Matrix Games Forums

Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

RE: AI for MWiF - USA

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [New Releases from Matrix Games] >> World in Flames >> AI Opponent Discussion >> RE: AI for MWiF - USA Page: <<   < prev  2 3 [4] 5 6   next >   >>
Login
Message << Older Topic   Newer Topic >>
RE: AI for MWiF - USA - 10/25/2007 3:56:18 AM   
doctormm


Posts: 124
Joined: 5/28/2004
Status: offline

quote:

ORIGINAL: composer99

Have we had much discussion on the CW-US strategic bombing offensive?

If not, we should.

The Millennium WiF Annual suggests that bombing should be early & often. My friend Pablo told me of a game he played against Andrew Rader where Andrew's CW/US combo would pound him for 18+ raids per turn in 43-44 (so figure between 9-15 pp lost per turn).

Any one else's thoughts on this?


My (more or less) Rule One as an Allied player is that victory as the Allies is about counter density/attrition. If you allow the Axis to build up huge forces, it doesn't matter that you can build up equally huge forces backed up by a stack of OChits. Germany will be strong everywhere. You'll only be able to grind Germany back a couple of hexes here and there. But if you keep attrition going, when the initiative shifts your way you'll be in a position to hit the weak spots of the Axis.

To me, an early strat bombing campaign is one way to get the attrition going, and it also forces Germany to put some effort into defending the factories, furthering the goal of lowering counter density at the front when the initiative swings.

(in reply to composer99)
Post #: 91
RE: AI for MWiF - USA - 10/25/2007 10:11:10 AM   
wosung

 

Posts: 692
Joined: 7/18/2005
Status: offline
Maybe the US/CW AIO should have different strategies for the strategic air offensive to choose from, maybe on a turn to turn base:

-day or night bombing
-oil or factories
-big wing or many raids
-escorted or not
-targetting the economy or axis fighters
-strategic air minded or open to other tasks (carpet bombing, ground strike, naval search, air transport).

Regards

(in reply to doctormm)
Post #: 92
RE: AI for MWiF - USA - 10/25/2007 11:22:00 AM   
haromar

 

Posts: 19
Joined: 10/22/2007
Status: offline
When commenting this post I assumed that sooner or later in the development process, the designers will put up their AI strategies, tactics etc to discussion.

At that stage, the strategic bombing discussion will be very interesting

Its a never ending issue ... Ask a few wiffers on that subject and you end up with loads of different ideas

While for the JP on China you can simplify the discussion to an extent, for the Wallies it depends so many factors, that I'm curious as to when the AI development process feedback will go under way.

On a separate note, few discuss the Ge strat campaign on France in 39. Getting France below the 5 BPs hinders him from building the Mech or/and the HQ for the summer 40.

(in reply to wosung)
Post #: 93
RE: AI for MWiF - USA - 10/25/2007 11:28:22 AM   
Neilster


Posts: 2890
Joined: 10/27/2003
From: Hobart, Tasmania, Australia
Status: offline

quote:

ORIGINAL: haromar

When commenting this post I assumed that sooner or later in the development process, the designers will put up their AI strategies, tactics etc to discussion.

At that stage, the strategic bombing discussion will be very interesting

Its a never ending issue ... Ask a few wiffers on that subject and you end up with loads of different ideas

While for the JP on China you can simplify the discussion to an extent, for the Wallies it depends so many factors, that I'm curious as to when the AI development process feedback will go under way.

On a separate note, few discuss the Ge strat campaign on France in 39. Getting France below the 5 BPs hinders him from building the Mech or/and the HQ for the summer 40.

There's already been heaps of discussion. Check out pages 2,3 etc for a general AI discussion as well as ones specific to each major power (and partisans).

Cheers, Neilster



(in reply to haromar)
Post #: 94
RE: AI for MWiF - USA - 10/25/2007 2:49:05 PM   
haromar

 

Posts: 19
Joined: 10/22/2007
Status: offline

quote:

ORIGINAL: Neilster

There's already been heaps of discussion. Check out pages 2,3 etc for a general AI discussion as well as ones specific to each major power (and partisans).

Cheers, Neilster



Oh God I checked the discussion, also for other MPs, wow.

Unfortunately (or maybe fortunately ) I've got no experience programming AI, so no clue as to all the comments /ideas are related .

I'll "study" the post a little to find out how the AI strategies or tactics are clustered or ares related to other issues such as optionals or game situation.

E.g. If I play (actually forced top play ) with the variable reorg optional, my Strategic Bomber campaign will probably look quite different than w/o that optional. At least different enough to warrant a change in AI procedures, however that change looks like.

Question: Are there like executive summaries / clustering / interdependencies flow charts of all these different AI posts?

after looking at all those posts.

(in reply to Neilster)
Post #: 95
RE: AI for MWiF - USA - 10/25/2007 8:36:06 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: haromar


quote:

ORIGINAL: Neilster

There's already been heaps of discussion. Check out pages 2,3 etc for a general AI discussion as well as ones specific to each major power (and partisans).

Cheers, Neilster



Oh God I checked the discussion, also for other MPs, wow.

Unfortunately (or maybe fortunately ) I've got no experience programming AI, so no clue as to all the comments /ideas are related .

I'll "study" the post a little to find out how the AI strategies or tactics are clustered or ares related to other issues such as optionals or game situation.

