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 - France

 
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 - France Page: <<   < prev  2 3 [4] 5 6   next >   >>
Login
Message << Older Topic   Newer Topic >>
RE: AI for MWiF - France - 4/13/2007 2:59:06 AM   
Shannon V. OKeets

 

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

ORIGINAL: npilgaard
Ah, I see - sounds good. Thanks for explaining.

quote:

ORIGINAL: Shannon V. OKeets
As the front line gets closer to Paris, the value of the hexes in the front line increases. that translates as the hexes behind the front line being more valuable than the hexes in the front line itself. However, the value of front line A versus front line B, is measured by the value of all the hexes protected by the front line (inclusive).


One more small comment: As the value of the hexes increases for the hexes closer to Paris, hexes somewhat away from it have less value, meaning that front line A has a higher total value than line B, if A is closer to Paris than B. It sounds reasonable, as there are much less ground to give at line A than B, and one has less option to withdraw/reorganise (in 'normal' meaning of the word, not in 'WiF-meaning' of turning over units ) the lines.
However, in the case of France the line B should probably be chosen in most cases (unless it is a bad defensive position due to breakthrough etc.), as it is important for the French not to give ground, as opposite to eg Russia during Barbarossa.

But maybe that is already included in the algorithm , since Paris has such a high 'must be protected'-value and is located quite close to the front even when fighting in Belgium, and therefore hexes in NE-France all have a relatively high 'must-be-defended'-value as well.

Regards
Nikolaj

I guess I wasn't clear, sorry.

Individual hexes close to Paris are more valuable than those farther away, because the closer the enemy gets to Paris the sooner he will be able to capture it.

However, a front line farther from Paris is more valuable than one close to Paris. That is because a front line farther away has more valuable hexes behind it (it is protecting more hexes). You could think of this as the front line farther away 'holds' more hexes.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 91
RE: AI for MWiF - France - 4/16/2007 10:26:29 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline 		While I was in airplanes this past weekend I worked on editing the strategic plans for the major powers. I will post several items over the next couple days seeking your comments, advice and help. Here is the first.

I am standardizing the structure for the Master Production Plan, and as an example, this is what I have for France. The use of Choice A and Choice Not A are where alternatives will be available for the AIO to choose from. I am being careful to differentiate between lending BPs and Lend-Lease, which phrase I am reserving for lend-leasing air units. So far, I don't have any instructions for the French AIO about lend-leasing air units.

As you might notice in the section on the Force Pool, more editing needs to be done.

What I am proposing with this structure is to lay out in 1.1.6.1 what is available without outside help, add in outside help in 1.1.6.2, then consider unit availability in 1.1.6.3. I'll probably include buildnig units ahead in 1.1.6.3 when I get to other major powers. This prepares the groundwork setting priorities in 1.1.6.4 and ultimately the scheduling of production in 1.1.6.5. Both of the last 2 sections will be quite dyanmic and depend on what is happening on the map (e.g., losses and enemy actions), though I would hope to have the strategic plan exert an influence on production despite the changing nature of the war. The AIO should not be overly sensitive to the vicissitudes of fate.
================
1.1.6 Master production plan (MC)

1.1.6.1 Resources and Convoys

France starts with 14 factories, 11 resources, and 10 convoys. The factories are all in Metropolitan France. The resources are located:
∙ 6 in Metropolitan France.
∙ 1 Algerian that can be brought in by rail through Gibraltar and Spain.
∙ 1 Senegalese that uses 2 convoys to get to Gibraltar, then by rail into France.
∙ 1 Iraqi Oil that uses 2 convoys to get to Marseilles.
∙ 1 Indo-Chinese that uses 6 convoys to get to Marseilles.
∙ 1 New Caledonian that is not used.

This makes 9 PP, which becomes 5 BP with France’s 1939 production multiple of 0.5. In 1940, France produces 7 BP and gets to save 1 Oil per turn.

Note that this depends on France either drawing the Territorial for Indo-China, or Indo-China not coming up on the Partisan roll. Also, using French convoys in the Mediterranean. leaves them vulnerable to attack by Italy.

Choice A
If the CW uses CW convoys to bring back the French RP east of the Mediterranean, then 4 French convoys can be used as losses takers in the Bay of Biscay or the Mediterranean.

