Pearl Harbor and AI

[bradfordkay]

" There isn't a fuel issue because of the replenishment TF near KB."
 
IRL this replenishment TF was emptied in the days before the PH attack. In stock WITP the KB starts out fully fueled and the replenishment TF is full as well. Is this still the case in AE?

[Splinterhead]

There's also the minor fact that KB used all of the available 800kg and, more importantly, all of the modified torpedoes in the 1st wave. The AI can continue to attack with weapons that did not exist and that's gamey.


Edit: If you don't want random, even impossible results, play JWE's Dec 8 scenario.

[pad152]

ORIGINAL: NightFlyer

I don't think it's gamey, in real life massive US fleets lingered for days off the Marianas, Leyte and Okinawa. Even when an invasion wasn't involved this occurred for example the US fleet pounding Truk lagoon for two days on February 17-18, 1944. Just because Nagumo was too dumb to take advantage of an opportunity that never came again doesn't mean the AI has to be dumbed down as well. Just my opinion and the US would have eventually won anyways....

You can't compare the US Fleet in 45 with the Japanese in 41. The Japanese Navy didn't have anything like the support/supplies/spare parts the USN had. Even day's super carriers can only carry enough supply/AvGas for about 3 days of combat ops then it has to go off station, replenish then return. There is no way Japanese carriers could conduct full combat ops for 3 or 4 days. Nagumo wasn't dumb, he was just cautious, he planed on a second raid but, the carriers weren't at Pearl as expected, He didn't know where the US carriers were, remember no one had every done what Nagumo did. If the Japanese stuck around for just a day, one or more of the US carriers would have been in range, then who knows what would have happened.

[Andy Mac]

OK Guys its quite simple there is a chance the AI will linger why did I do this its quite simple its what I would do if I was PBEM against you as Japan when designing the AI we tried to do it from the perspective of what would a PBEM pl;ayer do on the grounds that a harder nastier AI would give you a more difficult challenge.
 
I appreciate some people dont like it (although I am suprised its went down quite as badly as it seems to have) and it does make me wonder how much youy will like my other suprises [&:][&:] - I do understand but what am I to do - make the Ai play as hard as I can or make it exactly mirror history - most PBEM players would linger the fact that the AI only does sosome of the time seems to be going down badly.
 
I am really not sure what folks want but we did provide a scen starting on the 8th for those that want a `100% historic start.
 
 

[Erik Rutins]

It also only has a significant chance to linger if you choose non-Historical Start, so this is telling me that a lot of folks are choosing the non-Historical Start, where its chance to linger goes way up.

[Titanwarrior89]

Ive chosen the Historical start. There hitting pearl 7th-10th now. Trying to get every ship out but about 40 percent or more are Gone!

ORIGINAL: Erik Rutins

It also only has a significant chance to linger if you choose non-Historical Start, so this is telling me that a lot of folks are choosing the non-Historical Start, where its chance to linger goes way up.

[BeastieDog]

I love this challenging non-historical AI. I hope we see many single player mods with varying AI responses( Manila KB opening, Hawaii invasion etc). Thanks for the good work Andy!

[Andy Mac]

The AI always has a chance to linger for day 2 and a very small chance to linger for day 3

[AttuWatcher]

ORIGINAL: Andy Mac

OK Guys its quite simple there is a chance the AI will linger why did I do this its quite simple its what I would do if I was PBEM against you as Japan when designing the AI we tried to do it from the perspective of what would a PBEM pl;ayer do on the grounds that a harder nastier AI would give you a more difficult challenge.

I appreciate some people dont like it (although I am suprised its went down quite as badly as it seems to have) and it does make me wonder how much youy will like my other suprises [&:][&:] - I do understand but what am I to do - make the Ai play as hard as I can or make it exactly mirror history - most PBEM players would linger the fact that the AI only does sosome of the time seems to be going down badly.

I understand your logic behind this. People will always complain about something when it comes to AI because everyone has some predetermined idea how it's supposed to act.

For one person to improve this AI, make it act human, and keep within myriad historical scrutinies is challenging to say the least. It's not like we are playing Warcraft here where all the AI has to do is give you a challenge.

The balance between historical and challenging requires some trade offs.

[Andy Mac]

I just dont know what you want - a good and challenging game v the AI or a repeat of History which you will all rip apart - you can have either (although the latter would need about 4 months of solid effort on the scripts)
 
If its a total repeat of history then I guarantee you will see things like the unconquerable Java, Rabaul Death Spirals etc etc
 
Honestly I am now confused I thought folks wanted a challenge
 
p.s. Historic 1st turn is exactly that a historic 1st turn after turn 1 all bets are off and the AI is designed to try and keep you interested and give you a challenge
 
Going to go back to working on improving the scripts based on the feedback recieved so far - If you really want a 100% historic game then contact me and I will give a quick tutorial on building Ai scripts.
 
