Brutal-OXCE 2025-05-02

[Yankes]

I bump into bugs and even crashes every time I run OXC. It’s just grandpa project for more recent forks.

Just in case - I get crashes in OXC when using debug mode and teleporting my units all over the map. I guess it’s legit. My use case isn’t the same as for regular players.

Debug mode crash is not "valid" crash :> (probably I personalty fix that teleport with 2x2 units in OXCE because I needed to test somethings).

All crash should be fixed and where fixed in OXC, many cases it was backport from OXCE fix.
If fix is small then PR with it will be merged to OXC sooner or later (depending how much affect players).

[jnarical]

Debug mode crash is not "valid" crash :> (probably I personalty fix that teleport with 2x2 units in OXCE because I needed to test somethings).

All crash should be fixed and where fixed in OXC, many cases it was backport from OXCE fix.
If fix is small then PR with it will be merged to OXC sooner or later (depending how much affect players).

I got the same bug in both OXC and OXCE, where you could shoot with aimed to one target, with high >80% accuracy dozens of times without   a single hit.

As I was tinkering with accuracy code for long enough, I made a small “mod” where shot from laser pistol or rifle takes 1 TU, so I could easily confirm something is wrong, just by shooting many times without changing anything else.

I ain’t 100% sure I fixed that in my version of code, most probably I did. But I can’t make myself put any effort to debug that again, knowing that my another PR is out there for two years already, where I fixed a bunch of issues with visibility, making OXC behave closer to OG, particularly for cases when in OG one unit could see another, and in OXC, in the very same conditions - he is not.

[Yankes]

I got the same bug in both OXC and OXCE, where you could shoot with aimed to one target, with high >80% accuracy dozens of times without   a single hit.

As I was tinkering with accuracy code for long enough, I made a small “mod” where shot from laser pistol or rifle takes 1 TU, so I could easily confirm something is wrong, just by shooting many times without changing anything else.

This could be simply RNG fault, accuracy do not mean how often you will hit target but how spread its trajectories as you should know. Add some RNG basis, some distance and you could have cases when spots you can't hit targets in many hits.

I ain’t 100% sure I fixed that in my version of code, most probably I did. But I can’t make myself put any effort to debug that again, knowing that my another PR is out there for two years already, where I fixed a bunch of issues with visibility, making OXC behave closer to OG, particularly for cases when in OG one unit could see another, and in OXC, in the very same conditions - he is not.

I mean crashes not glitches. If you test long enough you will find many corner cases in engine (some could even be my fault ). Probably your PR get ignored as it need same amount of work on our side to analyze as you did to create it.
Because we need to check if it not introduce some bugs in other places or if its correct in first place.
This mean crash bug fixes are merged as adding one `if` is a lot easier to verify its correction even without running that code.

[jnarical]

This could be simply RNG fault

Imagine 20 shots in a row with 85% going through exact same voxel.

[DeltaEpsilon]

There's one behavior I noticed that, in my opinion, should be reduced.
In the first few turns, enemies tend to rush the Skyraider or whatever transport you have, clearly treating it as the mother of all covers. This behavior is very easily exploitable because they tend to stand in the same spots each fight, which, coupled with a decent transport, could allow your troops to kill them and go back into the safety of the transport.
Not only that, but it doesn't seem like the AI realizes those spots are heavily dangerous. Perhaps makes the area around the transport a no-go zone for a few turns?

[Xilmi]

There's one behavior I noticed that, in my opinion, should be reduced.
In the first few turns, enemies tend to rush the Skyraider or whatever transport you have, clearly treating it as the mother of all covers. This behavior is very easily exploitable because they tend to stand in the same spots each fight, which, coupled with a decent transport, could allow your troops to kill them and go back into the safety of the transport.
Not only that, but it doesn't seem like the AI realizes those spots are heavily dangerous. Perhaps makes the area around the transport a no-go zone for a few turns?

What Aggression-setting are you using?
The default is 1 and on 1 they should not "rush the Skyraider".
Using a higher aggression-setting nerfs the AI because it inevitably leads to worse play.

Maybe I should remove settings that make the AI play worse than it can. Otherwise players use them and will then complain about AI being stupid.

[DeltaEpsilon]

What Aggression-setting are you using?
The default is 1 and on 1 they should not "rush the Skyraider".
Using a higher aggression-setting nerfs the AI because it inevitably leads to worse play.

Maybe I should remove settings that make the AI play worse than it can. Otherwise players use them and will then complain about AI being stupid.

I'm using 2, though it also seems to rarely occur even with 1. Also, I beg to differ, aggression 2 makes the AI quite effective too because it doesn't let me to do the usual "throw smoke and wait a turn" deployment since some enemies inevitably end up going inside the transport and spotting my troops, leading to around ten bajillion grenades coming in, blowing up my item pile and possibly instakilling everyone, meaning I'm basically forced to deploy turn 1 and suck up the reaction fire. Or, in other cases, they also like to do suicide charges by going in with a preprimed grenade and basically blowing themselves up together with my troops. So it's definitely not entirely ineffective, but it is exploitable regardless for this particular point where they want to bunch up near the Skyranger. I don't mind the heightened aggression considering the enemies have the numerical and often damage advantage as opposed to more ambush-y aggression 1 type that works better for attrition.

[xenoroyal]

I played aggression 2 to get the hang of brutal AI because I wasnt learning much getting 12 soldiers killed at entrance to large scout

sometimes it was fun having 5 floaters ontop of your skyranger or getting rushed by snakemen and mutons,but in the end I was able to outplay the predictable behavior

thats why I think aggression 4 could be fun, if not the most challenging.
Id probably go through the unit aggression file for a playthrough of it and change cap setoids/floaters aggressions at 1 but keep snakemen and mutons at their natural 2.
would make for a varied playthrough i think