Oh interesting. I hadn't heard the reasoning before but it makes sense.

warcraft 3 and aoe 2 is 0,25 sec if i'm right, so it's ok

Interestingly, there is a difference between HP and armor bonus when it comes to healing: the armor doesn't reduce healing as it does taking damage. That would make an armor bonus stronger than a HP one. When Ptolemies had this cavalry hero that gave extra 40% HP to pikemen in his aura, you could heal the pikemen faster by moving them out of the aura. I remember playing some games from the times before the exponential armor system, and there seemed to be a problem where if a unit or building had more armor points than you had damage points, you wouldn't be able to cause damage to it anymore. And so there was a hard-coded minimum of 1 HP damage (one other detail is that nowadays the HP uses the fractional part, even if it's not shown). That could mean that you were better off using a greater number of very weak units (like archers) because you had the the guarantee of at least 1 HP per arrow shot.

turns are calculated 5 times each second, so depending on your input, there could be up to a 0.2 second delay for a unit to act on your orders. In laggy games with 8 players, this is exacerbated by slower players since simulation waits for the slowest players to "catch up". with substantial performance improvements, it could be lowered to 0.1 sec.

did you try in 1vs1 with someone of your region? you will see , it's good

Hi all, wanted to give a little rant.  For an RTS game where you are doing many actions very quickly and need a decent response speed from units having a reliable latency is needed. Its not fun for me when the there is a serious delay between my order and when it actually happens. In single player there is a delay as well but its much faster compared to pvp. Anyone else experience something similar?

About zram, I would suggest starting to use lz4 (it should be faster) rather than zstd, given you already have plenty of RAM. Then, you may want to consider upgrading to Debian testing/13, it has already less known bugs than 12 ( https://bugs.debian.org/release-critical/ ) and it can support building 0ad using its packages, other than natively having 0ad 0.27.0.