Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Buttons will always be faster or at worst tie with touch screen. It all depends on how the game recognizes touch screen input– if on touch, it ties, if on leaving touch, it can lose 1 frame compared to buttons.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
In “point and click” phases would it be faster to use the stylus + A button on the same frame to point and click on the same frame? It all depends on whether the stylus input is read when you touch the screen, or when you leave the screen. If the latter, stylus + A could save a frame.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Finding XY is easy. 4 bytes signed search, standstill and it’s unchanged. Start moving to the right, X is probably bigger (or you can just search “changed value”). Move left, X probably smaller. Stand still, unchanged. Change level*, repeat the process. When you have low enough results, in DeSmuME 0.9.11 you can mass-cheat RAM Search by highlighting multiple addresses. Click the top address, scroll down a bunch, hold Shift and click on an address to highlight them all from the first to that point. Don’t cheat too many all at once or things can get glitched/crash easily. After highlighting you just click Add Cheat and hold down the Enter key to mass-freeze all the addresses. If the game hasn’t crashed and you’re still able to move, you can delete all the highlighted addresses and remove the freeze cheat from the Cheats menu, because XY aren’t in there. If after freezing a bunch of addresses you’re stuck and can’t move, good job, inside that bunch is the XY you need to find. After you find X, Y is usually the address right after– if X is 0x02000000, Y will be 0x02000004, because X is 4 bytes big (0x02000000–0x02000003 = 4 bytes).
*You may want to skip this step if you can’t find the address with it. Sometimes addresses are different for each level, in that case you will need a pointer but I’m not gonna explain how that works. Most of the time the addresses are unchanging anyway.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
coldsun0630 wrote:
Link to video
Well, it seems the current version of youtube encode is not the newest. Here's the lastest snark's video of 4'97"93 which includes the comparison with the submitted 5'02"25.(log-in required) Posting for who didn't watched yet.
I thought he improved the BitS entry… but there’s no improvement going by that vid. Also none in BitDW. There are 2 frames on the Bowser fight in BitS though, lol.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Kurabupengin wrote:
ALAKTORN wrote:
Why are you TASing it then?
I don't really know what do you mean by that...
I’ve told you before that the starting point of a 2D game TAS is finding the XY character coordinates in RAM. Speed can be useful but the most accurate way to see speed is through Lua scripts, so I don’t use RAM for that.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Chef Stef wrote:
I'd also appreciate if you would quit insulting me.
Never have. All my insults were ifs that shouldn’t apply to you. You didn’t make it clear at all that you were gonna contact the MvsDK community, though; so you can’t pretend I should’ve known.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
All I’m saying is that you have no idea how hard this game has been destroyed to death by Prenz. If you want to make a TAS of this game, not contacting the best player of it who spent years upon years optimizing every single possible thing about the game and finding exploits all over, you’d be really stupid.
That’s true of any game, if you’re planning on doing a TAS of anything you should reach out to its competitive community for advice; but it’s especially true for this game for the simple reason that Prenz is the name of one of those competitive players.
100% strategies may have nothing to do with any% but that’s no excuse to not even get in contact with the game’s community. If you were to make a 100% TAS and didn’t look at Cyberscore’s records or hit Prenz up, I’d wager that’d be grounds for being banned for the sheer stupidity.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Score records are still speedruns, though. The timer gives points. If you calculate all possible points in a level and subtract them from the score record, you could figure out the time with which the level is ended. For your TAS that time should never be faster than yours, obviously.
I know nothing about the game but I know Prenz as a competitor and I am 100% certain that your TAS will be suboptimal shit if you don’t get in contact with him (he probably won’t help though) or somehow try to figure out some of his records to understand how to maximize the game. He may not tell you his tricks but he might still comment on whether what you have is optimal or not, so you could try asking him that if you can contact him.
Edit: By the way another competitor in that game (Crono) is also a TASer and has made a few TASes for this game, so you might wanna contact him as well. He’s nowhere near Prenz’s level though, I believe.
Edit: Here are a bunch of Crono’s MvsDK vids, maybe you can find something useful in them: https://www.youtube.com/user/CronoNes/search?query=mario+vs Edit: I think most of those videos are massively outdated, though…
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Are the Cyberscore records comparable to this TAS?
ALAKTORN wrote:
just noticed this thread… I’m not sure if they’re comparable at all, but Prenz is crazy on this game and his scores are here http://www.cyberscore.me.uk/game/22
I know this game has quite the version differences… earliest US release being the best as far as abusable glitches go, I believe
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Weatherton wrote:
They do in-fact, plan to share the movie file and actually submit it to this site.
My issue with that is the fact that they probably don’t know how to replicate the Japanese improvements, at least not yet. I’d rather they not submit a movie that has known suboptimalities to it.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Holy shit that was so annoying. Had to go through this so many times trying to manipulate the enemies, and every time I manipulated them away at one spot, they hit me at the next one. Jesus Christ. 19 empty steps, old was 25; 67 frames saved for a total save of 159 frames. 01:10.48 finish. I kind of think the TAS is perfect now but you can never know. Disregarding enemies it might be possible to do 1 or 2 steps better for level 5, but 19 is my record both with and without enemies.
Edit: Porco dio I fucked up the movie file gotta find it again.
http://tasvideos.org/userfiles/info/24955764634818867
I messed around with this a bunch. It was a TAS project file (even though it didn't start nor was supposed to end as one), so I removed the GreenZone files stuff and renamed it to plain .bk2.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
FractalFusion wrote:
This video is basically Sword Art Online (an anime about being in an MMORPG), cut in such a way that it makes references to elements typically found in RPG TASes (such as "going abnormally fast", event flags, cutscene skipping, and low-level runs).
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
RGamma wrote:
ALAKTORN wrote:
TASeditor wrote:
Figuring out the best route is just testing all possible routes.
That’s not very helpful.
As are your and my posts. Just post those kinds of things, when you have something better to offer, because this kind of problem is not exactly trivial (given there's an adversary)...
Lol, I thought it was funny. You shouldn’t care about the adversary. I don’t when making my strategies. Chances are it’s possible to manipulate them, or there is an equal solution done sligthly differently where the enemy isn’t in the way. Edit: The only level so far that can’t be done perfectly due to enemies is actually level 1. The more complex the level is, the easier it is to manipulate/avoid the enemies.
Experienced Forum User, Published Author, Player
(99)
Joined: 10/19/2009
Posts: 2527
Location: Italy
Zeupar wrote:
Oh, I thought that finding the optimal strategies for a video game was one of the most important parts of the process of "TASing intellingently". I mean, if one isn't good at that, their TASes will probably end up being among the "tons of suboptimal TASes published on the site".
Strategy finding has actually nothing to do with TASing. Most strategies used in TASes come from the RTA community (have you seen Paper Mario: The Thousand Year Door?), or are otherwise found in a non-TAS setting. Playing a game frame by frame doesn’t help very much with coming up with new ideas and effective ways to do things, what it helps with is to get all the inputs frame perfect and optimizing the game.
For another example, before TASing a level in Chase, I run through it in real time and figure out the best strategy. Only then I start TASing.