E.g. If I play (actually forced top play ) with the variable reorg optional, my Strategic Bomber campaign will probably look quite different than w/o that optional. At least different enough to warrant a change in AI procedures, however that change looks like.

Question: Are there like executive summaries / clustering / interdependencies flow charts of all these different AI posts?

after looking at all those posts.

Well, I have summarized and organized them into strategic plans for each major power. They aren't as polished as I will need them for data entry, but they are getting closer.

Pages for each major power:
China - 4 (very thin gruel)
CW - 47
France - 18 (significant portions of my polished version have been posted in the France AIO thread - somewhere towards the end)
Germany - 36
Italy - 29 (same comment as for France)
Japan - 15 (could use more)
USA - 24
USSR - 23

_____________________________

Steve

Perfection is an elusive goal.

(in reply to haromar)
Post #: 96
RE: AI for MWiF - USA - 10/25/2007 9:21:52 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline
Any strategic raid into the teeth of fighter cover, if you can manage it, should be as much about shooting down the defending fighter as getting the bombing raid through.

Strat bomb raids are good for hitting the Germans (and maybe the Italians) for .75 to 1.5 bp apiece. Shooting down the fighter is good for dinging them for 2-5 bp depending on what plane you shot down was and if you killed the pilot or not.

The other thing is that each turn the strat raids should come in two waves. The first to swamp and exhaust fighter cover in the region, the next to really try to pound all factories & oil into the ground.

_____________________________

~ Composer99

(in reply to Shannon V. OKeets)
Post #: 97
RE: AI for MWiF - USA - 10/25/2007 10:42:03 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
quote:

ORIGINAL: composer99
The other thing is that each turn the strat raids should come in two waves. The first to swamp and exhaust fighter cover in the region, the next to really try to pound all factories & oil into the ground.

Use HQ to reorg planes for that if possible. Air Action is best too for the first wave.

(in reply to composer99)
Post #: 98
RE: AI for MWiF - USA - 10/26/2007 4:30:50 AM   
Neilster


Posts: 2890
Joined: 10/27/2003
From: Hobart, Tasmania, Australia
Status: offline
"China - 4 (very thin gruel)"

Another famine there eh? Probably caused by one of those Japanese "rice offensives"

Cheers, Neilster

(in reply to Shannon V. OKeets)
Post #: 99
RE: AI for MWiF - USA - 10/26/2007 1:05:54 PM   
haromar

 

Posts: 19
Joined: 10/22/2007
Status: offline

quote:

ORIGINAL: Shannon V. OKeets
Well, I have summarized and organized them into strategic plans for each major power. They aren't as polished as I will need them for data entry, but they are getting closer.

Pages for each major power:
China - 4 (very thin gruel)
CW - 47
France - 18 (significant portions of my polished version have been posted in the France AIO thread - somewhere towards the end)
Germany - 36
Italy - 29 (same comment as for France)
Japan - 15 (could use more)
USA - 24
USSR - 23


The page http://www.eurowif.de/ offers targets, tactics and set ups for all MPs other than GE. Its by an excellent player, it can give you another view on some topics. Unfortunately its in German, but should be easy to translate. I'll try my best.

Whats lacking is later game strategy e.g. depending on the situation.

I'm also not sure how the overall AI will be behave, agressive, cautious, defensive etc. The above targets would change accordingly. This is not dependent on MP or Year, not even on game sitaution. E.g. some CW will try to get 4 corps into an attacked Greece at all circumstances while some want even consider it.

Have to do more reading on these threads to understand the overall AI picture.

Where are those summaries documented?

Thanks.

(in reply to Shannon V. OKeets)
Post #: 100
RE: AI for MWiF - USA - 10/26/2007 8:05:29 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: haromar


quote:

ORIGINAL: Shannon V. OKeets
Well, I have summarized and organized them into strategic plans for each major power. They aren't as polished as I will need them for data entry, but they are getting closer.

Pages for each major power:
China - 4 (very thin gruel)
CW - 47
France - 18 (significant portions of my polished version have been posted in the France AIO thread - somewhere towards the end)
Germany - 36
Italy - 29 (same comment as for France)
Japan - 15 (could use more)
USA - 24
USSR - 23


The page http://www.eurowif.de/ offers targets, tactics and set ups for all MPs other than GE. Its by an excellent player, it can give you another view on some topics. Unfortunately its in German, but should be easy to translate. I'll try my best.

Whats lacking is later game strategy e.g. depending on the situation.

I'm also not sure how the overall AI will be behave, agressive, cautious, defensive etc. The above targets would change accordingly. This is not dependent on MP or Year, not even on game sitaution. E.g. some CW will try to get 4 corps into an attacked Greece at all circumstances while some want even consider it.

Have to do more reading on these threads to understand the overall AI picture.

Where are those summaries documented?

Thanks.

They are my own notes and they undergo continual revisions (when I get the time). Posting them here isn't really feasible because they are so long. If you are interested, send me an email (SHokanson@HawaiianTel.net).

_____________________________

Steve

Perfection is an elusive goal.

(in reply to haromar)
Post #: 101
RE: AI for MWiF - USA - 10/26/2007 9:07:19 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline
In the CW AIO thread we have discussed convoy building and how it relates to Lend-Lease.

As the US is the principal lendor of the Allies, the question for the US AIO is how much to lend and when?

