Is the AI able to defend Japan?

[Rapunzel]

In september 1942 i invaded the northern Japanese Isle. 2 Marine Divs and a bunch of support units are advancing without any real resistance. There are going to be 5 more us divs on Hokkaido as reinforcements during the next weeks. The Ai seems to be nonresponding. It sill invaded Rockhampton (the ai lost there already 4 divs). Is the ai able to defend Japan or am I ruining my game now? I have the feeling, that I could conquer the main isle without any problems.

[wild_Willie2]

You just kicked the feet from under the AI...

It will not be able to manage to defend Japan before 44 or so....

If you want to play ahistorically, go PBEM...

[Stugots]

My guess would be no unless its in an ai script to defend.

Im playing ironman scenario against Japan. The ai keeps attacking me at Perth and keeps getting slaughtered before they even get off the ships. I would like to think that the ai would change its strategy but unless its in the script im not sure it will. Kind of frustrating. Maybe a dev of the ai can shed more light on your question. I know there was some other threads about the ai scripts dont know off hand what they were named.

Most will advise you to play a pbem game. Not sure i would want to make that commitment myself. I dont feel guilty if i dont send a tun to the ai for more that a couple days

[Lrfss]

Well IMHO, you would think that all scripts would at least defend each Nations Homeland rather well considering... No matter what time period...

[vettim89]

ORIGINAL: Lrfss

Well IMHO, you would think that all scripts would at least defend each Nations Homeland rather well considering... No matter what time period...

No offense but this is beyond the scope of historical reality. To be fair to the designers, teh scripts they wrote have to have some limit on them. At this point in the war the AI should not be scripted for HI defense because a human player would have murdered this invasion attempt. The AI is very strong but it cannot take into account every possiblity unless you want the turn resolution to last for several hours

That said, I agree that the AI has a hard time admitting defeat. My current game vs the AI has seen Midway assulted three times all ending badly for the IJN. I have the suspicion it will come again very soon. Its definitely the "NEVER SAY DIE AI"

[khyberbill]

I have the feeling, that I could conquer the main isle without any problems.

This is why I dont play against the computer. I may get the numbers wrong, but it took several IBM engineers and a fast computer to beat a strong human opponent in chess. In this case the program only had to look forward for the variables of 32 pieces on a board with 64 spaces. In AE, there are thousands of pieces and thousands of hexes and the computer used to determine moves, in my case, is a laptop. If you go much outside the scripts, you will probably have an easy time.

[Nikademus]

AI can do a decent job if a player plays along a more historical method in regards to early objectives. (aka, no jumping from Hawaii to Japan etc). AE has made an Aleutians/Kurilles gambit much more practical than in WitP so naturally against the AI this is easy to exploit as Player one is weak in that area at start and the AI is busy going through Andy's scripts that follow historical patterns.

[Andy Mac]

The AI should react to HI invasions in 42 = it may not react well in 42 as a lot of units are committed to other activites and they will take priority.

I have no way of telling the AI if player attacks base Y drop everything and defend its just not possible.

Keep pressing and the AI will eventually react.

p.s. ANY AI will become predictable - suprises that gave folks a good game 1st time through will become easy to defeat once you gtet the pattern especially as most folks are now in general aware of most of my suprises

[Andy Mac]

Send me a turn and I will take a look

a.mcphie@btinternet.com

[janh]

ORIGINAL: vettim89

No offense but this is beyond the scope of historical reality. To be fair to the designers, teh scripts they wrote have to have some limit on them. At this point in the war the AI should not be scripted for HI defense because a human player would have murdered this invasion attempt. The AI is very strong but it cannot take into account every possiblity unless you want the turn resolution to last for several hours

That said, I agree that the AI has a hard time admitting defeat. My current game vs the AI has seen Midway assulted three times all ending badly for the IJN. I have the suspicion it will come again very soon. Its definitely the "NEVER SAY DIE AI".

Yes, but that was a design decision to work with a scripted AI. It is also possible to create more dynamic AI's, that are based on in some sense on scripts that take dynamic variables for bases, units etc. That would have been much more powerful for such an excellent game, but would also have required more resources in terms of time and people to implement. However, this would in my opinion have taken this game to a truly new level. But maybe there is still hope and they'll in the future decide to create an "AI enhancement" addon, or WITP 3.

