How is this different from Tetris Attack?

You would think that Tetris Attack was just a re-skin of Panel de Pon, but it's not. The various artificial intelligences (AI) play a little bit differently between each version. It's these differences that force you to speedrun the hardest modes of each game a little differently.
 
First off, what's the same?
 
Both AIs scan from the top of the screen downwards looking for matches. This makes it beneficial to include a bunch of combos with your chain so that way even if your garbage is cleared on, the combo garbage will turn into normal garbage that your opponent hopefully will find a match in and scroll from the bottom of the screen to the top. Most opponents will be too slow to do this.
 
Both AIs will also tend to make combos to clear their garbage with. This can lead to sneaky wins that would otherwise not be possible.
 
So what's different?
 
After stage 8, Tetris Attack AI switch to prioritizing chains over combos. Chains will give more stop time, but they will be harder to find. You may be able to get away with sending less garbage on these opponents than the earlier ones due to them not clearing as many panels. Panel de Pon AI, on the other hand, will continue to look for combos in the later stages. This mindset will enable them to clear more panels quicker, likely forcing you to send more garbage than in the earlier stages due to their increased speed.
 
For that matter, Panel de Pon AI, in general, appear to be faster per stage than Tetris Attack. This is not confirmed, but should be straightforward to test.
 
The biggest difference, however, is what they do at the very start of the match. Tetris Attack AI will tend to raise their stack before making matches. Panel de Pon AI will do the reverse. What this means is that the Panel de Pon AI stack will generally be much lower on average, forcing you to send more garbage on average--like, sometimes 2-3 lines more on average, which takes quite a bit of extra time.
 
There is a tactic that can be used to slow down the Panel de Pon opponents. If you send garbage to them immediately, they will be forced to deal with the garbage immediately instead of depleting their stack. You can then send a more Tetris Attack-like amount of garbage to them that should fill their stack. When done correctly, the stage ends in about 15-18 seconds most of the time. Unfortunately, this turns out to be much too skill-intensive and inconsistent for standard play.

Memory Addresses

Addresses are based on BizHawk

000338 (1 byte)	Stage Timer (seconds)
00037E (1 byte)	Stage Timer (frames of second - 60 frames per second)
000418 (1 byte)	flag for determining when opponent's screen should be shaking (0=no, 1=yes)
000420 (1 byte)	counter for opponent's screen shaking
001860 (1 byte)	Player 1's bottom-left tile
001960 (1 byte)	Player 2's bottom-left tile
001A76 (4 bytes)	RNG value
001A80 (1 byte)	Player 2's CPU Speed
006D61 (2 bytes)	Player 1's Stop Timer
006D63 (2 bytes)	Player 2's Stop Timer
006D67 (1 byte)	opponent's screen shake level

S-Hard RNG Manipulation Theory TAS

https://www.youtube.com/watch?v=8FAC0hGlpAo

Until I can shape up this leaderboard...

Here is the Nico RTA leaderboard for this game:

http://niconamarta.sakura.ne.jp/index.php?game=%E3%83%91%E3%83%8D%E3%83%AB%E3%81%A7%E3%83%9D%E3%83%B3

الاسم	المؤلف
How is this different from Tetris Attack?	CardsOfTheHeart 8 years ago
Memory Addresses	CardsOfTheHeart 6 years ago
S-Hard RNG Manipulation Theory TAS	CardsOfTheHeart 8 years ago
Until I can shape up this leaderboard...	CardsOfTheHeart 8 years ago