Early in the war, when the US has a lower production multiple than the other Allies, it makes sense to lend somewhere between 2-4 resources or so to the CW (and/or France if it is still installed on the Continent).

After the US hits the 1.25 production multiple, it instead makes sense to go to lending build points. The absolute minimum, in my view, is for the US to lend to CW/USSR/Fr 10% of its build point output going forward. In actual fact lending between 10-20 build points to the other Allies from the US' entry into the war against all three Axis powers should be considered, with the exact amount to increase from 42 to 45 as the US' production increases.

Unless we've discussed this in depth and I'm wasting time, energy & brainpower.



_____________________________

~ Composer99

(in reply to Shannon V. OKeets)
Post #: 102
RE: AI for MWiF - USA - 2/22/2008 10:34:03 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline
Here is something I haven't worked on in a long time. I am bringing all the AIO strategic plans up-to-date with the annotations I made to them last year - the US is the last of the 8, so I am almost done with that task.

Here is my reworking of the fine analysis by Composer 99. Almost all of this is his.

What I want to do is fill in the Vital Hexes pieces. Feel free to double check the other sections shown here too. To understand this it is best to read about the sea areas, both defensive and offensive. I have created a universal structure for all 8 major powers, which is why I start with the Victory hexes. Oh, and this is just for the PTO. I'll get to the ETO next/soon.

========
GS 1.1 Victory cities, vital hexes, and sea areas

Pacific Theater of Operations
Victory Cities - Defense
Western Americas: Vancouver, Los Angeles, Panama
Eastern Pacific:
Hawaiian Perimeter: Honolulu
North Pacific: Dutch Harbor
South Pacific: Rabaul, Pago Pago, Canberra, Wellington
Southwestern Pacific: Kwajalein, Truk, Rabaul, Manila
NEI: Manila, Saigon, Singapore, Batavia
Japanese Perimeter: Vladivostok, Manila

Vital Hexes - Defense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter: Midway
North Pacific:
South Pacific:
Southwestern Pacific:
NEI: Canton
Japanese Perimeter: Hong Kong

Vital Sea Areas - Defense
Western Americas: West Coast, Mexican Coast, Gulf of Panama, East Pacific
Eastern Pacific: Gulf of Alaska, Mendocino, Clarion, East Polynesia, South Pacific
Hawaiian Perimeter: Central Pacific Ocean, The Marshalls, Hawaiian Islands, Christmas Islands
North Pacific: Bering Sea, Okhotsk Sea
South Pacific: Polynesia, New Zealand Coast, Coral Sea, Tasman Sea
Southwestern Pacific: The Solomons, Bismarck Sea, Marianas
NEI: Timor Sea, South China Sea
Japanese Perimeter:

Victory Cities - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
NEI:
Japanese Perimeter: Tokyo, Port Arthur, Shanghai, Taihoku

Vital Hexes - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
Japanese Perimeter:
NEI:

Vital Sea Areas - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
NEI:
Japanese Perimeter: Sea of Japan, Japanese Coast, China Sea

Land Regions
- Australia/New Zealand
- Fiji/New Hebrides
- Solomon Islands
- New Guinea
- NEI
- Philippines
- Marshall Islands
- Caroline Islands
- Guam/Marianas/Palau Islands
- Bonin Islands
- Taiwan
- Japan
- Mainland Asia South (Indochina, Thailand, Malaya,)
- Mainland Asia North (Korea, Manchuria, China north of the coastal mountains)
====================

_____________________________

Steve

Perfection is an elusive goal.

(in reply to composer99)
Post #: 103
RE: AI for MWiF - USA - 2/22/2008 11:04:18 PM   
Norman42


Posts: 244
Joined: 2/9/2008
From: Canada
Status: offline
quote:


NEI: Canton


I'm assuming that should be Batavia instead?

(in reply to Shannon V. OKeets)
Post #: 104
RE: AI for MWiF - USA - 2/23/2008 12:21:19 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: Norman42

quote:


NEI: Canton


I'm assuming that should be Batavia instead?

No.

The two sea areas I have labeled NEI (no other name came to mind) are: South China Sea and Timor Sea. So Canton is an important land hex adjcent to the South China Sea. Others should be Hainan, Hanoi, Palembang, ... This is what I haven't filled out yet and I am seeking advice.

Victory Cities are listed first, and are of primary importance. Vital Hexes are still very important, though secondary to Victory Cities.

The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes. In the Pacific a lot of the protection for hexes requires control of sea areas, hence the segmentation of all these into Western America, Eastern Pacific, etc.

When on the offensive, the AIO will look to control sea areas first and then occupy the hexes within a sea area. The sequence will be from the first on this list to the last: Western America first, ..., Japanese Perimeter last. These actually work as a spectrum, working roughly from east to west, with the US usually failing to defend the western end of the spectrum and having to counterattack to recapture it.

These can be viewed as defensive fall back positions when the Japanese are on the offensive (starting with NEI, followed by Southwestern Pacific) and as an offensive objective list running in the reverse direction (which historically started with North Pacific and South Pacific - the Hawaiian Perimeter having only been partially breached).

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Norman42)
Post #: 105
RE: AI for MWiF - USA - 2/23/2008 3:29:19 AM   
Norman42


Posts: 244
Joined: 2/9/2008
From: Canada
Status: offline


quote:

The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes.