Choice Not A
Using the French convoys as losses takers is not necessarily good.

1.1.6.2 Build Points and Trade Agreements

France can't receive any loans during the first turn because they are not an active Major Power in the Lending phase.

Choice A
Lend a few BPs from CW to France in N/D 39 and J/F 40. Then lend them back to CW the following turns. This often nets a transfer of 10-15 BPs to CW before France is conquered or Vichy declared .

Choice Not A
Don't lend French BP to CW, nor vice-a-versa. Both major powers need to build at full production, all the time. The CW needs to lay down its AMPH from turn one (unless a France first happens), to have offensive capability as early as possible in 1941 against Italy.

1.1.6.3 Force Pool and Lend-Lease

If there is an AT in the force pool, build it. If there is no Anti-Tank gun to build, the decision is more complicated and revolves around Infantry gearing. An AA gun or ART is a good build but then that is all that can be built (assuming it is the first turn). Do not produce more than 1 gun. They take a long time to arrive, are expensive, and France already has several of them. Instead build MIL, GAR, INF, and/or FTRs (with pilots).

There aren't many infantry units for the French to build. To start there is 1 MTN (+ div), 1 INF DIV, 4 MIL, 1 MOT (+ DIV), and 14 territorials. With the Senegal MIL in the force pool, there are four MIL left to build (5 total originally, 2 Reserves, plus the Senegalese). Territorials are not very attractive builds. Another INF and GAR arrive in J/F 1940, but that only results in 11 infantry-type units if we count the MTN and MOT and all the divisions. The MTN and MOT take 3 turns to arrive and are too expensive to build in 1939. Also these all cost 2 or 4 BP (until the 1940 INF arrives), which makes 5 BPs hard to spend completely.

The French Infantry gearing and thus their overall build plan is closely tied to which units are in their force pool. As for complete spending in a turn, Saved Build Points, should be a mandatory optional rule when playing solitaire (i.e., against the AIO) because production and gearing decisions are hard enough already. The pace of building the MIL sets up the critical Infantry gearing. Germany often tries its best to kill French units, which may let France build 2 + 3 + 4 infantry types in the first 3 turns.

1.1.6.4 Production Priorities

[still being edited]

1.1.6.5 Scheduling Production

[still being edited]


_____________________________

Steve

Perfection is an elusive goal.

(in reply to Shannon V. OKeets)
Post #: 92
RE: AI for MWiF - France - 4/16/2007 11:24:49 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline 		You should probably include some lending of resources and/or oil from the CW to France in the production plan. These are a lot more trivial in 1939-1940 to lend than build points are, and they can prove to be invaluable to the French (extra bps to build ftrs & pilots, in particular).

I would personally recommend 2-3 resources and 1-2 oil (depending on how generous the CW is feeling, how conservatively Germany is playing, and so on) per turn from N/D 1939 until the summer of 1940 (or later, depending on how well France holds). This gives them extra oil to re-organize with, assuming playing with oil.

_____________________________

~ Composer99

(in reply to Shannon V. OKeets)
Post #: 93
RE: AI for MWiF - France - 4/16/2007 11:26:55 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline 		Looks very interesting.

quote:

ORIGINAL: Shannon V. OKeets

If there is an AT in the force pool, build it.


I think that a few things play a role on this choice:
Firstly, German strategy. If they do a France-first and attack in the fall '39, then quick troops need to be built (MIL) so a decent line can be formed asap.
If German does the standard 'attack France in spring '40 ':
I also see the AT as a nice build - if it is to be built is should be done S/O '39 so it is ready for the attack M/A.
Alternatively, as you mentioned, building GAR/INF (or even MTN - it will arrive in M/A '40 as well) to get as many units on map as possible. Build GAR/INF first and then MIL, timing arrival for J/F or M/A.
An INF-gearing of 2-3 in M/A '40 (?) will be nice, to rebuild MIL/GAR/INF.

I have seen players building a number of FTRs instead, partly for protection against ground strikes, but for the primary overall goal of using France to do as much damage to Luftwaffe as possible, since they see France as lost anyway. Depends on whether one is really trying to defend Paris 'for real' to delay Axis (the lending of BP to CW is also an issue here), or to use France to do as much damage as possible while it lasts, I guess.

