AI in the game

[Ekaton]

I don't have that much time now and probably won't be able to play a human opponent. I wonder, (1) how good is the AI in this game? Is it likely to be doing something reasonable, or will it tend to do something ridiculous, like assault islands with no strategic value or fail to supply units properly?

(2) Is historic AI setting going to result in a somewhat historically accurate gameplay and recreation of Japanese strategy? How does it differ from other settings?

(3) Will mods such as DaBigBabes going to break the AI? Will it be able to work properly despite using it?

[PaxMondo]

AI works with most, but not all mods.

Stock mods - the AI will give beginners a reasonable game on VERY HARD for about 6 months.

Ironman mods - these are designed for AI play and give the best AI game. Playing these on VERY HARD you can get well into '44 or later if you do not 'game' the AI. The AI is NOT perfect, but apart from a few flaws plays pretty well. The Nasty version is a challenge, the Nasty, Nasty version is sadist. [;)]

[Yaab]

Ekaton, read this thread about updating AI and AI historical variants.
tm.asp?m=4257473&mpage=1&key=&#4257473

[Lecivius]

Also, look to Andy's improvements. I have found his latest AI update slightly devious [8D]

[m10bob]

I have used the AI for years, and the length of the campaigns themselves help prevent you from memorizing what the enemy will do, to some extent.

In preferences, I allow the AI advantages and time allowances as well.

One of the stock scenarios was intended especially for AI, the second I I believe?

I suggest each time you play, play using totally different strategies as well.

[Zecke]

eseee m10Bob, long time with no experience wisdom.

I play always HISTORICAL...dont know yet admiral, but i figure out, that the AI from WITP is into Admiral; so is a female; definetly

[urtel]

Let me share mine 2 grand campaign against AI experiences it may help you.

1st game I play as Japan 2nd scenario vs Allied AI all time on historical difficulty:
- I lost game from soviets in august 45, I mean they breach mine defense on 2 from 6 directions and I have nothing more to put on they path(0 divisions everything was used somewhere else)
- I stop US in Marshals, pretty much sunk all of they combat fleet and most CV/CVEs, submarines was still make some trouble but it was bearable
- I stop British in Burma and Chinese are confident in inner china
- 4E are wiping mine frontline bases but they was out of reach of industry
It was nice game for me, I was total noob and most of time I use to manage economy, plane research etc, but I did not use this modern all R&D factories size 30 etc tricks, so I mange to push Frank line for few months and that is all,
This give me nice air war and headache with training pilots, I also do just limited expand of FP groups so mine IJN pilot pool was on breaking point for most of time. So AI was good for me, there was some nice moves by US combat units but AI have tendency to desire some Island and he will send all of it's fleet in not really organized way to that same target. In those cases I will let him to win if island is not in what I plan to defend (Canton Isl. on example).
But I did fight for Marshals and it was some funny moments when I though I will lose it,s battle take around 1 year in 44'.
So if u fell like noob I thin this is best experience you can have, also check is there AndyMac updates for AI for scenario you want to play.

I plan in future to play as Japan against AI I will like to play DBB-C mod if it work with Allied AI. I like scenario 2 because I like to produce more late war destroyers(and I do not see rationale if Japan economy is in good shape why they will not do it) and few more IJN fighter groups make big difference(this can be replaced by resizing land IJN fighter groups using carriers but I do not like that). On other side now I m not noob so I will put AI on hard, and maybe 2 days a month to very hard.
I also plan to invest more in R&D and get some 45 late planes into game, not 2 years early but up to 6 months will be mine goal. If I stop/hold soviets until 31.12.1945 I will call that full win ..I expect it will be funny game just to find time to play it..

2nd game was me Allied vs Japan AI, DBB-C mod limited cargo capacity, on historical with every month I put difficulty on hard for 2 days:
- in this game I land on Okinawa in late 44, take Burma and merge with Chinese all way to coast. So I physically separate Japan from SRA
- in this game AI stop active defending somewhere around middle 44', this mean no fleet action of any importance from AI side
- I did sunk a lot tankers with subs and everything in SRA beside Palembang was in rage of mine 4Es so I guess AI was out of fuel

Now u ask why I not ramp up difficulty when I see AI is breaking, to be honest I become tired by re-activate of bases I was long time ago pacify and cut off ( like Rabaul and Truk). Every time u ramp up difficulty level those bases will magically get supplies and then ur back lines will be attacked by Nells and Betties, I will personally like to give AI gazillion of fuel and more troops and planes and ships but on Home Islands so AI need to move them to front and I have chance to cut those lines, because that is how I perceive fighting in ww2, cutting supply lines and everything else is secondary to it.
So this game was not fun to me like 1st one I will like to have option when Japan can get xxx amount of everything(including ships) in home islands but AI need to find way to move all that to front line, I hate AI get supplies to isolated bases/troops.
If I will again play agains Japan AI I will probably play hard/very hard difficulty, but i will also plan to take every big base on main advance path or path will planned to be out of Betty range from bases I "isolate".
In this game AI focus on Geralton in Australia and after I sunk couple of bad organized amphibious attempts I just let AI to take base then some time later I retake it.
Also for some time AI send some medium size empty tankers to island which do not have oil ( one of those east from Timor) for which I fail to find rationale, beside those 2 things AI was behave reasonable for most of time.