This is why I thought it was Batavia, since it is listed in "Vital Hexes - Defence". All the other hexes listed in:

quote:


Vital Hexes - Defense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter: Midway
North Pacific:
South Pacific:
Southwestern Pacific:
NEI: Canton
Japanese Perimeter: Hong Kong


...are cities/hexes the US and allies control at start and can allocate defensive assets to (early in war when on defence mode). Since the US doesn't control Canton(unless this has changed somewhere? I'm pretty sure Canton was on the Japanese side of the 1939 start line) I was wondering how it would 'defend' this Vital Hex while on defensive mode unless it captured it from the Japanese on Offense mode first? Would Canton not be "Vital Hexes - Offense" in that case?

Perhaps I am misunderstanding your system.




(in reply to Shannon V. OKeets)
Post #: 106
RE: AI for MWiF - USA - 2/23/2008 3:44:57 AM   
Mziln


Posts: 1107
Joined: 2/9/2004
From: Tulsa Oklahoma
Status: offline

quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: Norman42

quote:


NEI: Canton


I'm assuming that should be Batavia instead?

No.

The two sea areas I have labeled NEI (no other name came to mind) are: South China Sea and Timor Sea. So Canton is an important land hex adjcent to the South China Sea. Others should be Hainan, Hanoi, Palembang, ... This is what I haven't filled out yet and I am seeking advice.

Victory Cities are listed first, and are of primary importance. Vital Hexes are still very important, though secondary to Victory Cities.

The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes. In the Pacific a lot of the protection for hexes requires control of sea areas, hence the segmentation of all these into Western America, Eastern Pacific, etc.

When on the offensive, the AIO will look to control sea areas first and then occupy the hexes within a sea area. The sequence will be from the first on this list to the last: Western America first, ..., Japanese Perimeter last. These actually work as a spectrum, working roughly from east to west, with the US usually failing to defend the western end of the spectrum and having to counterattack to recapture it.

These can be viewed as defensive fall back positions when the Japanese are on the offensive (starting with NEI, followed by Southwestern Pacific) and as an offensive objective list running in the reverse direction (which historically started with North Pacific and South Pacific - the Hawaiian Perimeter having only been partially breached).


WESTPAC (Western Pacific), ALFSEA (Allied Land Forces South East Asia), FEAF (Far East Asia Force), or FEA (Far East Asia).

_____________________________


(in reply to Shannon V. OKeets)
Post #: 107
RE: AI for MWiF - USA - 2/23/2008 4:11:41 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: Norman42



quote:

The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes.


This is why I thought it was Batavia, since it is listed in "Vital Hexes - Defence". All the other hexes listed in:

quote:


Vital Hexes - Defense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter: Midway
North Pacific:
South Pacific:
Southwestern Pacific:
NEI: Canton
Japanese Perimeter: Hong Kong


...are cities/hexes the US and allies control at start and can allocate defensive assets to (early in war when on defence mode). Since the US doesn't control Canton(unless this has changed somewhere? I'm pretty sure Canton was on the Japanese side of the 1939 start line) I was wondering how it would 'defend' this Vital Hex while on defensive mode unless it captured it from the Japanese on Offense mode first? Would Canton not be "Vital Hexes - Offense" in that case?

Perhaps I am misunderstanding your system.


You have it right. Canton should be in the Offense Vital Hexes - my mistake. You might want to check the others too.

If a hex belongs to a neutral country, (e.g., Netherlands) then it should be listed as under Offense. Should the Axis DOW the neutral, then the AIO will move it from Offense to Defense. But then all hexes that are taken by the Axis move to the Offense lists automatically. What these lists are really doing is laying out the starting positions as Offense versus Defense. I guess I could automate this based on which countries are at war and who controls what, but I want to see the starting assignments clearly for a lot of reasons.

---
Batavia is not an Vital Hex because it appears as a Victory Hex.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Norman42)
Post #: 108
RE: AI for MWiF - USA - 2/23/2008 4:12:27 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: Mziln


quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: Norman42

quote:


NEI: Canton


I'm assuming that should be Batavia instead?

No.

The two sea areas I have labeled NEI (no other name came to mind) are: South China Sea and Timor Sea. So Canton is an important land hex adjcent to the South China Sea. Others should be Hainan, Hanoi, Palembang, ... This is what I haven't filled out yet and I am seeking advice.

Victory Cities are listed first, and are of primary importance. Vital Hexes are still very important, though secondary to Victory Cities.

The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes. In the Pacific a lot of the protection for hexes requires control of sea areas, hence the segmentation of all these into Western America, Eastern Pacific, etc.

When on the offensive, the AIO will look to control sea areas first and then occupy the hexes within a sea area. The sequence will be from the first on this list to the last: Western America first, ..., Japanese Perimeter last. These actually work as a spectrum, working roughly from east to west, with the US usually failing to defend the western end of the spectrum and having to counterattack to recapture it.

These can be viewed as defensive fall back positions when the Japanese are on the offensive (starting with NEI, followed by Southwestern Pacific) and as an offensive objective list running in the reverse direction (which historically started with North Pacific and South Pacific - the Hawaiian Perimeter having only been partially breached).


WESTPAC (Western Pacific), ALFSEA (Allied Land Forces South East Asia), FEAF (Far East Asia Force), or FEA (Far East Asia).

