Post subject: Kirby's Avalanche
Player (105)
Joined: 6/7/2005
Posts: 290
Location: New York
I did a speedrun of this game and have already submitted it. I was wondering if anyone knew of any tricks or might be able to help me out with the randomizations in the game? What determines what set of blobs you get for a match? When is it determined? When is the placement of the boulder blobs determined? Is it possible to make the blobs settle in their spot faster then just laying them out onto other blobs?
Soft Blue Dragon
Former player
Joined: 3/30/2004
Posts: 1354
Location: Heather's imagination
I have a better idea.
someone is out there who will like you. take off your mask so they can find you faster. I support the new Nekketsu Kouha Kunio-kun.
Former player
Joined: 8/1/2004
Posts: 2687
Location: Seattle, WA
Have you tested what keystrokes on certain frames/pausing can do to the randomness. THe usual culrpits here are frame # and odd inputs (as in left/right or using buttons with no use).
Boco wrote:
I have a better idea.
That's a pretty shoddy idea.
hi nitrodon streamline: cyn-chine
Player (105)
Joined: 6/7/2005
Posts: 290
Location: New York
thanks Zurreco. I have tried the random button with no use pushing, that one seemed to do a little with the boulder blobs. i haven't tried pausing and using left+right yet...i will have to give it a shot
Soft Blue Dragon
Joined: 2/2/2005
Posts: 29
Location: Finland
In my experience the blobs you get in a level are all randomized on the start of the level and there is no way you can manipulate them during gameplay. The boulder placement however seems to be based on where you put your last pair of blobs before the boulder falls. And i did some testing with left+right & up+down, but they didn't seem to do anything (not atleast during gameplay) Even if there isn't much luck manipulation, you can still plan all your games in advance since the exactly same blobs will fall no matter what, this should give you a bit of a advance atleast :)
Player (105)
Joined: 6/7/2005
Posts: 290
Location: New York
In my experience the blobs you get in a level are all randomized on the start of the level and there is no way you can manipulate them during gameplay. The boulder placement however seems to be based on where you put your last pair of blobs before the boulder falls. And i did some testing with left+right & up+down, but they didn't seem to do anything (not atleast during gameplay) Even if there isn't much luck manipulation, you can still plan all your games in advance since the exactly same blobs will fall no matter what, this should give you a bit of a advance atleast :)
Thanks for your input on this. I do know that the randomization starts at the beginning of each match, but i don't know when. as far as i can tell it could be what time you get on the match before it. In that case I'd have to probably make some matches better in order to get a better set of blobs for the next match. there are a couple levels i know i could improve on..and perhaps improving on them would help me improve on the succeeding level.
Soft Blue Dragon
Player (206)
Joined: 5/29/2004
Posts: 5712
Or it could make the next level worse!
put yourself in my rocketpack if that poochie is one outrageous dude
Player (105)
Joined: 6/7/2005
Posts: 290
Location: New York
yes i know...i am hoping that will not be the case
Soft Blue Dragon
Post subject: Kirby's Avalanche!
Joined: 6/19/2005
Posts: 180
This will be my first TAS, and I am currently working out testing and calculations to help me. This game, simply put, is a clone of Puyo Puyo. Colored beans fall from the sky into a 6x12 grid, and when 4 or more are chained together, they disappear. Whenever beans are cleared in groups or chains, they produce grey trash beans that fall in groups on the enemy's side. The objective is to make the opponent unable to place any more beans on their side of the screen. I'm trying to work out formulae how much trash is produced by groups and chains. I've got some for a few basic things, but if someone can help me come up with a definitive group of equations, then that would be absolutely wonderful. So far, I've worked out that beans cannot be manipulated once they are set at the beginning of a level; they always come down in a set order. However, the outcome of the previous level can be manipulated (either score or time, I'm not sure) to produce a different set of beans. This means that I can catalog the exact order of beans and come up with a way to quickly and efficiently remove a large group from the field. Current guesstimates say that a total of 7 or 8 rows of trash beans would be enough to kill my opponent, but 9 or 10 would destroy them royally. That would translate out to around a 5 or 6 chain possibly, as high as 8, though. If anybody can help me, then I will be very grateful. :) Currently working on the equations for trash beans. For now, we have a formula we believe works. n(n+K)/7, where n is the number of beans cleared, and K is a constant dependant on the level of the chain. The constants for the chains are: 0: -3 1: 4 2: 12 3: 28 4: 60 Memory addresses to watch are: 7e0775: Trash beans produced by the current clear (0 if beans not being cleared) 7e0777: Trash beans waiting to drop 7e070e: Remainder to the equation (Holding down to speed up beans increases this number) When this number reaches a certain amount, it will add 1 trash bean to the next group cleared. BIG thanks to Nitrodon for helping me out with these values. [Edited once due to fluidity/content reasons] UPDATE! (so soon, I know. :P) Test run of the first level can be found here. http://dehacked.2y.net/microstorage.php/info/1235/Kirby%27s%20Avalanche%20%28U%29%20%5B%21%5D.smv It completes a single combo that kills the enemy (barely) and ends input. It's just a test run for now, and it just barely makes it before the opponent can do anything. I still need to get frame perfection down, and I do need to get everything down to the exact frame.
Joined: 10/24/2005
Posts: 1080
Location: San Jose
You do know that my real time run (actual speedrun) (recorded on an emulator in real time w/out savestates and rerecords) got 29 seconds on PBS... I would post the SMV, but it's pointless... because the current run does it in 24 seconds! There has been 3 runs for this movie, all of them published. See here: http://tasvideos.org/forum/viewtopic.php?t=2464
<agill> banana banana banana terracotta pie! <Shinryuu> ho-la terracotta barba-ra anal-o~
MarbleousDave
He/Him
Player (13)
Joined: 9/12/2009
Posts: 1559
This game and Dr. Robotnik's Mean Bean Machine are both the same game of Puyo Puyo with different re-skins.