Yea but how is the AI?

[tiredoftryingnames]

The start of the scenario has historic landings being done by the japanese with as close as possible the units involved. These TFs are loaded and have their objectives set to give them bonuses for the first turn. After that the AI plans and conducts operations based on tons of criteria. In Burma you're going to see the same units over and over as they are stationed on the mainland. But what unit gets picked to hit an island or strengthen a base will vary. The AI splits units also to get max coverage. Three bases with weak or no opposition will get the 3 parts of a large unit and this has improved greatly over time with us sending saves in of what the AI was doing. It use to fragment units to the extreme and scatter a division all over the pacific. It's better now from the tests I've run. Not perfect mind you but it doesn't have the ability to think and isn't tied to the structured use of units that humans tend to do especially those of us that like to use units in a historical manner. But the exact unit name doesn't matter so much when you're fighting the AI most of the time. You just need to know how many men, guns and vehicles. But if it'll make you feel better in AI tests i've seen the Allied AI send units to different islands in different games.

[Joel Billings]

ORIGINAL: ZOOMIE1980

But it's the Big Picture stuff, the Strategic side of the AI that I'm afraid is going to be disappointing, at least after the first run or two through a scenario. It looks like I'm going to know exactly where, when and how the AI is going to attack me after a two runs through any scenario, right down to the exact same composition of ship types in the TF's.....

Should I admit I was laughing out loud when I read this post? Ok, I was. If you play each scenario 2 times through for each side, by my calcuations you will probably have played for about 8000 hours. At 40 hours a week, that's almost 4 years. Is that enough gameplay for $70? If you take the game seriously the first time you play the main campaign, you won't know what the AI is going to do. Should be a fun first year of playing WitP. Sorry, I just couldn't resist.

[freeboy]

Joel,
thas continues to look GREAT... thanks again.. and remember for each posting member there probably hundreds of former board game war in the pacific players who will love this monster er I mean user friendly game[:D]

[denisonh]

SO, are you telling me that War in the Pacific is massive in scope [:D]

ORIGINAL: Joel Billings

ORIGINAL: ZOOMIE1980

But it's the Big Picture stuff, the Strategic side of the AI that I'm afraid is going to be disappointing, at least after the first run or two through a scenario. It looks like I'm going to know exactly where, when and how the AI is going to attack me after a two runs through any scenario, right down to the exact same composition of ship types in the TF's.....

Should I admit I was laughing out loud when I read this post? Ok, I was. If you play each scenario 2 times through for each side, by my calcuations you will probably have played for about 8000 hours. At 40 hours a week, that's almost 4 years. Is that enough gameplay for $70? If you take the game seriously the first time you play the main campaign, you won't know what the AI is going to do. Should be a fun first year of playing WitP. Sorry, I just couldn't resist.

[ZOOMIE1980]

ORIGINAL: tiredoftryingnames

The start of the scenario has historic landings being done by the japanese with as close as possible the units involved. These TFs are loaded and have their objectives set to give them bonuses for the first turn. After that the AI plans and conducts operations based on tons of criteria. In Burma you're going to see the same units over and over as they are stationed on the mainland. But what unit gets picked to hit an island or strengthen a base will vary. The AI splits units also to get max coverage. Three bases with weak or no opposition will get the 3 parts of a large unit and this has improved greatly over time with us sending saves in of what the AI was doing. It use to fragment units to the extreme and scatter a division all over the pacific. It's better now from the tests I've run. Not perfect mind you but it doesn't have the ability to think and isn't tied to the structured use of units that humans tend to do especially those of us that like to use units in a historical manner. But the exact unit name doesn't matter so much when you're fighting the AI most of the time. You just need to know how many men, guns and vehicles. But if it'll make you feel better in AI tests i've seen the Allied AI send units to different islands in different games.

Well that's encouraging news based on past reports on here of how the strategic part of the AI is going to function.

What's your recollection of how the AI (in this case the Japanese during conquest phase, but in general as well), say deploys his carriers. In UV you always knew almost EXACTLY what the composition of an enemy carrier TF was going to be and it never altered much, CV, CV and most of the time maybe CVL. IT did that no matter how many CV's it had available or valuable the target. I'd just hit him with a four CV TF every time and wipe him out! I degree of randomness, based on CV availablity, value of the target, etc.... would be nice. Maybe one encounter is with a CV, CV TF and the next a CV, CV, CV, CV, CVL "death star". Another thing that I found boring and tedious in UV was while playing Japan, the US AI consistently sent a two CA or so bombardment TF from Nomeu to hit Lunga about once every other week. All I had to do was sit a two CV TF a couple of hexes from Lunga on the right day of the week and nail them each time they came up. Stuff like that really ticks me off in wargames....the damned AI's shouldn't be THAT predictable.

Some degree of rigidness is expected and is reasonable. Even the human player, playing Japan is pretty much going to attack, all out, in the same places with roughly the same relative strengths of force because the strategic goals pretty much require it.

But another question, is what happens after Japan has finished the main conquest phase and established the defensive perimeter? Is there some randomness in where it may choose to "go on an adventure"? Attack Midway, attack Fiji or American Somoa, land in Oz, go on a major assault in China to try and seriously knock China out of the game by borrowing units from the PI or whatever, or not adventure anywhere and elect to play pure defense? Or is that pretty rigiddly scripted as well?