Thanks. I like Far East Asia and will replace NEI with this better name.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Mziln)
Post #: 109
RE: AI for MWiF - USA - 2/23/2008 6:01:15 AM   
brian brian

 

Posts: 3191
Joined: 11/16/2005
Status: offline
you might want to create another category - Oil Hexes - that you could weight differently in a game using or not using Option 48

(in reply to Shannon V. OKeets)
Post #: 110
RE: AI for MWiF - USA - 2/23/2008 12:55:52 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
The way this will work in practice, is that the AIO, when on defence, will worry about protecting its Victory Cities first. If they look ok, it will protect its Vital Hexes. In the Pacific a lot of the protection for hexes requires control of sea areas, hence the segmentation of all these into Western America, Eastern Pacific, etc.


Imho, this approach works well for land-theatres (Europe, China etc.), but regarding the Pacific the risk of defending sea areas in order to defend victory cities should be weighted, as loss of especially CVs will delay the US naval-strength take-over, since rebuilding them takes so long. Also, 2 corps+1 div (or to some degree just 1 white-print corps) often forces the Japanese player to do a significant commitment to take just a single victory city, even if out of supply.
So, I think that it is important that the US don't commit its main fleet (especially the CVs) early on when risking overwhelming Japanese retaliation. Better to use LBAs (or single SCS) to try to stay in supply, and then use the fleets to hit smaller japanese fleets in remote areas where main japanese fleet can only reach low box, or after main japanese fleets have been committed.

quote:

ORIGINAL: Shannon V. OKeets
When on the offensive, the AIO will look to control sea areas first and then occupy the hexes within a sea area. The sequence will be from the first on this list to the last: Western America first, ..., Japanese Perimeter last. These actually work as a spectrum, working roughly from east to west, with the US usually failing to defend the western end of the spectrum and having to counterattack to recapture it.

These can be viewed as defensive fall back positions when the Japanese are on the offensive (starting with NEI, followed by Southwestern Pacific) and as an offensive objective list running in the reverse direction (which historically started with North Pacific and South Pacific - the Hawaiian Perimeter having only been partially breached).

Maybe take overall US Pacific strategy into consideration. Some options:

- slow advancement, retake victory cities and important hexes (maybe even most ports/cities) as the sea areas are controlled.
This could be done from either Pacific (Honolulu, Truk, etc.); or from Australia (primary supply source, so keeping control of Pacific sea areas less important) and then striking directly against the Indonesian oil and China seas.

- quick advancement - head for inner Japanese areas (and/or oil) quickly, leaving non-crucial cities/hexes to be taken later, when Japanese are busy defending the sea areas around Japan. With such a strategy: McArthur + Phil. TERR in Manila (must be rebuilt immidiately if lost) (maybe even + 1 DIV) important - if they survive then there is an available major port right in the japanese heartland. Also, the port enhancing TRS units become important, as major ports, such as Truk, can then be left until later, saving a lot of time.
Such an approach requires a strong naval force which can ensure victory - if doing an overall early Japan strategy (going in seriously from (42) 43) then early US build must be geared towards this. If doing Europe-first (e.g. Italy in '43) and leaving serious attack on Japan to '44 then time is important and the quick-advance strategy may be better than a slower 'securing-all-before-advancing'-strategy.

- war of attrition - go (early on) for the South China sea area in force (from Australia), preferably with SUBs/LBA only at first, threatening the Japanese oil convoy lines and forcing the Japanese to defend here. Thus the US will lose quickly rebuilt LBA and SUB while Japan will lose naval units, and hopefully oil (maybe even do some strat-bombing of the oil-fields from eg. Makassar/Menado), and then later, when strong enough, or if Japan commit elsewhere, strike with CVs/navy. This strategy obviously works best if Japan is somewhat low on oil (i.e. no Persian and preferably no Burmesian oil fields)

Just some thoughts.

_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 111
RE: AI for MWiF - USA - 2/23/2008 1:11:29 PM   
brian brian

 

Posts: 3191
Joined: 11/16/2005
Status: offline
standard - leave McArthur in Manila. I never understand why the richest country in the game would want to evacuate a unit perfectly positioned to delay the enemy.

(in reply to npilgaard)
Post #: 112
RE: AI for MWiF - USA - 2/23/2008 2:54:17 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline
quote:

ORIGINAL: Shannon V. OKeets
Here is something I haven't worked on in a long time. I am bringing all the AIO strategic plans up-to-date with the annotations I made to them last year - the US is the last of the 8, so I am almost done with that task.


Looks good - makes commenting on specific parts of strategy much easier/structured. Looking forward to see plans for other MPs.
You are very good at bringing such somewhat 'loose' things into a tight and useful structure - but that comes with the job, I assume :)

quote:

ORIGINAL: Shannon V. OKeets
Here is my reworking of the fine analysis by Composer 99. Almost all of this is his.

What I want to do is fill in the Vital Hexes pieces. Feel free to double check the other sections shown here too. To understand this it is best to read about the sea areas, both defensive and offensive. I have created a universal structure for all 8 major powers, which is why I start with the Victory hexes. Oh, and this is just for the PTO. I'll get to the ETO next/soon.


