RE: AE Land and AI Issues (Full Version)

All Forums >> [New Releases from Matrix Games] >> War in the Pacific: Admiral's Edition



Message

mariandavid -> RE: AE Land and AI Issues (11/30/2009 5:04:39 PM)

Andy: I concur with Hans on the one 'blot' on the AI performance - the tendency of CV TF to remain on station when they should move well away. Playing against your Ironman (as nasty creature indeed!) I checked one case - a TF that started with the usual 50 F, 50 DB, 50 TB was still hanging around Suva with its strength down to 40F, 4B, 0TB! Is it possible to code a withdrawal trigger in terms of time and strength - ie override with return to base after a random 2-3 days, a random less than x% of strength?



Wirraway_Ace -> RE: AE Land and AI Issues (11/30/2009 5:49:31 PM)

Andy,
trying to understand how and why the British 70th Div is organized the way you have it. It looks like you have one unit titled 70th Div that appears to be the 16th Bde + all the division troops and the other two bdes (14th and 23rd) are split out. There is no rebuild unit option.

I admit I don't know the Division's history well. My memory is many of its units became Chindit formations. Just trying to follow your logic.

thanks.



Andy Mac -> RE: AE Land and AI Issues (11/30/2009 6:47:46 PM)

Are you seeing the not withdrawing when understrength in 1094 i thought we had fixed it ??

70th Div HQ has all the Divisional Assets and disbands when the 3 compnoent bdes concvert to Chindit

Its not designed to rebuild

I would have liked to have had a special convert to Chindit option for 70th Brit Div but it jhust wasnt possible to justify the coding time for one conversion.

So each Bde converts to a Chindit TOE and the Div Assets are disbanded to the pool at a point in time



Wirraway_Ace -> RE: AE Land and AI Issues (11/30/2009 6:56:19 PM)

Andy, re the 70th Brit Div, I may not be close enough to its withdrawal date to see the flag yet. I will check the unit withdrawals list tonight.



Wirraway_Ace -> RE: AE Land and AI Issues (11/30/2009 7:09:28 PM)

Andy, re 70th Brit Div withdrawal in version 1.0.2.1094. It is 1 Apr 42 in my game and the division is scheduled to withdraw in 426 days (1 Jun 43). So it looks like it can fight in the 1st Arakan (as I understand some of its units historically did) before becoming Wingate's personal army...



mariandavid -> RE: AE Land and AI Issues (12/1/2009 9:26:58 PM)

Andy; Let me check over the next few turns. The TF that was floating endlessly around Suva and safe because its 40 crack F would wreck the few DB and TB I have left (!) seems to have vanished. I want to see what happens with the second one that seems to appear and disappear near Pt. Moresby. Whatever happens at the very least your patch change has created uncertainty.

One other AI 'blemish' I have noted; when attacking India it seems to only use one attack route, staggering north of Akyab rather than at the same time advancing inland against Imphal or launch an amphib attack in the vicinity. That makes it easy to check (at least so far he hastily added!)



Andy Mac -> RE: AE Land and AI Issues (12/1/2009 9:35:50 PM)

[:D][:D]

The AI scripts dont get updated in the patch unfortunately (I dont think at least) so you are seeing the code improvements

Re the India attack variant lets just say I spent the best part of a week on that script alone play it on scen 1 and its a lot tougher play it on scen 10 and its downright scary.....

of course I am not going to tell you which one it is that would be unssporting on the AI





bsq -> RE: AE Land and AI Issues (12/2/2009 10:46:31 PM)

Hope this is the right place to post this observation and that the AI guys don't take umbrage - so here goes...

Are there two sets of rules concerning aspects such as supply usage and attrition due to island stacking limits being exceeded?

By this I mean one for the AI (hence posting here) and one for a human player?

I ask because I am noticing that the AI seems not be as heavily penalised for overstacking

(worked examples - I over-stacked Canton by 100% and lost all excess troops in around 10 turns (600 per turn) whereas the AI currently has Iwo Jima over 100% over stacked and is losing only around 50 excess troops per turn - the latter despite constant attention from 12 CV, 6 CVL, 12 BB, numerous CA, CL, DD and 6 Sqns of B-24's for 2 months)

Also the AI seems to be able to run a deficit budget in terms of supply.  Notice this when a location runs out of supply, but gets subjected to air attack - it will fight back by drawing on non-existent supplies creating negative values for supply.

