From movie rules:

Must be clearly definable as a game, which has achievable goals

Should of waited til 4/1.

Oh no, it is gonna be rejected, I'm gonna do a TAS of Hebereke's Popoitto.

Alright, So instead, I did a TAS of Hebereke's Popoon.

Samsara has outline a number of items that are clear red flags for optimization problems.

welcome back samsara!

Samsara has outline a number of items that are clear red flags for optimization problems.

Would you mind taking a look at the movie yourself and verifying them? I saw you in the chat for the RTA run I linked, so I'm assuming you know enough about the game and speedrunning it to be able to confidently confirm/deny my suspicions.

This is noticeably suboptimal in a lot of places, and I've only watched the first stage.

• The rerecord count is ridiculously low for a game like this. Piece order can be manipulated before each stage, but not in the middle of a stage, so a TAS would want to at least make sure that the piece order is suitable for a fast stage completion.

[*] Speaking of fast stage completions, this is a Puyo Puyo style game (not exactly the same, but similar enough), and yet combos aren't being used or even set up in this TAS. The game has them and accounts for them. Not only is it far more impressive to set up a massive chain in advance, but it should also be much, much faster than just going for 3-in-a-row clears over and over.

[*] The input looks noticeably sloppy. Autofire looks like it's being used in places just to turn pieces, leading to situations where you use 5 B presses to turn a piece twice. At the beginning of the first stage, you turn and position pieces before you start fast dropping them, which loses time, and it looks like the only reason you stopped doing that is because you started holding Down for the entirety of the rest of the stage. It looks like you just did the entire run in "real time", but at a low speed, as opposed to using frame advance or TAStudio. The menuing at the beginning is optimal at least, but the random Start presses while waiting for "PRESS START!" to appear tell me that you didn't bother trying to correct them. It's only a minor nitpick, but it still leans into the sloppiness of the input. Nearly all of your text clearing input is overshot. The last actual input of the run happens at frame 142940, but the A button continues to be held for 5 more seconds after that, not to mention the almost 2 full minutes of blank frames at the very end.

[*] This was done on the default difficulty (Monkey). Hardest difficulty should be used for a TAS like this, as it's much more impressive to see the domination of a difficult CPU opponent as opposed to an easy one. The faster drop speed on higher difficulties may even make the run shorter, since attack pieces don't drop in until a normal piece lands.

[*] On that note, the most directly comparable RTA run I could find (there's a much faster one on speedrun.com but it's on a different version of the game (Arcade), and appears to be on a different mode altogether)

To sum up, going for optimization in this game seems to be a little odd...depending on what needs to occur to wipe out the opponent. The biggest issue that I see is a lack of executing a plan, which the current plan is just taking on the computer by thinking one move ahead, rather than accounting for situations to avoid opponent obstacles, setting up for combinations, and trying to push the opponent off the screen early by noting the CPU's position and circumstance. On the first round here...the strategy used push the CPU to the very last space possible, which takes a longer time...especially for the cut scene of removing all the drops.

As a quick example, our currently published Puyo Puyo 2 run is 36 minutes shorter than this (with overshot input/blank frames removed), but it uses about 6x the number of rerecords. It also has one less variable to account for compared to this game: The CPU's blocks don't disappear individually after a battle so it doesn't matter how many they have on screen.

Hello and welcome to TASVideos! I appreciate the efforts went into the movie. A good tool assisted speedrun or superplay can take a lot of time to create. So first, I got rid of one of the 5 Start button presses at frame 528 in TAStudio (my very first time using it! It really comes in handy!). When I did delete that Start press, it seemed to affect the piece RNG and caused the movie to desync earlier. So theoretically you can shuffle the piece RNG with different inputs to get the most desirable piece orders possible. Is there any particular reason the default difficulty setting (おさる/Monkey) was used over たつじん/Expert in a TAS? Most Puyo Puyo-like puzzle games I know increase the drop speed in addition to higher A.I. levels. At the first round, it looked like you built a few stairs (placing columns of the same color). Couldn't you setup chains from those stairs? The SFC port of Puyo Puyo 2 was the best version to use (in my opinion), due to the added ability to skip the opponents' defeat "animation". I would like to see this falling piece puzzle game get optimized further. Popoon could end up at Vault tier, as with Wrecking Crew '98.