Looks very good overall. A few comments below.

quote:

ORIGINAL: Shannon V. OKeets
Vital Hexes - Defense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter: Midway
North Pacific:
South Pacific:
Southwestern Pacific:
NEI: Canton
Japanese Perimeter: Hong Kong


Maybe I have listed to many 'vital' hexes below (many are not 'vital' but merely 'nice to control' :) - if needed I can make a harder priority ): don't place land units in all of the below - that will costs to many units - only in most important hexes or where likely invasions are to come (also: depending on available US land units).
Also, in general: if not threatened, most of the hexes located well beyond the japanese controlled areas don't need land units to protect them - only if they can be reached by Japanese invasion forces - but keep a SCS in a 1-2 sea areas wide zone beyond the current front line, to avoid japanese surprise invasions with DIV on fast SCS on out-of-supply islands, to keep supply chains, and - especially if playing with 'in the presence of the enemy' - to reduce Japanese strike range into 'hinterlands')

For each area: in order of priority:

Western Americas:all factory/resource/major port hexes
Eastern Pacific: (Papeete)
South Pacific: Auckland, Brisbane (and other Australian Major Ports), Port Moresby (if planning to strike north from Australia), Suva, Port Vila, Nouma (no supply-chain needed)
Southwestern Pacific: minor ports at Marshall islands
NEI: if planning to go for the oil/from Australia: Makassar/Menodo (will require troops ready to sail in as soon as Japan DoW on NEI, though)

quote:

ORIGINAL: Shannon V. OKeets
Vital Sea Areas - Defense

If playing with 'Food in Flames' it may be worth it to try to maintain a convoy line (with help from CW) from Australia, either through Tasman Sea/New Zealand coast (if Japan is heading west) or (more difficult) through SW Indian Ocean (if Japan is leaving India/Indian Ocean alone).

quote:

ORIGINAL: Shannon V. OKeets

Victory Cities - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
NEI:
Japanese Perimeter: Tokyo, Port Arthur, Shanghai, Taihoku

Vital Hexes - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
Japanese Perimeter:
NEI:

Vital Sea Areas - Offense
Western Americas:
Eastern Pacific:
Hawaiian Perimeter:
North Pacific:
South Pacific:
Southwestern Pacific:
NEI:
Japanese Perimeter: Sea of Japan, Japanese Coast, China Sea

As noted in above post, imho this depends on which US strategy is chosen.
I will try to split up in 5 strategies: slow or forced advance from Honolulu, slow or forced Advance from Australia, attrition against oil.

General: secure control of all cities/hexes/sea areas as listed for 'defensive' above, except where noted below.

- Slow or forced advance from Honolulu:

Slow advance: take all

If forced advance: take all quickly, except those in ( ) - leaving to later mobbing up ('44 or '45 when Japans focus is elsewhere - often they are then out-of-supply or (if lucky) left ungarrisoned (sp?) ). The list below assumes playing with naval supply units - if not, then at least one major port must be taken in each sea area, before advancing. With naval supply units it is often enough to take one port (major or minor - pick one that is lightly defended - the Japanese can't protect it all. If none available, then maybe even try to 'jump over' one sea area and advance to the next - that requires an upper hand in naval strength, though).
In all cases: if hex in ( ) only lightly or un-defended then take it.

Priority:
Western Americas, Eastern Pacific, Hawaiian Perimeter, Southwestern Pacific (or maybe South Pacific), Japanese Perimeter.
Then take the rest.

Victory Cities - Offense
North Pacific: (Dutch Harbor)
South Pacific: (Rabaul), (Pago Pago)
Southwestern Pacific: Truk (see below), (Rabaul) (Kwajalein, if at least one other minor port in Marshalls controlled).
Japanese Perimeter: (Vladivostok, Port Arthur: depending on USSR strength in Asia: leave it to them or maybe (Port Arthur) to the Chinese), (Shanghai, Taihoku) (Tokyo - often no need for invasion on Japanese homelands)
NEI: Manila or Canton, (Saigon, Singapore, Batavia).

Vital Hexes - Offense
South Pacific: Auckland, Brisbane (and other Australian Major Ports), Port Moresby (if planning to strike west against the oil), (Suva, Port Vila, Nouma)
Southwestern Pacific: one minor port at Marshall islands, Truk or Saipan, If Manilla in US hands: Legaspi.
Japanese Perimeter: when reaching this, it might be time to start mobbing up.
NEI: Manila or Canton, (Saigon, Singapore, Batavia). If lightly defended: Menodo, Sorong, Makassar. (Oil fields)

Vital Sea Areas - Offense
Hawaiian Perimeter: (Central Pacific)
South Pacific: Coral Sea (Polynesia, New Zealand Coast, Tasman Sea: only ( ) if Hawaii perimeter + Solomons Sea is secure)
Japanese Perimeter: + South China Sea, (Sea of Japan, Japanese Coast)

- Slow or forced advance from Australia:

Slow advance / forced advance: comments as above.
General: Garrison Honolulu heavily as this strategy will leave it vulnerable (2 wp corps+AA gun, FTR) and maintain supply (through Christmas Islands or Hawaiian Islands) (NAV ready to intercept is nice here).

Priority:
Western Americas, Eastern Pacific, South Pacific, NEI, then Southwestern Pacific or straight for Japanese Perimeter.
Then start mobbing up, and take other areas.

