Well, I think this is a bug:
You can see, I beat this person on both sides and I get a loss for some reason:
Recently I changed a lot in many systems to make it perfectly symmetrical and still some randomness somewhere, however now its really minor
However I agree β something is weird with that player. need to dig
1 Like
Everything looks ok. I spied that user and lose against many players. My suggestion that there were some problems when I changed components and systems and maybe it caused cache issued on the simulation side. Lets watch and wait. If it repeats then I dig deeply. Ping me please if you notice something.
Hmmmm, I need to check one weird thing and possible exploit with anomalies.
ok. So there is a trick β if you are using hero abilities in an event handler then they block each other but you can build in the same time. Iβm going to fix this, so be careful with using hero methods inside event handlers.
Fixed it and I hope I havenβt broke anything 
So as βbuildβ method takes 1 frame if you have enough mana and βx frame while enough manaβ and it blocked the main thread. However it was possible to cast abilities in parallel in the guard thread, so you could build towers and cast abilities in the same time. That was not designed such way, as the idea that βabilities cost you timeβ.
If a player on the human team casts an anomaly, that is added to findActiveAnomalies() for the ogre team on the same tick.
So if both sides have code like:
for anomaly in anomalies:
if isEnemyPull(anomaly) == True:
castTeleport()
castPull()
Then the human side would cast the pull, but the ogre side will notice it and cast teleport in the same frame, violating symmetry.
Is there any way to detect what towers the opponent is building?
You can use this:
Thanks a lot! This helps me! 
1 Like
Btw, the nerf to the archer tower and the buff to cannon didnβt get added to docs
@Bryukh Is this a potential source of breaking symmetry? (Along with the anomalies thing I posted earlier)
Nope, this is βgoodβ randomness based on a seed and for the same pair of players the seed would be the same. There is still some really minor symmetry which can affect the match result if you are really close in base hp. Also some players can use βMath.randomβ and this randomness we donβt control (on purpose).
You are right. Fixed. Thank you!
These are not clones. Just schools or group tournaments. The ladder doesnβt affect on the final tournament results.