Matrix Games Forums

Last night Jamie and I had a chance to play our PBEM game in pseudo HTH mode as we meet each friday night for a team mega board game and we processed our UV PBEM turns while our teammates were doing their part in the boardgame.

I had a single ship AP TF set on continuous supply from Rabaul to Lunga (CS: Lunga). I had been sending troops and supplies to Buna and Dobadura via fast transports and under no circumstances had I ever run a normal transport TF naywhere near those two locations.

After loading my turn I noticed a transport TF about 4 hexes east of Dobadura. Examination revealed that it was teh CS:Lunga TF with a destination set for BUNA. Immediately I thought WTF!

We are near the end of the first week of scenario #16 and niether of us has any carriers that could cause a TF to react wildly and set a new destination for itself as a result of running from carriers.

Assuming it was a program glitch with autocontrol I changed the setting from CS:Lunga to "player control" and manually reset the destination for Lunga hoping it would get out of LBA range before it got pounced on.

What do you think happened in the next resolution phase? The damned task force defied my new orders and headed for Buna anyawy ending up in the Dobadura hex come morning. Of course every plane in PM came out to pummel the AP and it promptly sank giving valuable dive bombing experience to the American A24s.

Does anyone have any idea what happened? I can understand if I had left it on autocontrol and there was a glitch with the auto control that it might do what it wanted to do instead of what I wanted it to do. However, since I took it off of auto control and issued completely new orders I cannot fathom why my AP commander chose to kamikaze his ship.

I thank you for your donation to Davey Jones's locker.

Also, we meet on Thursday nights, not Fridays.

Must have been a bad night for the Axis....

Jamie

ORIGINAL: Bigdog128

I thank you for your donation to Davey Jones's locker.

Also, we meet on Thursday nights, not Fridays.

Must have been a bad night for the Axis....

Jamie

Where did I say anything about Friday night? I said LAST night and I posted on Friday morning, which in most people's book would indicate Thursday night.

DO NOT TRUST THE COMPUTER TO CONTROL ANYTHING!  Of course if you'd like to help the enemy.....

ORIGINAL: tocaff

DO NOT TRUST THE COMPUTER TO CONTROL ANYTHING!  Of course if you'd like to help the enemy.....

OK Now. Exactly what part of "I took it off of computer control, issued it new orders and it STILL WENT WHERE IT DARN WELL PLEASED AND IGNORED MY ORDERS" didn't you read?

I can accept that with computer control it might go haywire......what is perplexing and "maddening" is that it did it AFTER I retook control.

Hans, I just had something simular happen. I think, if you look at Buna, it will be very low on supplies. I think the computer will take over a CS taskforce if it thinks it needs to send supplies somewhere. Remember that you have put the Tf on computer control and it is doing what it thinks is best. As far as continueing, it is pretty well known that some of the data fields do not get cleared properly. The best thing to do in that situation is to create a new TF ( with a new number ) and give it the orders you want.

I hear you, I play with CS on a regular basis in my games to try and get a handle on it and it seems the AI will divert it to a base running low on supply every time I play. Best thing to do if your using CS keep an eye on it every turn.

Both of you seem to be on target.

What must be happening is the computer is taking a TF set to Continuous Supply (CS) and reverting it to Routine Convoy while leaving the display reading CS.


Which means that the computer can't be trusted and my illustrious opponent Todd, who is shouting from the rooftops not to ever let the computer control anything, should indeed be listened to.

Well I'm a little horse now from yelling and I rarely have anything to say of value, but..............the AI should never be trusted with anything unless you want headaches.

In my just finished game w/IKE the allied CS/anywhere worked fine.  All the ak's would continue to run the same route over and over.  Must be different w/ the japs.

Nope, did the same thing to me playing as the Allies. Basically, I don't use CS anymore if there are any occupied bases along the convoy route that are getting low on supplies (the AI doesn't seem to mind if an uninhabited base is low on supplies, fortunately). It works fine on a route from, say Townsville to Rabaul, but I have seen convoys careen towards other Allied bases if they are occupied and low on supply.

I'll agree with anyone that when the AI does these strange things it is really, really annoying. That said, the AI represents a massive and complex multi-layered command system for BOTH sides, and giving any kind of order should not be taken as a guarantee that it will occur.

The true test of anyones ablities is not just the "perfect" plan, or out thinking the "enemy; but how you cope with the errors made by your own staff, officers & men along the way. Tocaff, is partly right, in that it is best not to trust the AI with any part of your command structure; but the functions that you can leave to it are those that usually cause the least problems, generally speaking. Of course, sometimes, even these problems will prove to be major. Thats life.

The big problem for us as players, is understanding if the AI action is reasonable in the situation, even if it's not what we ( personally ), would have done OR if it is a game BUG ( error ).

Past experience with this game ( via players comments ), leads me to say that what ever we consider to be the answer here, the truth is we do not know ( for sure ).

The only ( apparent ) certainty, is we are unlikly to get any response from those who should know if it is or is not a bug. And as for a patch...

The other definition for AI is ALL IMPORTANT. This is the engine that powers your carrier and land-based airstrikes, determines whether to send a single, overwhelming force of aircraft towards an insignificant (or already-sunk) target or send a bunch of tiny formations (3 planes at a time) against a fighter-protected target where they get chopped to pieces. Then there's the issue of the AI playing favorites. Ever notice how often that, when playing the AI versus a human player, the AI side gets better cohesion on airstrikes, with fighters sticking to the bombers like glue or, conversely, putting up an impenetrable CAP over a TF or base? The key to the game is figuring out how to a) do the same yourself and b) short circuit the AI's ability to do this. There are ways that work devastatingly well, but certainly are not historically accurate. Never confuse this simulation with reality - it's a game, and when you start thinking of it as such and how you can beat the AI (whether as an opponent or as the manipulator of your automated actions like carrier airstrikes), then you might have a change of winning. And the techniques you learn in circumventing the AI's flaws will win you games against human opponents as well, because they, too, are at the mercy of this ALL IMPORTANT game engine. As Joel Billings, one of the game's designers, told me, they purposely put quirks into the AI to reflect such things as the faulty catapult on a single Jap cruiser at Midway that enabled the US CVs to get in a devastating airstrike while remaining undetected. When you think of the AI, think of a 5-year-old with a gun and you'll have some grasp of how to treat this capricious dungeon master.

It's true that a lot of what can happen in ALL computer games can be put down to an element of AI "cheating". This particular game has several different AI's built in and they are all making choices that YOU would not make if you had the chance. Either option is a possibility for the problems encountered here. Instead of "blaming" the AI, consider what other logical reason/s may have caused the problem. If you cannot think of any you may ( or may not ) be right in saying the AI is "cheating". If you can think of one, you have ( maybe ), an answer that says otherwise.

Just a different perspective.