Matrix Games Forums

The AI seems so prescient I have to believe it is cheating, unless there is a mechanism within the Naval Search mission that allows search planes to spot CAP.

If I put up no CAP at PM the AI sends 3 Zeroes and 7 Betty's (typical example). If I decide my morale has finally climbed high enough to have a go at CAP and I put 32 fighters up the AI sends 100 Zeroes and 7 Betty's.

Why don't I seem to be able to catch the small group of escorts with a larger CAP force. If I put up substantial CAP, the AI simply trumps me. However, when I put up no CAP, the AI seems to know this and doesn't waste the escorts.

Has the AI been given an advantage that I won't have to worry about when playing human opponents? It would be nice to have things work as a sort of guesing game so that I at least have a small chance of success by guessing right and catching a small group of escorts with an overwhelming CAP force.

It just seems like every time I decide to put up a substantial CAP force the AI gets wind of it and strikes with one of those 100+ Zero escort raids.

What setting is the diffulty of the game set to for the AI?

It's not the AI as in the computer opponent. There is a generalised thing with the "tactical AI" which sends an appropriate number of escorts depending on the CAP, which is always known to it. Or sometimes the strike will not even go off due to a large CAP, particularly on Naval Attack. This applies equally to both sides, in whatever mode. There is even a bit in the manual about the phenomenon, though doubtless it's incomplete and inaccurate.

Where I believe the AI may cheat a bit in UV is with the effects of Airfield Attack, which always seemed way stronger than anything I could muster using the same types and numbers of aircraft. Or perhaps this is just a function of the way it continually bulks up base forces and other eng units beyond the ToE, thus providing itself with "insta repair capability".

ORIGINAL: tocaff

What setting is the diffulty of the game set to for the AI?

Hard

I think that in the hard setting one of the advantages the AI gets (cheating) is intel.  It knows things that you wouldn't.

The "TacAI" always knows what the defending CAP is. Was my last post invisible?

ORIGINAL: Captain Cruft

The "TacAI" always knows what the defending CAP is. Was my last post invisible?

No, your post wasn't invisible. I was just answering a direct question from another poster. I didn't see any questions in your post that needed a response.

Your statement that it works for both sides is only true IF the AI plans it's misiions BEFORE that info is revealed to it, just as the player has to. Granted, the players TacAI will have to ultimately decide if it wants to fly the planned mission in the face of the CAP it has become aware of after the players commits it to that mission at that location, but the player has no knowledge of the CAP at a location before assigning the mission, so the AI shouldn't either!

The ability of the AI to be aware of CAP levels is one of the things built into the game to make the AI actually have some 'I'.

Can you imagine the criticism if the AI sent unescorted raids in against 100+ CAP. The problem is that if it meets the target criteria, the AI is dumb enough to do it again the next turn (and the next turn, and the next....). This would make a very short and unsatisfying game. This issue seems to only get raised when players get frustrated when the AI invariably fails to fall for any of their 'CAP traps'.

There is some historical justification in this approach that any historical commander would make it a priority to be VERY aware of the enemy activity levels in their immediate area. There would be many methods of achieving this (not all of which are included in the game), such as monitoring radio traffic, sighting reports from various sources, etc.

I have a reference (Angels Twenty, Park E. University of Queensland Press, 1994. ISBN 0 7022 2555 X) where the author describes how his first P-39 mission was leaving Port Moresby at the crack of dawn to do a weather recon run over the target for the main raid following a couple of hours behind. He would have certainly reported any CAP encountered.

I think both Todd and Reg are correct, although it's hard -- if not impossible -- to find this info in the UV pdf.

Re "hard' setting: a WitP post explained that at this setting, the computer AI "peeks" at your deployment, giving the impression that it's behaving smarter; but in reality it isn't any brighter, it just sees more of your forces -- than you can see of its -- and acts accordingly.

Against the AI and in PBEM games, I've seen my opponent send bombers w/o any escort whenever I neglected to set CAP for its target. Conversely, my bombers will not launch against a target w/heavy CAP if I don't supply them w/sufficient escort. Of course, the escort needs to have good morale/fatigue ratings, or they won't fly either, which was the subject of another recent UV thread.

Although I have never heard of "TacAI," it may be related to the Max Detect Levels (MDL) and other DL explanations in the UV pdf, which explain how your opponent can detect your forces. In fact, a "mouse-over" of an enemy base will reveal info re the # of planes, units, etc., but that intel may not be entirely reliable.

I will say that all questions from the new UV players have proved a learning experience for this vet player as well.

ORIGINAL: Joe D.

I think both Todd and Reg are correct, although it's hard -- if not impossible -- to find this info in the UV pdf.

I tried to find this info also ( and failed ).

Re "hard' setting: a WitP post explained that at this setting, the computer AI "peeks" at your deployment, giving the impression that it's behaving smarter; but in reality it isn't any brighter, it just sees more of your forces -- than you can see of its -- and acts accordingly.

From what I can tell from the UV rules pdf, the computer AI "peeks" at your CAP levels at ALL settings, but proberly gets more concise info at harder settings.

Although I have never heard of "TacAI," it may be related to the Max Detect Levels (MDL) and other DL explanations in the UV pdf, which explain how your opponent can detect your forces. In fact, a "mouse-over" of an enemy base will reveal info re the # of planes, units, etc., but that intel may not be entirely reliable.

The "mouse over" of any enemy base before ordering an attack is useful for several reasons, but is essental for a "possible" CAP level. This is how your current Intel on any enemy base is "delivered".

ORIGINAL: DEB

ORIGINAL: Joe D.
... In fact, a "mouse-over" of an enemy base will reveal info re the # of planes, units, etc., but that intel may not be entirely reliable.

The "mouse over" of any enemy base before ordering an attack is useful for several reasons, but is essental for a "possible" CAP level. This is how your current Intel on any enemy base is "delivered".

I found that sending a recon plane before the "mouse over" gives more accurate results, but since detection levels are recomputed several times a turn, this information can get old fast; it's not so much "fog of war," but that old news is unreliable.