I am all for giving the AI every chance it needs, but this type of advantage is an open cheque book.  How can I be sure when the AI is going to adopt creative accounting making any assault on that location pointless or at the very least not cost effective.  I am on the point of sticking 16 subs around Iwo-Jima and moving on as clearly the AI is not playing by the same rules that constrain me.  This is not the only location where this is occuring, but I particularly wanted IJ to see if my B-29 losses could be reduced by the use of long range fighters.




Andy Mac -> RE: AE Land and AI Issues (12/2/2009 11:57:12 PM)

AI doest get exemption from overstacking rules I see the opposite in my AI games when the AI is out of supply no aircraft fly so I know its not getting help.

Sounds like a bug



bsq -> RE: AE Land and AI Issues (12/3/2009 11:36:14 PM)


quote:

ORIGINAL: Andy Mac

AI doest get exemption from overstacking rules I see the opposite in my AI games when the AI is out of supply no aircraft fly so I know its not getting help.

Sounds like a bug

Strike missions don't fly. CAP does. Ground units fire Flak. Coastal guns kill my landing craft, reload (with what?) and kill the next wave and the next and the next... Checking the numbers they are doing this with pretty generous supply 'overdrafts'. Makes it hard where to plan to go next because no matter how tight the blockade, the ability of the AI to creatively account (whether by accident or design), negates any degree of planning and I might as well invade location by location with a 10:1 local numerical advantage (which takes the fun of planning campaigns away as all that needs to be done is to amass overwhelming force and strike - which slows the pace of the game down as there is only enough lift capacity to conduct one op at a time)



Andy Mac -> RE: AE Land and AI Issues (12/3/2009 11:41:55 PM)

Havent even seen CAP fly as the units become all disabled

The rest is same as it always has been units fight at 25% effectiveness when out of supply but they still fight thats not an AE AI cheat thats as it has been since year dot of stock

p.s. its the same for players




Andy Mac -> RE: AE Land and AI Issues (12/3/2009 11:42:47 PM)

If you have an example send me a save but this doesnt sound like an AI cheat



Andy Mac -> RE: AE Land and AI Issues (12/4/2009 12:38:11 AM)

I amn specifically looking for feedback from anyone who has attacked.

India
NZ
Fiji
or Australia v the AI

Andy



gingerbread -> RE: AE Land and AI Issues (12/4/2009 6:45:57 AM)

Expanding SPS size 0 ports is slow as stated by designers, but it seems that it is enough to get to 10% to get a level increase - can anyone confirm?

/g



Andy Mac -> RE: AE Land and AI Issues (12/4/2009 9:39:59 PM)

Sorry I dont understand the question



FOW -> RE: AE Land and AI Issues (12/6/2009 9:22:57 AM)

Re 5th British Division

This unit arrives in Mombasa in late May 42, with a withdrawal date of 200+ days (can't recall the exact detail.
But this unit is permanently restricted to India Command(R).
What's the point of it? - you can't change command, or transport, or even disband it !!!
Is this correct (in database terms)?

(Patch 2 Beta, Grand campaign, Allies vs Jap AI, started under original version)



Andy Mac -> RE: AE Land and AI Issues (12/6/2009 12:34:26 PM)

The unit was removed in patch 1 OOB update



FOW -> RE: AE Land and AI Issues (12/6/2009 5:04:40 PM)


quote:

ORIGINAL: Andy Mac

The unit was removed in patch 1 OOB update


Thanks Andy - I suppose if I'd started a new game it would not be there - I'll just ignore it then



tbridges -> RE: AE Land and AI Issues (12/8/2009 5:24:40 PM)

When LCUs are given orders to move to a new location that is several hexes away, and there is more than one way to get to that location, how does the unit decide which path to follow? In other words, if the destination is three hexes away directly through a jungle or 7 seven hexes away if it chooses a roundabout route by sticking to a road - what happens?

Does it compute the lowest possible movement cost and take that route?



BigJ62 -> RE: AE Land and AI Issues (12/8/2009 11:15:25 PM)

It uses the fastest route and while avoiding enemy.



tbridges -> RE: AE Land and AI Issues (12/9/2009 2:07:36 AM)

Thanks



erstad -> RE: AE Land and AI Issues (12/9/2009 4:27:03 AM)


quote:

ORIGINAL: BigJ62

It uses the fastest route and while avoiding enemy.


