About AI...

[bradfordkay]

"I think the resupply routine is really weak in this area, as I have seen the AI dump off troops for an invansion, and if it doesn't take the objective in a timely manner it will just leave the troops there to starve without resupply."

Rather similar to what the Japanese did IRL... okay, they sent in about 1/10th the supply needed, as opposed to none.

[Shark7]

ORIGINAL: bradfordkay

"I think the resupply routine is really weak in this area, as I have seen the AI dump off troops for an invansion, and if it doesn't take the objective in a timely manner it will just leave the troops there to starve without resupply."

Rather similar to what the Japanese did IRL... okay, they sent in about 1/10th the supply needed, as opposed to none.

Yes that is true. However the problem is that I usually play as Japan against an ALLIED AI, so it should be resupplying its troops to be historically accurate. [:-]

[Pyrrhos1976]

Is this problem systematic ?

[decaro]

ORIGINAL: treespider

Reading through this thread I don't believe anyone actually answered Pyrrhos1976's question....

For a small short scenario that is limited in scope is the AI at least somewhat competent???

... "W/the WitP/UV engine, the longer the scenario, the lamer the AI as it progresses."

Therefore, the shorter the scenario, the less lame/more competent the AI. By inference, the answer is "yes".

[Shark7]

I don't know if the problem is based on the GE itself, or on the way that I play the game.

Any time I see an AK/AP task force headed anywhere, I will dispatch my carriers and as much LBA as is feasible to sink as much of it as possible. I just wonder if this is affecting the AI's decision making on whether to send follow on supply TFs or not, being that I usually end up with a large number of air and surface forces in the attack area. It could just be that it is doing the smart thing by not trying to force a lightly escorted tranport TF into the teeth of the dragon, so to speak.

Which is why I was wondering about other players experiences in this situation.

[decaro]

ORIGINAL: Shark7

... Any time I see an AK/AP task force headed anywhere, I will dispatch my carriers and as much LBA as is feasible to sink as much of it as possible. I just wonder if this is affecting the AI's decision making on whether to send follow on supply TFs or not, being that I usually end up with a large number of air and surface forces in the attack area. It could just be that it is doing the smart thing by not trying to force a lightly escorted tranport TF into the teeth of the dragon, so to speak.

Isn't WitP based on the PacWar game engine, which had (bomber) Air Zones of Control to prevent the passage of routine convoys; these same convoys would "never be run into areas where naval and air combat were to be expected."

In my experience, this never really worked in PacWar, so why would it be expected to work in WitP?

[Shark7]

ORIGINAL: Joe D.

ORIGINAL: Shark7

... Any time I see an AK/AP task force headed anywhere, I will dispatch my carriers and as much LBA as is feasible to sink as much of it as possible. I just wonder if this is affecting the AI's decision making on whether to send follow on supply TFs or not, being that I usually end up with a large number of air and surface forces in the attack area. It could just be that it is doing the smart thing by not trying to force a lightly escorted tranport TF into the teeth of the dragon, so to speak.

Isn't WitP based on the PacWar game engine, which had (bomber) Air Zones of Control to prevent the passage of routine convoys; these same convoys would "never be run into areas where naval and air combat were to be expected."

In my experience, this never really worked in PacWar, so why would it be expected to work in WitP?

Well if that is the case, then it does explain why the AI never makes any attempt to resupply assault forces.

[Gem35]

If you play on hard setting the AI is always supplied so it doesnt really matter, add to the fact the AI is scripted and will do the same things every game again it doesnt matter. Name a game who's AI can think and react like a competent human being.
If you want that challenge start a pbem with somebody.
How many threads are there about how lame the AI is?
Shut up already.[;)]

[decaro]

ORIGINAL: Gem35

... Name a game who's AI can think and react like a competent human being.

BoA, GoA, etc.

[Pyrrhos1976]

Other questions about AI. In your experience:
1) has AI a tendency to scatter his carriers or is he able to concentrate them for striking ?
2) how does he manage the different TFs (bombardment, transport, air) for an invasion ?

[decaro]

The AI is capable of forming TFs for bombardment, transport, etc., but I've never seen it coordinate or concentrate any of them together for an invasion.

[Pyrrhos1976]

ORIGINAL: Joe D.

The AI is capable of forming TFs for bombardment, transport, etc., but I've never seen it coordinate or concentrate any of them together for an invasion.

[X(] But how can it succeed an invasion ??

At least, after the landing, is it able to support the offensive (reenforcement, air or bombardment TFs) ?

[decaro]

ORIGINAL: Pyrrhos1976

ORIGINAL: Joe D.
The AI is capable of forming TFs for bombardment, transport, etc., but I've never seen it coordinate or concentrate any of them together for an invasion.

