baddap1's NES Super Mario Bros. Bowser's Jumping Challenge in 07:54.25, and #4619: Baddap1's NES Super Mario Unlimited in 13:03.46? Usually, a submission of a hack/alternate goal is looked down upon simply because it doesn't offer anything new in comparison to the vanilla game, regardless of how well optimized it is (in the case of the first example). I'm having a hard time finding an example of a submission that was viewed poorly only because it was a hack/alternate goal, and not because of some other significant factor such as the hack itself being subpar, or the optimization. TPP Anniversary Red would not fall under the above concern, especially if it were played glitchless 100% due to all the additional requirements completely new from vanilla Red. I think even a simple "glitchless beat the game once" category would qualify, as, per my original post, we have no published moves of Gen I that do not heavily abuse some strong glitch. I'd like to make a reference to #3651: bahamete's SNES Super Mario World: The Second Reality Project Reloaded in 34:40.18 and its acceptance comment here - a 100% run of that game would deserve to obsolete the current movie and would definitely stand out against vanilla SMW or even SDW, but a movie that simply beat the game was deemed enough to accept. (As a side note, I do not think the simple "glitchless beat the game once" category would qualify as standing out if a glitchless run of Gen I were already published (unobsoleted), but there isn't one.)
FractalFusion wrote:
That being said, I think that hacks should be TASed on their own merit, and not solely to get around some limitation in the original. Since the idea of TASing a hack is to show off the hack, "glitchless 100%" is the first thing that came to mind for this hack. Of course "glitchless" will always be subject to debate; personally I think that the specifics of what constitutes "glitchless" should be left simply to the author's discretion and let the TAS speak for itself; otherwise there is too much red tape.
Agreed on all of this.
FractalFusion wrote:
I think a good 100% would be completing all important events exclusive to this hack (gym leader rematches, Dream Red, Mew event, Battle Tent, catch all the Pokemon, and E4/Champion rematches).
Quick note: Excluding the Battle Tent, all of this including Dream Red is already a prerequisite for unlocking the E4+Champion rematch and the final boss after that. An end goal could be condensed to "Beat the final boss" or "Beat the final boss and also beat the Battle Tent once". (Identity of final boss not mentioned directly here in case of spoilers.)
FractalFusion wrote:
By the way, even though this hack seems a bit too similar to the original game, there are ROM hacks with movies that are accepted on this site that are hardly anything more than replacing the main character, such as: [2521] Genesis Knuckles in Sonic the Hedgehog by marzojr & WST in 13:05.29 [2752] Genesis Sonic 3 & Amy Rose by WST & marzojr in 29:27.58 [2514] Genesis Tails in Sonic The Hedgehog by marzojr in 13:21.87
Reply to Pokota: I believe what FractalFusion meant by this is that it's not necessary for scenery to change in order for a hack to be accepted, if there are enough other merits. The equivalent here would be a Pokemon hack that doesn't change the maps. (And even then that's not completely true; a few areas are brand new in this hack, such as the Battle Tent and the entire Mew event.)
grassini wrote:
Or can we accept the arbitrary category of "nido family ohkoing with horn drill x accuracy the whole game or just luck manipulate the ohko because it's tas"?
FractalFusion already covered this; Nidoking spamming OHKO moves is not a category but rather simply the fastest known way to beat Gen I glitchless. And "glitchless" or "no memory corruption" if you prefer is certainly not an arbitrary category. Under your suggestion it would be like calling the SMW run "uses Orb and Cape".
grassini wrote:
i think it's a bad idea,because the category removes absolutely all optimizations conceived in the speedrunning community.
grassini wrote:
Your suggestion just seems like those people who want a OoT glitchless intended route speedrun,all tricks removed just run right for great justice all over again.
Absolutely false. Many optimizations are still being discovered in the speedrunning community for glitchless runs of all kinds, even when glitched runs exist. Look no further than PokemonSpeedruns.com and deanyd.net for just two examples. And it's certainly not the case that all optimizations come in the form of glitches.
Joined: 12/29/2007
Posts: 489
Somewhat of an update: http://www.reddit.com/r/twitchplayspokemon/comments/31udel/pokered472ips/ Looks like the final version of the Anniversary Red patch has been released. I asked the ROM hackers about which glitches were fixed and it looks like the only ones are the Old Man glitch and the Poke Doll skip glitch (i.e. the only ones Twitch Plays Pokemon could readily abuse). Brock Skip/Brock Through Walls and Trainer-Fly seem to be intact.
Skilled player (1741)
Joined: 9/17/2009
Posts: 4981
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
Zowayix wrote:
Somewhat of an update: http://www.reddit.com/r/twitchplayspokemon/comments/31udel/pokered472ips/ Looks like the final version of the Anniversary Red patch has been released. I asked the ROM hackers about which glitches were fixed and it looks like the only ones are the Old Man glitch and the Poke Doll skip glitch (i.e. the only ones Twitch Plays Pokemon could readily abuse). Brock Skip/Brock Through Walls and Trainer-Fly seem to be intact.
Thread states that Brock skip was patched though: http://www.reddit.com/r/twitchplayspokemon/comments/31udel/pokered472ips/cq5m6x8
Appears to have been fixed by Pigu by forcing the NPC to talk to you before you can get behind him.
Also did they fix the battle tent bugs?
Joined: 12/29/2007
Posts: 489
jlun2 wrote:
Thread states that Brock skip was patched though: http://www.reddit.com/r/twitchplayspokemon/comments/31udel/pokered472ips/cq5m6x8
Below that comment, Pigu replies that it wasn't fixed after all, and Revo replies back that he's not going to further update and that a fix would "just annoy speed runners". Looks like a miscommunication that was corrected and then confirmed; Brock skip is not patched.
jlun2 wrote:
Also did they fix the battle tent bugs?
Yes. TPP did (finally) successfully use it several times toward the end of the run.
Skilled player (1741)
Joined: 9/17/2009
Posts: 4981
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
Zowayix wrote:
jlun2 wrote:
Thread states that Brock skip was patched though: http://www.reddit.com/r/twitchplayspokemon/comments/31udel/pokered472ips/cq5m6x8
Below that comment, Pigu replies that it wasn't fixed after all, and Revo replies back that he's not going to further update and that a fix would "just annoy speed runners". Looks like a miscommunication that was corrected and then confirmed; Brock skip is not patched.
jlun2 wrote:
Also did they fix the battle tent bugs?
Yes. TPP did (finally) successfully use it several times toward the end of the run.
Annoy speedrunners? That's a rather odd reason considering they fixed things like the pokedoll glitch/oversight. lol
Joined: 12/29/2007
Posts: 489
To be fair, Poke Doll and Cinnabar coast are only minor glitches and were disabled specifically because they were the only ones Twitch Plays Pokemon could reasonably pull off. Having Brock Skip or Trainer-Fly disabled would have had much, much larger effects on speedrunners than either of those two.
Joined: 1/1/2016
Posts: 1
I just wanted to be fair here and say that 4.7.3 will be the last version ever of this hack. It was posted almost a half a year ago and is the version on all physical cartridges of this hack. Source code can be located here: https://github.com/TwitchPlaysPokemon/pokemonAnniversaryRed if someone wants to exploit the Battle Tent or something by knowing how it behaves. Also: It's not easy to fix Trainer-Fly because it seems to just be part of the engine itself (being able to pause the same frame before you get spotted).