Does Allied AI think outside the box????

[Macclan5]

ORIGINAL: dr.hal
.

But Manila was certainly a valid target and before Yamato's plan, WAS the target. Given the number of subs and support ships there, I think it was still a challenger to the Pearl plan. Here too historical knowledge impacts the play as we KNOW the US's modern torpedoes are "cr#p" and thus the subs less of a threat than was known to the Japanese at the time. But if one thinks in the "long" view of history, the subs DO become a real threat later in the war.... so it's good to take them out NOW!!! But to my point, I don't want to screw up the AI's play too much by using "thinking outside the box"!!!

In fact I think I totally agree.

In the long view Japan not attacking Pearl... dominating other areas such as the PI... even encouraging confusion and a 'hastily planned response' before the USA had time to learn the lessons of Coral Sea, Midway et al is a great start to a game and what if.

I would even argue that the USN "would have" sent an under-supported response fleet from Pearl to such Japanese actions. Under-supported in terms of Air Coverage and Carrier coverage because the value of the CV task force was largely unknown by the USN at the time.

Here you have the historical basis of what if. The Kantai Kessen will happen more likely in 1942 as opposed to Dec 7 if the USN - AI sends "enough" in response to you..possibly you may have to force this issue by proximity to Hawaii for example (??!)

The trick is ... as you indicate... understanding that the AI will not deviate from priority of linear progression after that.

The AI will learn its hard knocks lesson... and then likely proceed to invade and secure its SLOC. Pago Pago / Suva / Noumea / Guadalcanal / etc (??)

I have never played the AI as Japan. I assume (with some randomness) it will pursue a semi historical corridor of advance

Its more of a question of you repeatedly delivering a Kantai Kessen. [8D]

Your enjoyment will then be dictated by how much you would enjoy that.

[Macclan5]

ORIGINAL: Chickenboy

The AI is what it is. It's designed to be crushed without mercy. In my opinion, NOT doing this is not playing to the capabilities of the engine or oneself.

Which is most fair to say. [&o]

I fully understand your Point of View and frankly I think I will be intrigued enough especially once I set out to play Japan and learn the Empire OOB, and industry, the challenges.

My own point is exactly that your approach is great, fun and should be all you want it to be. [8D]

There is no wrong or right in this debate I think.

I tend to be of the historical bent towards this game. What fascinates me personally is the OOB - deployment - the tools that the actual generals worked with AND the intangible considerations / restraints they had in the face of uncertainty.

I can at best only approximate the last. House rules, limited re-load restrictions, awaiting the correct tools to accomplish what they did in history. I hope to use the same tools more effectively as do all players.

I tend to measure this historical considerations both in terms of score (Victory Points) but also in terms of minimizing certain scores such as LCU losses, airframe losses, ship losses.

--

Imperfect AI ?

I prefer to say linear and at least tethered to those intangible restraints of history.... it may be the same thing.

I just don't considered it to be broken and given the complexity of all the variable I think it is quite amazing given its age.

I cannot wait until such time as I am able to import "IBM Big Blue" app - share ware into the engine; once it learned through playing a few dozens of games from all of you LOL ....

Even if the game engine is old some day soon such an AI app will indeed be sufficiently modular to

[rustysi]

Personally I look at it this way. The devs got it right one more time. AI is AI to me and in this day and age we can seek and find opponents anywhere in the world. So the time, money, and effort went into the game itself and did they ever get it right. I don't mean to detract from those whose efforts went into the AI, but I understand there's only so much one can do with it. Given all the different openings and later moves a player may undertake we could program the AI 'til the cows come home and never cover everything. So as I say, to me its just a learning tool to play out my 'what ifs'. To defeat it is no real deal, it is what it is.

[Ian R]

ORIGINAL: Alfred

1.  The AI does not ever teleport task forces.

2.  The AI will generally not teleport individual anchored ships but will do so under a specific condition.


The script will set out certain circumstances under which a specified type of task force will be deployed by the AI.  The task force will be populated from anchored ships in the "trigger" port.  However, if the requisite type of ships are not available in the "trigger" port, the AI will teleport anchored ships to meet the script requirements.


Alfred 

Yes.

If you watch the message box while the programmed opponent is organising its turn, you may see a message along the lines of "Move ships to active locations" or something similar.

However, I think it also possibly moves damaged ships sometimes. There are reports of heavily damaged Repulse/POW being in Singers on day x, and Colombo on day x+1.

[Ian R]

ORIGINAL: dr.hal

the subs DO become a real threat later in the war....

And the AI ignores withdrawals.

I don't have a problem with the AI taking shortcuts. I don't even think of it as cheating.

Like Gräfin I am playing ironman nasty. As the allies - its December 1943 and I only recently cleared the Aleutians, and took Midway, Milne Bay and Lunga back. I won't spoiler things by recounting what else I had to take back before I could even do that. The Marines are now prepping for the Gilberts, but it'll be another 6 months of training my new CV air groups before I'm going there. [redacted] Island lagoon is going to be the wreck diving capital of the world. I've got US carrier battle groups hanging out at backwoods places like [redacted], because the multiple IJ carrier raiding groups are like the Borg - they seem to appear out of a subspace conduit, 'assimilate' everything in their path, and then slip past you and away to replenish - and they don't do dumb things like sailing up to a group of level 7/8/9 allied airbases and remaining on station for too long.

Fortunately, my mod of this goes to 1948, and I put in all the ships from the suspended/cancelled USN building programs (although I left out the last 6/32 Essex from the 44 program). I'm going to need them to penetrate the inner defence ring, but that can wait until the Soviets have liberated Manchuria and China.

Its a very challenging experience.

[PaxMondo]

Anyone who has been here in these forums for over a year knows, or has access to know, all about the AI. Thus there are really only two options:

CB's "fowl" play style of punishing it
or
GZ's playing a steroid version of the game.

Both work. Both are effective. Both are fun.

Me?

I play a version of Andy's Nasty, Nasty release updated by using recombinant DNA, gamma rays, and of course the old standby: toxic dip. [;)]

Anyway, it is fun. It provides a great, albeit different, competition as compared to a PBEM opponent. One of these years, likely after my youngest has completed his education, I will even get past '43 before life intervenes for so long that I cannot recover my thought chain and have to start over. [:@][:@]

[:D][:D][:D]

[PaxMondo]

As for thinking "Out side of the box". That is up to Andy. So, yes, some of the AI variants, yeah, they have some real differences that you won't expect. And the AI system is so robust that you can create almost anything given time to code it up and a group of testers.

The biggest issue for anyone who is NOT a dev are the scripts that fire anytime after about 1 Feb 42. Why? Testing. non-devs have no easy and accurate way to test those without spending a lot of time to create new starts in the editor. Anyone who has attempted that knows the amount of time that takes. And the later in the game you go, the more time it takes to create one.