[X(] But how can it succeed an invasion ??

At least, after the landing, is it able to support the offensive (reenforcement, air or bombardment TFs) ?

It doesn't always succeed, and rarely ever supports an invasion. Why do you think PBEM is so popular w/this game?

[Pyrrhos1976]

I know.
And on the defensive, is it worst ? I have the impression that AI sends TFs with CA and BB to bombard ennemy invaders without air support: in the Marianas scenario, the japanese BB (Musashi, Yamato) are destroyed in suicide mission (june 1944, it's a little too soon for this kind of mission...), while japanese CVs remain far from allied CVs and offer no air protection at all.

But maybe, in the Marianas scenario, is it partly the consequence of the overwhelming allied air power: there's no solution for the japanese player (as historically)... and if this is the AI then

[Zebedee]

The AI is just as dumb as the AIs in any other very complex strategy games. Not really a criticism, just a comment on how difficult it is to script an AI which behaves in any way rationally.

But for the first couple of grand campaigns, it can be quite good fun as you're just learning the game too.

Oh about scattering CVs - yes the AI does tend to throw in reinforcements piecemeal. So once you've broken the first Air combat TF, you'll see the AI send single CV tfs steaming up to Truk with only a cruiser and dd for company. It will bomb every island on the way there and run out of fuel half-way there. You'll find that by mid-1942 the 'big fist' of Japanese CVs can go anywhere and do anything without having to worry about Allied airpower unless a stray bomb from a B-17 flying at 18000 ft interferes.

After your first grand campaign against the AI with each side, consider implementing house rules to help even things up a bit. eg I force myself to pay political points in order to use a unit outside of its 'normal' area of operations, even if it isn't required by the rules (eg a division assigned to Southern Area has to be paid for in order to be used in Burma Area).

Other than that, there's not much to be done. Hopefully the new expansion with its moddable AI modules should add a bit more spice. But even then, I'm not expecting miracles.

[hvymtl13]

The AI gave a good showing in the Gilbert Islands, and somewhat but increasingly less so in the Marshalls. I couldn't say if the AI's actions in the Gilbert Islands was scripted or what but it was effective in those battles.
 After Makin Tarawa and Kwajelein were occupied and built up it seemd the AI did not detect the danger or make the neccesary adjustments to it's tactics to protect surface fleets and or invasion fleets. I believe it is at least a capable opponent in the Gilbert/Marshall Island Battles.  But doesn't know the meaning of the phrase -one to many trips to the well..
 
I've never played the japs but from US player perspective it seems to concentrate it's CV forces fairly well. At least into Early 43.

[jwilkerson]

ORIGINAL: Mike Scholl

In my experiance, the AI is never competant. But it's incompetancy shows up less in a smaller, shorter scenario.

I disagree.


For me the amazing thing about the WITP AI only shows up in the campaign game. When I play Allies against AI Japan and Japan can actually take the SRA in a historically reasonable timeframe.

[jwilkerson]

ORIGINAL: Joe D.

ORIGINAL: Pyrrhos1976

ORIGINAL: Joe D.
The AI is capable of forming TFs for bombardment, transport, etc., but I've never seen it coordinate or concentrate any of them together for an invasion.

[X(] But how can it succeed an invasion ??

At least, after the landing, is it able to support the offensive (reenforcement, air or bombardment TFs) ?

It doesn't always succeed, and rarely ever supports an invasion. Why do you think PBEM is so popular w/this game?

PBEM is not popular. According to the experts

Joel Billings says: "By the way, my guess is that 80% of WitP buyers play AI only"

http://www.matrixgames.com/forums/fb.asp?m=1629218

[Mike Scholl]

ORIGINAL: jwilkerson

ORIGINAL: Mike Scholl

In my experiance, the AI is never competant. But it's incompetancy shows up less in a smaller, shorter scenario.

I disagree.


For me the amazing thing about the WITP AI only shows up in the campaign game. When I play Allies against AI Japan and Japan can actually take the SRA in a historically reasonable timeframe.

I made that point earlier in the thread. But how often can you play the first 5 months of the war from the Allied side before getting bored? After that the poor AI "loses itself in the forest" and begins bumping into the trees. Not it's fault perhaps, but not much fun either.

[Mike Scholl]

ORIGINAL: jwilkerson
PBEM is not popular. According to the experts

Joel Billings says: "By the way, my guess is that 80% of WitP buyers play AI only"

http://www.matrixgames.com/forums/fb.asp?m=1629218

Could well be correct..., but what are the figures for those who continue to play WITP? While it's obvious that this forum over-represents the PBEM crowd, it's not at all obvious that they don't constitute at least 50% of those who still have the game on their systems...