Yes, blind spots include unused tech, tactical and strategic positioning the latter being probably the most critical component of the game (planetary improvements, tech progress and ship design being mostly streamlined)sven wrote: My own impression is that AI has always had big problematic blind spots, and I think those are hurting the game no matter how I balance it. Mistakes in fleet routing are the category of issues that most has my attention right now -- particularly in mid/late game, when you have large empires and/or big alliance blocks, the AI's tendency to send ships scrambling from one side of the map to the other, missing out on all the critical battles that are taking place while they're in transit, is a real problem. That's true under the old balance as well, though it's perhaps more pronounced now, because games are lasting longer.
Pre-coin rebalance, the AI did a decent job as long as it wasn't in alliance and its systems have been clustered together and not stretched out to much.
Now, it struggles to build ships and thus its flaws are more apparent (on hard I have seen single Yoral torpedo destroyers trying to take on two star bases + PD)
Most importantly: from what I have seen so far, it does not weight system importance based on ownership
Specifically: an allies system seems to be considered as its own and does not take into account the diplomatic relationships of its allies to assess their system weights
Attached an example where we are looking at Phidi right after I have offered up alliance as Orthin and joined its war against Tinkers
You can see that it has prioritized two of my systems to defend while its own core systems are right next door to a huge chunk of enemy territory. Also note that Orthin are only in war with Tinkers and has virtually no conflict with its other neighbors. The Orthin systems have no PD while Phidi systems have them, so this might be the reason
It has also marked a Tinker system to attack which is the most heavy defended system of the Tinkers (two star bases and PD) instead of going for Regulus which has no defenses whatsoever
Thus regarding positioning:
1. AI should probably weight ally systems lower and only defend any of them if it has at least one system in its core cluster which is defended. This is to never leave its core system cluster undefended as simply sending them over to an ally taking above 4 turns to get there could easily mean defeat for them when attacked
2. Defense priority should weight ownership, threat to system (considering diplomatic relationship), system value, system defense (PD, bases and ships: existing and inbound) and proximity. System defense seems to largely overweight the other attributes
3. Attack priority does not seem to take planetary defenses and bases into account, or if it does, it prioritizes proximity too much. Primary target should be both in close proximity and have low defense at the same time - the two should roughly have the same weight