All as in the list for the above strategies, except:

Victory Cities - Offense
South Pacific: (Rabaul)
NEI: (Manila, Saigon, Singapore, Batavia)
Japanese Perimeter: Manila

Vital Hexes - Offense
South Pacific: Auckland, Brisbane (and other Australian Major Ports), Port Moresby (if reasonably easy to take - otherwise use Sarwin and ensure supply by having unit in Timor Sea and HQ in hex west of Cairns), (Suva, Port Vila, Nouma)
NEI: If lightly defended: Menodo, Makassar, (Sorong), (Oil fields)

Vital Sea Areas - Offense
Southwestern Pacific: (The Solomons) Bismarck Sea (Marianas)

Straight for the oil:

As 'slow advance from Australia', except:

Vital Hexes - Offense
NEI: Menodo, Makassar, other hexes on Celebes (allows for more LBA - primarily NAV and FTR, but maybe also a few strat bombers) (Sorong), Oil fields.
As LBA within range of South Chinese Sea is a priority:
if China reasonably secure: Nanning (or maybe even Hanoi) (place Chinese NAV or FTR here, with Stilwell also US LBA)
if India/Bay of Bengal secure: Singapore (allows for protected SUBs
and LBA)

Edit: when objective is met for this strategy (cut off Japan for supply) or when otherwise required (e.g. shifting from focus in Europe to greater focus on Japan, or Japan has been hurt enough by attrition to make US advancement feasible, then overall strategy can be shifted to 'advance from Australia'-strategy above, or, if Japan has neglected garrisoning the Pacific, maybe even make a quick redeployment of fleets and using reinforcements from the US West Coast to start a surprise-'advancement from Honolulu'-campaign).


Other strategies may be worthwhile (e.g. reconquering India/Burma or (in the end-game) China, or invasion of Japanese home islands) - in that case the above will need adjustments.

< Message edited by npilgaard -- 2/23/2008 3:19:18 PM >


_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 113
RE: AI for MWiF - USA - 2/23/2008 2:57:13 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline
quote:

ORIGINAL: brian brian

standard - leave McArthur in Manila. I never understand why the richest country in the game would want to evacuate a unit perfectly positioned to delay the enemy.


I also do that. However, some players like to have an extra 4-reorg. HQ in Europe, especially early on or when doing multiple OC impulses in a turn; or use him in Asia for naval OC, reorg. (i.e. LBA) or supply.

Edit: also, if not playing with TERR (or DIV) then McArthur is the only corps in Manilla, leaving it relatively easy for the Japanese to take it anyway.

< Message edited by npilgaard -- 2/23/2008 2:59:53 PM >


_____________________________

Regards
Nikolaj

(in reply to brian brian)
Post #: 114
RE: AI for MWiF - USA - 2/23/2008 8:42:37 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline
Nikolaj,

Excellent stuff, thanks. I'll work on integrating it into what I have already.

What you haven't seen is the time line for the US to take control of the Pacific TO. Composer99 developed that in detail and it is a later section of the Strategic plan for the US. Your comments are a welcome addition to his and I think once I integrate them they will make for a very strong strategic plan for the US in the Pacific.

My fundamental structure here is to define important hexes for the AIO to concentrate on, both when defending and when on the offense. Right now I only have 2 tiers of priority: Victory Cities and Vital Hexes. I expect the time line and milestones section of each strategic plan to give more importance to some of these. For instance, the need to take out ports when establishing a bridgehead, to take out facories when conquering a country, and to take out resources when striving to reduce enemy BPs and/or increase your own.

Do you think that Vital Hexes need to be differentiated any more? Possibilities are: priority 1 versus priority 2, to conquer of a territory, to establish/deny airbases, .., others?

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 115
RE: AI for MWiF - USA - 2/23/2008 9:33:19 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline

quote:

ORIGINAL: Shannon V. OKeets
Excellent stuff, thanks.


Thanks

quote:

ORIGINAL: Shannon V. OKeets
My fundamental structure here is to define important hexes for the AIO to concentrate on, both when defending and when on the offense. Right now I only have 2 tiers of priority: Victory Cities and Vital Hexes. I expect the time line and milestones section of each strategic plan to give more importance to some of these. For instance, the need to take out ports when establishing a bridgehead, to take out facories when conquering a country, and to take out resources when striving to reduce enemy BPs and/or increase your own.

Do you think that Vital Hexes need to be differentiated any more? Possibilities are: priority 1 versus priority 2, to conquer of a territory, to establish/deny airbases, .., others?

Yes, a differentiated priority would be nice, I think.
E.g.:
If doing a forced advance in the Pacific then victory cities are not necessarily of higher priority to conquer than e.g. some (major or naval-supply-unit-suitable) port. Later on the Victory Cities must of course get top priority.
In general, when establishing control over an area Vital Hexes (or other lightly defended, useful hexes) may be more important at first in order to get fleet/air bases etc. before going for the main objective of the Victory City.

At the end-game: Victory cities get more important than casualties/Vital Hexes: after all the winner is determined on those cities (in the last impulses: make all-out attack on those red cities).

Also, changes in he situation on board may affect priority. E.g. when trying to get upper hand in a new sea area first objective is to get airbases and then naval bases (preferably major port or naval-supply-unit minor port), and those have high priority. But when that is secured other objectives in that area (or adjacent areas) may get higher priority (other hexes/cities, hunting CONV, attrition combat etc.).

