TETRIS® THE GRAND MASTER 4 -ABSOLUTE EYE-

TETRIS® THE GRAND MASTER 4 -ABSOLUTE EYE-

View Stats:
[RESOLVED] Player vs. Player - Error with Garbage TYPE 2 logic
When sending garbage in the traditional TGM style (TYPE 2), the garbage can be received in the incorrect order. This affects all Player vs. Player, however the bug is most obvious in Konoha so I will use that for my example.

Please look at the Player 2 screen to see the problem:

https://gtm.steamproxy.vip/sharedfiles/filedetails/?id=3464675209

And here is a video showing the gameplay which produced the bug:

https://youtu.be/g0qTeQ4yhVM

In this example, a second attack is sent while the first attack is being received. The second attack does not wait it's turn, but instead it cuts in line and immediately sends damage. This creates an impossible pattern for Big Mode style gameplay.

I think the issue is because garbage is implemented as a "stack" instead of a "queue". I am sure Arika's programmers are skillful, but to make sure the meaning is clear (after translation) here are references in Japanese:

https://ja.wikipedia.org/wiki/%E3%82%B9%E3%82%BF%E3%83%83%E3%82%AF
https://ja.wikipedia.org/wiki/%E3%82%AD%E3%83%A5%E3%83%BC_(%E3%82%B3%E3%83%B3%E3%83%94%E3%83%A5%E3%83%BC%E3%82%BF)
Last edited by colour_thief; 22 Jul @ 9:51pm
< >
Showing 1-5 of 5 comments
Kitaru 15 Apr @ 6:08pm 
+1, good demonstration of the race condition.

I agree that it seems like attacks are being processed on a Stack; I noticed that if two attacks are sent before the opponent places a piece, the opponent receives attacks in a Last In First Out order. At the time I felt this differed from the mechanics and strategy of previous titles, but did not look into it more at the time.

Upon reviewing the behavior in TGM2, it seems that previous titles processed attacks in a Queue; if two attacks are sent before the opponent places a piece, the opponent receives them in First In First Out order.

While the priority is of course to fix the race condition, I would also like to say that I prefer the strategic implications of the attack queueing rules from TGM1-3.
zaphod 16 Apr @ 8:23am 
i actually like the ability to interrupt an attack with a second attack, as it's a way to send trickier garbage, but it's just too broken in big mode when it can cause a half a line like this.
I have played 1000+ hours of TGM versus and I can assure you that a stack is worse than a queue. Yes it's a "way to send trickier garbage", but it's far too easy to execute and doesn't have any depth. It's a shallow strategy with no variety.

In contrast, the queue mechanic requires lots of risk vs reward trade-offs and requires deep skill to achieve tricky garbage.
Kitaru 17 Apr @ 3:19pm 
Just to make sure it's clear where the Race Condition bug is happening, I've created a diagram with the attack segments highlighted. The diagram on the left is the expected/desired result, while the diagram on the right is the result of the new attack being inserted in-between the existing one.

https://gtm.steamproxy.vip/sharedfiles/filedetails/?id=3465709628

The issue is caused by the 2 rows from the L-mino attack being pushed to the Stack before all 4 rows of the T-mino attack have popped. This type of attack pattern is effectively impossible to defend against with Big tetrominoes.
Kitaru 22 Jul @ 6:25pm 
I think this was indirectly addressed with the changes in version 1.7.0, as the garbage rising animation phase was removed; any pending attack is added to the playfield instantly. As such, it should no longer be possible to interrupt the pattern of previous attack with a new one. This issue can be considered resolved for the time being.
< >
Showing 1-5 of 5 comments
Per page: 1530 50