STC : Wrong number of skulls are dropped

I have witnessed a bug in the number of skulls dropped. The bug can be observed in the following situation :

  1. Player 1 does a combo which drops skulls on player 2’s side.
  2. Game continues.
  3. Player 1 does another combo which brings his total score to exactly a multiple of 70*6.

According to the rules, I would expect that the number of skulls are dropped so that the nuisance points are brought down to 0.

On step 3), I observed during the game that one less line of skulls was dropped compared to my expectations.

I believe this might be due to rounding errors in the calculation of the nuisance points.

I suggest the bug would be advantageously resolved by storing the “70*nuisance points” with an “int” type instead of storing the actual nuisance points with a floating point type.

As of now, the bug is game breaking because we can NOT keep track of the opponent’s nuisance points if we can’t predict how many skulls are dropped.


I have seen this too.
Sometimes 420 points drops 6 skulls on the other side, sometimes they do not drop.
I suppose it depends on how the 420 was created, and how the rounding/truncating is done.

Thanks, we will investigate today.

1 Like

It should be fixed, but it’s hard to reproduce, can you please confirm?

1 Like

Thanks !

It is also hard for us to reproduce for me.

Let’s consider that the problem is fixed for now, and I’ll update the topic if I observe it once again.

I have programmed my bot in a way that it crashes automatically when this problem arises. It could help a lot if we could see easily in the match history that our bot crashed. At the moment, a srash is indistinguishable from a straight loss.

I’ve observed the same issue too in the past days. I’ve also noticed that there was an instance where the skulls dropped only a turn later, when no match has happened at all. Weird.

It looks fixed, thanks.