Priority may be affected by current strategic or tactical goal - e.g. if objective is to take Truk, then invading one of the adjacent island hexes (which have 0 defending notional factors if out-of-supply) might be priority before going for Truk itself, so the ARM MAR can attack from that hex and allow a blitz attack on Truk (or, if Japanese AA-gun there, then marines can participate in the attack from that hex)

For sea areas, e.g.:
- going for air superiority: air bases which allows for LBA in a high search box (which hexes suitable would depend on range of own LBA (to get most LBAs in as high a box as possible) and hexes distance to a hex dot in the sea area)
- going for naval/air superiority: as above, but if enemy has strong air presence (which requires own LBA presence), then the fleet probably won't go in a higher search box than ones own LBA anyway, meaning that a major port in the sea area in question is not such a high priority. Otoh, if the navy is to do the damage, then a high box is crucial, which often requires a major port in same or adjacent sea area (depending on speed of the task force).
- going for invasion: if tac-bomber present/needed: airbase within tac range priority - same regarding LBA FTR range. If strong CV presence, then LBA less important. If weak target or hard to capture a port, then invasion from 2-box may be an option. If wanting to invade from 3-box then a port must be taken (or available), depending on range of AMPH/TRS. Also, LBA (or CV) cover is required to reach the 2 or 3 box, if enemy forces are likely to engage.


So, maybe a dynamic priority system - or partly fixed, partly dynamic, e.g. each hex/city has a fixed priority value ('default value') and also a current value, based on the circumstances atm.
One option could be to include two sets of 'current' priorities: tactical priority (most of what is mentioned above, i.e. securing air bases before the strike against the main objective) and strategic priorities (taking that important major port). Tactical objectives would change often (depending on own and enemy forces present - e.g. loss of just a few units could change this (no need to take air base XX if the LBA planned to land there are both destroyed), range of naval units/LBA, weather, reinforcements arriving, etc.) while strategic priorities would change less often (overall strategic plan, time schedule, enemy major actions (e.g. DoW), serious losses (e.g. getting superiority or inferiority in a sea area), etc. )

Again, just some thoughts - not very structured, though

_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 116
RE: AI for MWiF - USA - 2/23/2008 9:50:51 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline
Victory Cities and Vital Hexes are not the only hexes the AIO is interested in controlling. Instead, they serve as focal points. Obviously, to take Paris you must first take hexes adjacent to Paris, and to take those hexes, you must take hexes adjacent to hexes adjacent to Paris, ...

So, if a Vital Hex serves as a focal point, then capturing hexes adjacent to it will be a priority for the AIO. Single hex islands are different of course. For multi-hex islands, looking for a good invasion hex will be considered, the closer to the focal point hex, the better. That's assuming the focal point can't be invaded directly.

I think of Victory Cities and Vital Hexes as magnets, that draw the AIO forces towards them. Once captured, the AIO moves on to the next one - with the degree of its juiciness dependent on how easy it is to capture, how important it is, and how defensable (once it is taken).

This design should work for both land and sea operations: North Africa, USSR, China, France, Pacific, Mediterranean, ...

My motivation for this design is that I think it is a reasonable facsimile of how I play WIF.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 117
RE: AI for MWiF - USA - 2/23/2008 10:03:40 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline
Sounds very good!!

I assume, that if focusing on one objective, and another suddenly presents itself as being fairly easy to get (because of enemy mistake, losses etc.), then the AI will make a (maybe temporary) shift in objective.

What about the details on how to capture the chosen objective (such things as mentioned above e.g. establish air bases, invade, get adjacent hexes etc.) - do you have a 'structure' for this as well?

Regarding the tiers of Vital Hexes/Victory cities: I think that at least two are needed: 1) must be taken before proceeding. 2) take if lightly defended/doesn't require to many resources, otherwise wait until later/until easier to get (Victory cities would eventually go from 2) to 1), as they need to be taken before the game ends ).


_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 118
RE: AI for MWiF - USA - 2/24/2008 1:22:40 AM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline

quote:

ORIGINAL: npilgaard

Sounds very good!!

I assume, that if focusing on one objective, and another suddenly presents itself as being fairly easy to get (because of enemy mistake, losses etc.), then the AI will make a (maybe temporary) shift in objective.

What about the details on how to capture the chosen objective (such things as mentioned above e.g. establish air bases, invade, get adjacent hexes etc.) - do you have a 'structure' for this as well?

Regarding the tiers of Vital Hexes/Victory cities: I think that at least two are needed: 1) must be taken before proceeding. 2) take if lightly defended/doesn't require to many resources, otherwise wait until later/until easier to get (Victory cities would eventually go from 2) to 1), as they need to be taken before the game ends ).


1 - yes, targets of opportunity will be taken advantage of. But I am especially leery of oscillating between two objectives and never taking either (as happened to a French Division between Ligny & Quatre Bras).

2 - 2 priorities it is then for Vital Hexes: high & low (I think simple labels are best).

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 119
RE: AI for MWiF - USA - 2/24/2008 4:08:48 AM   
brian brian

 

Posts: 3191
Joined: 11/16/2005
Status: offline
"Dug-Out Doug" can always go to Europe, once the Japanese forcibly relocate him to the force pool. If they don't, you've probably won on the Pacific map anyway, and Europe will be a bit easier.

(in reply to Shannon V. OKeets)
Post #: 120
Page:   <<   < prev  2 3 [4] 5 6   next >   >>
All Forums >> [New Releases from Matrix Games] >> World in Flames >> AI Opponent Discussion >> RE: AI for MWiF - USA Page: <<   < prev  2 3 [4] 5 6   next >   >>
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.984