Andy

[AttuWatcher]

I would have built the AI the same way (like playing a PBEM player) had I been in your shoes Andy. I think it's the way to go for the game to provide the widest range of enjoyment and the most lasting challenge upon repeats.

[undercovergeek]

you did a great job dont worry - youve got what? 2-3 unhappy people while the rest are out there trying to fight off the AI - please dont pander to the 'but the AI did this to me' crowd - thats exactly what you were tasked to do!! i say congrats and look forward to many more surprises - if you dont want historical why the hell are you playing non-historical?

[Sheytan]

Frankly I dont want to see anything about this game changed unless there is a real and compelling reason to do so, IE bugs etc. As was mentioned people can as easily play the Dec 8th scenario as well. The game has only been out a few days and I see debates about mines, AAA etc.

Lets see how the game plays out before we jump all over about our pet peeves. I like this game, it is simply outstanding.

[Splinterhead]

I'm playing scenario 8 so my opinion is irrelevant, but here it is. I don't believe the engine can simulate the Pearl Harbor attack very well because it allows the Kates to use more heavy weapons than existed in real life. KB lingering for days while repeatedly attacking the US fleet with weapons that didn't exist is not a sign of good AI, it's fantasy. Having the KB bound by the historic logistic reality during the Pearl Harbor attack shouldn't mean you have to be bound by history later.

That's my opinion, and it only applies to scenario 1 as scenario 2 assumes the Japanese are better prepared so all bets are off.

[Keifer]

error

[sfbaytf]

I've not played the first few turns yet, but from what I've been hearing I don't see what the problem is-especially if non historical is checked. I want a mean, dirty and nasty AI that doesn't play fair-I don't mean cheating, just something that isn't afraid to kick me when I'm down. If you lose a carrier or 2 as the allied vs the AI in the first few turns in a long campaign game then that just makes it more challenging.

[Keifer]

ORIGINAL: Andy Mac

I just dont know what you want - a good and challenging game v the AI or a repeat of History which you will all rip apart - you can have either (although the latter would need about 4 months of solid effort on the scripts)

If its a total repeat of history then I guarantee you will see things like the unconquerable Java, Rabaul Death Spirals etc etc

Honestly I am now confused I thought folks wanted a challenge

p.s. Historic 1st turn is exactly that a historic 1st turn after turn 1 all bets are off and the AI is designed to try and keep you interested and give you a challenge

Going to go back to working on improving the scripts based on the feedback recieved so far - If you really want a 100% historic game then contact me and I will give a quick tutorial on building Ai scripts.

Andy

I like the tweak to make the AI stick around a few days.

If you are going to improve the scripts, make the AI nastier.

I just started a historical start scenario and I've got 4 BB's sunk and 4 with float damage over 50. I am eager to see if the KB sticks around and finishes off my 4 damaged battlewagons. Just in case I am evacuating every ship with less than 30 float damage and keeping my CV task forces well away until it is safe.

What would be really nasty would be for the KB to sail away on Dec 8, refuel at Kwaj, and then hit Pearl again a few weeks later.

keep up the good work

[Cap Mandrake]

What do you guys want the AI to do?

Recreate historical moves so you can anticipate everything and beat up on a bunch of lousy logic gates or kick your ass every once in a while so you dont get bored two weeks into the game.

[AttuWatcher]

ORIGINAL: Splinterhead
KB lingering for days while repeatedly attacking the US fleet with weapons that didn't exist is not a sign of good AI, it's fantasy.

This is an OOB issue, not AI. If they would run out of ammo then they would, I'm assuming , go home.

Does this happen with historical first turn on? With it off all bets should be off, just like scen 2.

[Keifer]

ORIGINAL: Mark Weston

ORIGINAL: Jim D Burns

ORIGINAL: Mark Weston

If camping out at Pearh Harbour is causing a player problems, then that's pretty much the definition of the AI doing a good job.

It's also indicative of the flak model failing utterly. No way could a CV group remain on station day after day hitting targets over and over... Historically they would have run out of planes due to flak losses within a day or two.

Jim

I'm not buying this. Where during WWII did AA demonstrate anything like the ability to destroy a six-carrier air group in two days?

At the Battle of Santa Cruz the Shokaku and Zuikaku air groups were massacred in one day of battle, mostly due to us AA.

I could easily see the Japanese KB air groups losing 50% in a few days of air attacks on Pearl harbor.

Historically on Dec 7 "Of Japan's 414[48] available planes, 29 were lost during the battle (nine in the first attack wave, 20 in the second), with another 74 damaged by antiaircraft fire from the ground." (from wikipedia)

In just the hour between the 1st and 2nd waves, US AA over Pearl became over 100% more effective.