[1.2 beta] Sector AI STILL building Engineering labs everywhere

[Asuzu]

Sector AI ignores all science tiles and just slaps Engineering labs everywhere on "respect tile resources".
Please fix this. It's been since day 1.
Its a confirmed bug that engineering lab is first on build list, and it's still there.
My physics is at 40, enginering is at 70, and it keeps slamming engi labs everywhere.

 

[Mkennedy1983]

Still happening in 1.2 release.

 

[Hakugard]

Can also confirm it's happening in version 1.2.1 as well.

 

[Zemurin]

Same here. I end up with 1:1:1.5 ratio of sciences, very annoying.

 

[Birdy123]

Same here!

 

[Defiler99]

When I conquer a new spacefaring species, all their homeworld labs are engineering labs as well. This may help explain why it takes the AI so long to figure out Tachyon Lances, as well.

 

[Seomis]

While it defaults to engineering, it actually goes by which of the three research resources will yield the highest bump on a given upgrade. So if you have the "Prismatic Lenses" empire modifier, the sector AI will always build physics labs.

 

[ForzaA]

Thank you for your reports.

We'll probably want to reevaluate how the AI makes its decisions - As Seomis mentions, the AI strongly favours the "highest bump"
As an individual evaluation, that's probably exactly what we want. At the same time, we probably need to do some (more) work on maintaining some semblance of balance in research.

 

[Defiler99]

As an individual evaluation, that's probably exactly what we want. At the same time, we probably need to do some (more) work on maintaining some semblance of balance in research.

To take a specific example, let's say you have a bunch of empire modifiers etc etc, and in the end you have +200/+180/+160 physics/society/engineering income.

The "highest bump" algorithm will keep taking advantage of those bonuses or other influences, and diverge more and more.
If instead the metric was "pay off the highest deficit first", the AI would be looking at the following:
Physics: 200 is the highest of the three, so the physics deficit is 0.
Society: 200 - 180 = deficit of 20 points
Engineering: 200 - 160 = deficit of 40 points

In this situation, the AI would build an engineering lab, to attempt to catch up. Once Engineering passed Society, it would build some biolabs.
The above should pretty much converge on equal research, as far as I can see. I think "reduce the deficit" is closer to the typical human strategy.