There is things which kill AI fast, you need to watch your self not to those those. On example if u have small ships in Port Moresby(AMCs) AI will lose all his good pilots and gazillion of planes(but planes will be replaced if u play on high difficulty level) from Rabaul in attempt to torpedo that few AMCs,
so u need to watch not to do that if you want to have good fight when time come.

[BBfanboy]

urtel, I think that should be Port Moresby in your last paragraph. It is near Rabaul while Port Arthur is way up in the Yellow Sea. [:)]

[Yaab]

Port Moresby was renamed to Port MacArthur obviously.

[urtel]

ORIGINAL: BBfanboy

urtel, I think that should be Port Moresby in your last paragraph. It is near Rabaul while Port Arthur is way up in the Yellow Sea. [:)]

Yea, thx I fix it, it was long post from head...

[Zecke]

URTEL: some things of your speach i suffer in my games against the AI

BUT if you play the original witp (2004)(I have it) most of yours complains will be resolve.

THE WIPT 2004 (original) is totally another version from the new ones.

For example WITP2004 the IA ships can run out of fuel.

[Ian R]

Now u ask why I not ramp up difficulty when I see AI is breaking, to be honest I become tired by re-activate of bases I was long time ago pacify and cut off ( like Rabaul and Truk). Every time u ramp up difficulty level those bases will magically get supplies and then ur back lines will be attacked by Nells and Betties, I will personally like to give AI gazillion of fuel and more troops and planes and ships but on Home Islands so AI need to move them to front and I have chance to cut those lines, because that is how I perceive fighting in ww2, cutting supply lines and everything else is secondary to it.

This is why I prefer ironman nasty nasty at 'normal' difficulty level.

Andymac has put in things like extra oil produced in Tokyo, so it is more challenging, but you can also take the historical path of isolating bases and letting them wither on the vine.

[Rafid]

I agree with Ian. I vastly prefer to play an ironman scenario at "normal" difficulty level (I set to hard a few days a month) over scen 1/2 on hard or very hard.

a) The AI really needs the extra ships it gets in ironman
b) I think the visible logistic helpers in ironman (additional resource and light industry in many places) work just as well as the invisible bonus of the hard AI in helping with the logistic burden. However the invisible bonus can't be destroyed and hence disables strategic approaches such as cutting places of (as urtel pointed out).
c) The industry bonuses of the hard AI likewise disable many viable strategies. Cutting of the oil in 43-44 against a hard Jap AI: Pointless, the bonus makes the Home Islands close to self-sufficient and the unloading bonus means the stockpile is already sky high and will last till 1960. Same with HI points, so the only kind of strategic bombing that works is airframe factories.

To go back to Ekaton’s second original question: The AI’s strategy is based on scripts and these are close to (but not exactly) historical behavior. Hence the AI will be somewhat historical regardless of difficulty setting. In the end “Historical” difficulty setting should just read “Normal”.

Bottom line: With the right scenario/difficulty setting you can get a lot of entertainment from the AI.

[Yaab]

To play against the AI is glorious.

[PaxMondo]

The reason to play the AI at higher than normal is that it does not do a good job of supply movement, particularly to islands. Even though Ironman adds supply generation to a number of places, it isn't everywhere. So, even in Ironman, if you play at normal or lower, you need to realize that islands on the AI side will run out of supply with all of the detriments that entails.

As for supply 'magicing' in at high difficulty levels, as Leningrad and Moscow proved conclusively, there is no such thing as perfect containment around a large city. Some supply will get through. This is how I rationalize the result. Every player plays his game, but just wanted to be sure that you understand the basis behind the recommendation to run the difficulty up to at least HARD periodically for the AI.

[urtel]

ORIGINAL: Ian R

Now u ask why I not ramp up difficulty when I see AI is breaking, to be honest I become tired by re-activate of bases I was long time ago pacify and cut off ( like Rabaul and Truk). Every time u ramp up difficulty level those bases will magically get supplies and then ur back lines will be attacked by Nells and Betties, I will personally like to give AI gazillion of fuel and more troops and planes and ships but on Home Islands so AI need to move them to front and I have chance to cut those lines, because that is how I perceive fighting in ww2, cutting supply lines and everything else is secondary to it.

This is why I prefer ironman nasty nasty at 'normal' difficulty level.

Andymac has put in things like extra oil produced in Tokyo, so it is more challenging, but you can also take the historical path of isolating bases and letting them wither on the vine.

Damn I must try that, nasty on normal(historic) vs Japan AI, that will be nice game I just need to find time...

[adarbrauner]

ORIGINAL: PaxMondo



As for supply 'magicing' in at high difficulty levels, as Leningrad and Moscow proved conclusively, there is no such thing as perfect containment around a large city. Some supply will get through. This is how I rationalize the result.

You may have meant Stalingrad rather than Moscow?

[PaxMondo]

The Germans attempted encirclement, never succeeded, and never cut supply at Moscow.

What I meant to point out was that in the modern era, surrounding a metro city and attempting to starve the army within has not worked IF the indigenous population has not fled. Conceptually, it seems easy. In practice with real people involved it is rather more difficult to actually effect. Think it through and you should see why fairly quickly.