And to be fair to the player, this is a game and a game ought to allow you to try things that were not done or believed to be impossible at that time. Probably Japanese main islands is beyond possibility, or least extremely costly, but this would have been the game that should have allowed a player to test it out. And not just "in another minor alteration" repeat history. But due to the scripted nature, the scenario designer has to incorporate that option, which surely makes the interplay between the myriad of other scripts that Andy put into the scenarios hard to predict and engineer without having a dynamic framework. I think presently the scripted AI is the only weakness on this game, and I would so much appreciate it if that changed someday. Until then I will enjoy reading the AARs off and on. Otherwise PBEM is obviously a much more promising experience, and there the game seems to excel given you have a worthy and reliable opponent. And the time to commit to it, as you will surely make enemies quickly if you play scenarios only off and on and quit with you have tested what you wanted to try...

[Lrfss]

ORIGINAL: vettim89

ORIGINAL: Lrfss

Well IMHO, you would think that all scripts would at least defend each Nations Homeland rather well considering... No matter what time period...

No offense but this is beyond the scope of historical reality. To be fair to the designers, teh scripts they wrote have to have some limit on them. At this point in the war the AI should not be scripted for HI defense because a human player would have murdered this invasion attempt. The AI is very strong but it cannot take into account every possiblity unless you want the turn resolution to last for several hours

That said, I agree that the AI has a hard time admitting defeat. My current game vs the AI has seen Midway assulted three times all ending badly for the IJN. I have the suspicion it will come again very soon. Its definitely the "NEVER SAY DIE AI"

Well no worries here! I think the game works fine for the most part considering it's size alone... I suppose you can improve upon it forever in any case assuming you had unlimited time and money...lol! I have also seen several mulitiple attempts by the AI to take several places back which were doomed to fail though only twice not three times as of yet...lol

Well if I read Andy Mac's statement correctly, given time the AI well at some point react to this invasion though it may be a lil belated?

When you consider how long it takes to play through to '45/'46 Vs the AI and all the things that will and can change with each game not including how many scripts there are(?), there still is a very good challenge here! Hats of to the designers and such always!

Later,

Lrfss

[Andy Mac]

We did the best we could with what we had - all we promised is no worse than WITP for the AI

Andy

[Rapunzel]

ORIGINAL: Andy Mac

Send me a turn and I will take a look

a.mcphie@btinternet.com

My turn is on the way. Thx for taking a look. I will simply avoid such an early invasion the next time.

(BTW ich play pbem too, the ai game is just for the time between turns)

[castor troy]

ORIGINAL: Andy Mac

We did the best we could with what we had - all we promised is no worse than WITP for the AI

Andy

the AI is working very well Andy, compared to WITP. It was definetely one of the steps forward when comparing WIPT and AE. Doing "very strange" things like an invasion of Japan in 42 shouldn´t be a point to judge on the AI.

[Mifune]

We did the best we could with what we had - all we promised is no worse than WITP for the AI

Actually I wish to praise for a excellent work with A.I.

[Andy Mac]

OK had a look I see what you did.
 
The AI is reacting but because it hasnt yet competed all its other jobs the forces used by those scripts are held in reserve
 
I am more worried about Sumatra and Burma in that save still looking at it

[Andy Mac]

You are playing variant 9 of the AI scripts
 
The AI is commirtted to another theatre and its counterattack scripts are not set to be triggered in this timeframe (I basically assumed when I did the AI that no one would invade the HI in 42 - you did and did without going through the Aleutians so you bypassed the minefield scripts I can adjust but it will take some time if we ever do another data patch I will set up so the AI will in fact react

[Andy Mac]

Interesting as I crawl through and analyse whats going on you went all out straight for the HI in 42 with only 2 Divs and managed to pick the Japanese off base by base.
 
Interesting idea how did you stop the AI in Burma Btw ??

[Andy Mac]

ah nm got it XI Bomber Command

[Andy Mac]

That script may be to aggressive for Scen 1 as players are now further up the learing curve probably still ok for Scen 2 and Ironman but that particular script may be to aggressive