I guess the basic question is just how rigid is the scripting for each side? Are there multiple "possible branches" in the strategic scripts where one or the other may be chosen based on either an entirely random choice or some sort of weighted choice based on the current situation, or does it follow a single script with only small deviations on force composition and maybe a small degree of randomness in TF composition, exact timing, etc...??

[ZOOMIE1980]

ORIGINAL: Joel Billings

ORIGINAL: ZOOMIE1980

But it's the Big Picture stuff, the Strategic side of the AI that I'm afraid is going to be disappointing, at least after the first run or two through a scenario. It looks like I'm going to know exactly where, when and how the AI is going to attack me after a two runs through any scenario, right down to the exact same composition of ship types in the TF's.....

Should I admit I was laughing out loud when I read this post? Ok, I was. If you play each scenario 2 times through for each side, by my calcuations you will probably have played for about 8000 hours. At 40 hours a week, that's almost 4 years. Is that enough gameplay for $70? If you take the game seriously the first time you play the main campaign, you won't know what the AI is going to do. Should be a fun first year of playing WitP. Sorry, I just couldn't resist.

Oh, I agree, wholeheartedly. The sheer massiveness of the main scenario may make the "strategic predictablility" of the AI moot. Who cares if it takes you a year to play one campaign??

But many will play the first four-to six months of the major campaign or some the small campaigns, and there, it DOES become a major issue. I hope it is not as predictable as the PATHETIC UV AI....

[Joel Billings]

ORIGINAL: ZOOMIE1980
I hope it is not as predictable as the PATHETIC UV AI....

Me too...[:)]

[ZOOMIE1980]

Well based on what I've seen in this thread, it looks promising!

[Ron Saueracker]

Hate to admit it but I tried to get cute and the AIis spanking me for my attempt. Left the ships in Manila too long waiting for enough P points to evac the Marines and along came a spider...two actually. Ryujo and some other scow's Kates are having a field day. Oooops.[:)]

[tondern]

ORIGINAL: Ron Saueracker

Hate to admit it but I tried to get cute and the AIis spanking me for my attempt. Left the ships in Manila too long waiting for enough P points to evac the Marines and along came a spider...two actually. Ryujo and some other scow's Kates are having a field day. Oooops.[:)]

Sooo. If you used them as bait and just happaened to have a CV group near by . . .

My beef with the AI is not so much its lack of agressiveness, or even lack of variability, but in its inability to avoid repetitive pointless losses of ships and planes - reference Gnome's post. In a long war this really kills the side doing it. But it does make it more fun to play the Japanese.

Tondern

[kev_uk]

Its all very well having a detailed game, but if the AI is poor then the whole point is moot. I dont fancy playing a PBEM game of the full scenario which would take, well, a very long time....ok well i would play it, but am really looking for a good solo game as well.

This will be the crunch whether I pay out nearly $70 or not.

If its like UV AI then probably not.

Cheers

[Von Rom]

Will any of the testers be playing a game and doing an AAR against the AI soon?

[Mr.Frag]

Will any of the testers be playing a game and doing an AAR against the AI soon?

Which side do you want me to loose at? [:D]

[kev_uk]

play as US, full game. PLEASE!!

Cheers

Kevin

[Mike Scholl]

ORIGINAL: Mr.Frag
Which side do you want me to loose at? [:D]

How about you taking the Allies and giving us a look at what the AI is capable of in
either a December 7th start, or a May, 1942 one? The second might even be pre-
ferable, as the "coded" initial expansion is finished and the AI has to make some
actual Strategic decisions while it still has the "pre-Midway" edge?

[Huskalator]

How about a South Pacific Scenario May '42 similar to the one in UV

That would give us an excellent way to compare and contrast the AI in UV and the AI in WitP.

[Mike Scholl]

More or less what I was talking about as the second choice..., but I'd like to see it as
a "full map" varient to get an idea of the AI's strategic grasp. Will it use more of the
fleet to support the Port Moresby Effort? Will it chose a different line of attack entirely?

[Damien Thorn]

I'm interested in seeing how the AI is on offense. Usually most game AI's are OK on defense. It is when it tries to attack that it usually gets its head handed to it. How is the AI as the Allies in 1944 or 1945? I assume it is ok as the Japanese in 1941-42, since a lot of that could be pre-scripted but bu 1944 (playing from 1941) the setup would be totally unpredictable so it would be interesting to see how the AI can handle that.

[ColFrost]

ORIGINAL: kev_uk

Its all very well having a detailed game, but if the AI is poor then the whole point is moot. I dont fancy playing a PBEM game of the full scenario which would take, well, a very long time....ok well i would play it, but am really looking for a good solo game as well.

This will be the crunch whether I pay out nearly $70 or not.

If its like UV AI then probably not.

Cheers

Boy, you guys are harsh. I must be a moron, because I have yet to beat the AI as Japan yet in UV.

[Huskalator]

ORIGINAL: ColFrost

Boy, you guys are harsh. I must be a moron, because I have yet to beat the AI as Japan yet in UV.

Build up Guadalcanal as fast as you can, stock it full of Bettys, keep a large surface force there, and keep your CVs at Shortland. Easy victory, and fun for a while until the Allies run out of ships.

The allies will always send its forces straight at Guadalcanal. The one-two punch of the CVs and the superbase takes care of any Allied CVs. The surface force doesn't even need to win a battle, just damage the bombardment TFs so the Bettys can tear things up. The AI will send attack after attack, mindlessly, until it runs out of ships.