Related question - if you are one hex away it appears to always move directly. Is there a way to make it take the quickest path rather than the shortest, other than moving one hex at a time? Particularly an issue when a road through bad terrain bends.



HansBolter -> RE: AE Land and AI Issues (12/9/2009 10:14:50 AM)


quote:

ORIGINAL: erstad


quote:

ORIGINAL: BigJ62

It uses the fastest route and while avoiding enemy.


Related question - if you are one hex away it appears to always move directly. Is there a way to make it take the quickest path rather than the shortest, other than moving one hex at a time? Particularly an issue when a road through bad terrain bends.




Could the devs please, please take a look at this as it doesn't seem to be working as designed.

Units trekking across Auatralia via the rail/trail route to Darwin inevitably go off the trail and into heavy terrain when given an order to move from the rail head to the next town.

If you don't give them hex by hex marching orders they go astray. There are no enemy units on the continent that could or should be affecting the route determination.



BigJ62 -> RE: AE Land and AI Issues (12/9/2009 12:17:34 PM)

This a feature called direct hex movement when setting destination to an adjacent hex.


quote:

ORIGINAL: erstad


quote:

ORIGINAL: BigJ62

It uses the fastest route and while avoiding enemy.


Related question - if you are one hex away it appears to always move directly. Is there a way to make it take the quickest path rather than the shortest, other than moving one hex at a time? Particularly an issue when a road through bad terrain bends.





BigJ62 -> RE: AE Land and AI Issues (12/9/2009 12:19:04 PM)

What version are you referring to? We made improvements to this in patch 2.


quote:

ORIGINAL: HansBolter


quote:

ORIGINAL: erstad


quote:

ORIGINAL: BigJ62

It uses the fastest route and while avoiding enemy.


Related question - if you are one hex away it appears to always move directly. Is there a way to make it take the quickest path rather than the shortest, other than moving one hex at a time? Particularly an issue when a road through bad terrain bends.




Could the devs please, please take a look at this as it doesn't seem to be working as designed.

Units trekking across Auatralia via the rail/trail route to Darwin inevitably go off the trail and into heavy terrain when given an order to move from the rail head to the next town.

If you don't give them hex by hex marching orders they go astray. There are no enemy units on the continent that could or should be affecting the route determination.




HansBolter -> RE: AE Land and AI Issues (12/9/2009 1:16:17 PM)

Was referring to Patch 1. Just downloaded patch 2. Will watch for the improvement. Thanks.



bsq -> RE: AE Land and AI Issues (12/18/2009 10:03:55 PM)

Some device issues

LVT(A)4 (Device 1188) only has a build rate of 4 PCM, given that 1840 were built for the US Army and USMC, the rate seems very low (http://www.wwiivehicles.com/usa/amphibious/lvt.asp).

The 17 pdr seems to take the least powerful of it's ammunition types to determine it's anti armour, whereas M3 equiped vehicles (M26 and M36) take the most powerful of it's ammunition types - why the discrepancy. Either the figures should be 220/206 or 120/140, not the 220/140 currently used (M3/OQF 17 Pdr). As things stand the 6 pdr is a more capable weapon (which it was not).




Chris21wen -> RE: AE Land and AI Issues (12/22/2009 7:01:07 PM)

2nd USMC Air wing is shown as a US Army unit.



khyberbill -> RE: AE Land and AI Issues (12/22/2009 11:04:36 PM)

quote:


LVT(A)4 (Device 1188) only has a build rate of 4 PCM, given that 1840 were built for the US Army and USMC, the rate seems very low (http://www.wwiivehicles.com/usa/amphibious/lvt.asp).


How many of those 1840 were sent to Europe?



witpqs -> RE: AE Land and AI Issues (12/25/2009 7:22:09 PM)

Andy,

Question re Ironman scenario.

A while back it was discovered in the main GC scenario that a bunch of Soviet bases gradually starve over the long run because they can't supplied while the Soviets are inactive (I think it's the ones that are off of the road/rail net). IIRC some kind of adjustment was made to fix/mitigate the problem.

Was the fix applied to the Ironman sceanrio?



Page: <<   < prev  62 63 [64] 65 66   next >   >>

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
1.078125