AI Research Oddities

[Uncle_Joe]

After playing a handful of game as different Powers, I am wondering as to the AI's research in some cases.

The ones I really think need to be looked at are:

1) Germany researching Light Fleet ASW...I'm at a completely loss here. I cant even fathom a situation where this is worthwhile unless they are holding the Med and even there, its marginal at best.

2) Germany researching Tac Bomber Torpedo...to level 3...Again, while I understand the desire to want to kill Allied Transports, it rarely does so (may 1 or 2 per game with the bombers). Given that, I think the AI's research could be better spent elsewhere (like on Uboats or better Surface Fleet Evasion or something that accomplish the same goal, but better).

3) Lack of research for Carrier Air Group Evasion. Both the WAllies and Japan tend to ignore this one for most of the game meaning that Carrier Air in general is not really accomplishing all that much. Aircraft losses are WAY too high without it. I'm starting to think that BB AA should be reduced by one at start to prevent this one from being an absolute necessity every game.

There have been a few others that have raised some questions, but the above three are the biggest ones. With a few tweaks to the AI research, I think it could prove far more challenging. Japan does an OK job of managing the Pacific war if given the 200 Transport help and some supply help. But what kills it every game is the hideous losses their aircraft take to enemy AA fire.

The German AI also does a fair job in many places given appropriate help. But again, I think it could be even more efficient if it didnt squander precious research on what are undoubtedly 'luxury' items

-data captured-

[Dalwin]

I can see putting torpedo on the German Tacs, though I would usually go with ship attack instead. ASW for the Germans is downright silly, however.

I never play with FoW off, so I don't know if the Chinese are wasting a few of their precious production points on starting projects that they are unlikely to ever finish.

[aletoledo]

good points, I never have been \out teched' by the AI, even against the WA ai, pretty much as you've described.