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 94
RE: AI for MWiF - France - 4/17/2007 1:23:17 AM   
Shannon V. OKeets

 

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

ORIGINAL: npilgaard
Looks very interesting.

quote:

ORIGINAL: Shannon V. OKeets
If there is an AT in the force pool, build it.


I think that a few things play a role on this choice:
Firstly, German strategy. If they do a France-first and attack in the fall '39, then quick troops need to be built (MIL) so a decent line can be formed asap.
If German does the standard 'attack France in spring '40 ':
I also see the AT as a nice build - if it is to be built is should be done S/O '39 so it is ready for the attack M/A.
Alternatively, as you mentioned, building GAR/INF (or even MTN - it will arrive in M/A '40 as well) to get as many units on map as possible. Build GAR/INF first and then MIL, timing arrival for J/F or M/A.
An INF-gearing of 2-3 in M/A '40 (?) will be nice, to rebuild MIL/GAR/INF.

I have seen players building a number of FTRs instead, partly for protection against ground strikes, but for the primary overall goal of using France to do as much damage to Luftwaffe as possible, since they see France as lost anyway. Depends on whether one is really trying to defend Paris 'for real' to delay Axis (the lending of BP to CW is also an issue here), or to use France to do as much damage as possible while it lasts, I guess.

Regards
Nikolaj

Thanks. I am still thrashing out what goes where between sections 1.1.6.3/4/5. That the availability of an AT is in doubt shuold be in 1.1.6.3. That it is a good unit to build belongs in 1.1.6.4, and the scheduling of when to build it belongs in 1.1.6.5.

I am uncomfortable with you refering this strategic plan as mine. Yes, of course it is, in that I will make the final decision as to what goes into it. However, the reason this is taking me so many edit passes is that I am amalgamating all the advice that was posted to the AIO strategic forum threads. Those contains a mishmash of different points, mostly in no specific order. Once I am done, the fact that many hands were involved in the final product will be pretty obvious. Though I expect it to be crystal clear and precise, since software will have to process it as data.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 95
RE: AI for MWiF - France - 4/17/2007 1:39:22 AM   
Froonp


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

1.1.6.3 Force Pool and Lend-Lease

A MECH is also a good built in S/O 39.

(in reply to Shannon V. OKeets)
Post #: 96
RE: AI for MWiF - France - 4/18/2007 2:43:46 AM   
Incy

 

Posts: 336
Joined: 10/25/2003
Status: offline 		For the production plan, some things that should be added:

France needs to have som oil LL as soon as possible. France has HQs, air and navy that all tend to get flipped, but only one oil. France needs to be careful with oil in sep/oct +nov/dec, and needs extra oil as LL nov/dec. It's normally better to store oil in Syria rather transport it to france (IF uk maintains a cp route around africa (which I believe they always should) AND Syria is reasonably safe). 2 CP to send the oil to france is better used elsewhere.

It's common to LL CW (and US if possible) resources to france. France will often have a higher PM. If the situation dictates it, CW can always designate some recources that neither France nor the CW can recieve (the malaya resources, for example). This allows France/CW to optimise production according to PM rounding and/or developing situation (like France loosing factories due to axis advances or CW getting a dent in their CP line).

I find it a really silly idea to not plan CP lines togheter when there are several active powers from the same side (like France and CW). CP should be used where they provide the best possible overall resource (and regular) supply. France should normally use some of it's CP to transport CW resources (that might in turn be lent to France). There are only a few differences between the two countries CP's that should affect placement:
-future value (french CP are vulnerable to conquest/vichy)
-possibility for different status with respect to war with italy
Both these factors favours placing french CP mixed with UK CP in vulnerable spots in med and atlantic.

When choosing builds, as several people noted: if playing with 2d10, units that provide defensive blitz bonus should be first priority during sep/oct 39

< Message edited by Incy -- 4/18/2007 2:46:10 AM >

(in reply to Froonp)
Post #: 97
RE: AI for MWiF - France - 4/18/2007 12:13:55 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline 		Interesting points and I'll work on including them in the document I am finishing up presently.

As for treating the CW and French comvoys as interchangable, I'll visit that when I get to editing the CW strategic plan. Another difference in how convoys are are handled by CW and France is maintaining reserve convoys for replacing holes that occur in the convoy pipeline. That depends on the availability of ports (hence using CW convoys as reserves is usually better).

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Incy)
Post #: 98
RE: AI for MWiF - France - 4/19/2007 10:46:18 AM   
npilgaard

 

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

ORIGINAL: Shannon V. OKeets

I am uncomfortable with you refering this strategic plan as mine. Yes, of course it is, in that I will make the final decision as to what goes into it. However, the reason this is taking me so many edit passes is that I am amalgamating all the advice that was posted to the AIO strategic forum threads.


Yes - bas wording of mine

quote:

ORIGINAL: Shannon V. OKeets

As for treating the CW and French comvoys as interchangable, I'll visit that when I get to editing the CW strategic plan. Another difference in how convoys are are handled by CW and France is maintaining reserve convoys for replacing holes that occur in the convoy pipeline. That depends on the availability of ports (hence using CW convoys as reserves is usually better).


Also, moving convoys requires naval moves. Sometimes France can spare a naval move if it is possible to take a combined action, but often a land is needed to move units to optimize the defensive line, reorganise units, etc. CW usually take at least one naval action (or more, and/or combined actions) per turn anyway, which usually makes it easier to 'repair' the convoy lines with CW units.

Another thing to consider is where to place French naval units - the more remote French colonies have a higher chance of becoming Free French (iirc - I haven't played with RaW Vichy for quite some time).

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 99
RE: AI for MWiF - France - 4/19/2007 5:07:13 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline 		The colonies with the best chance of going Free French are in Equatorial Africa - at least, those are the ones with the best chance that also have proximity to Lend-Lease sources and the main theatre of action (Europe).

I find it's not a bad idea, when playing RAW Vichy, to start some spare convoys (if you have any), cruisers, and good subs in Equatorial Africa, (mostly) guaranteeing the Free French have a decent fleet.

_____________________________

~ Composer99

(in reply to npilgaard)
Post #: 100
RE: AI for MWiF - France - 5/8/2007 12:56:29 AM   
fallgelb

 

Posts: 28
Joined: 5/8/2007
Status: offline 		I think the French AI should not ommit the use of french units (and more important: french moves) later in the game after the conquest/vichyfying of france). Priority for cw can be to get france bps for a factory in the colonies and to deliver BP to get some TRS, NAV, TAC4, MarDiv for use vs. italy or japan.

(in reply to composer99)
Post #: 101
RE: AI for MWiF - France - 5/8/2007 4:29:50 AM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline 		Actually, the Free French are best off building land units. The CW and the Americans can build plenty of sealift. Sure, they can (and will) have big armies, too, but they have to play the odd naval, air, or combined impulse from time to time. A big Free French army can do lands every impulse, and it will be really useful once France is liberated. The odd bomber and/or fighter is a good idea, too, though.

It's all about hitting the Axis where they are most constrained - action limits!

_____________________________

~ Composer99

(in reply to fallgelb)
Post #: 102
RE: AI for MWiF - France - 5/8/2007 7:04:40 PM   
npilgaard

 

Posts: 175
Joined: 5/3/2006
Status: offline 		I think the question of which FF units to build depends on the role FF is assigned. If FF has a big army, as suggested above, then it will actually be able to make use of land actions every impulse.
However, if FF has a smaller force (e.g. 1 factory built/maybe receiving no or few BP), then imho it is better to build the specialised units, which can be of use, eg. MAR (useful to invade and retake axis controlled French areas when CW or US doing naval/combined), maybe a TRS (allows US/CW land unit to be transported during US/CW land action), and som LNDs (as suggested above - allows ground strike without expending US/CW air moves), as wel as a few land units (to fill in gaps or critical hexes when US/CW doesn't have the land moves available).
Also, if Axis is going heavily for the CW convoy lines (especially in '41/'42) then it is nice to be able to use FF naval moves to move a few SCS, if the convoy protection is not ready at the start of a turn, and CW would prefer not doing a naval as first impulse.

An alternative idea (just an idea - haven't really thought it through in detail   ): build almost exslusively FF LND's (they get some nice bombers from '43, + LL US LNDs) and HQ(s), and then have FF doing air impulses, ground striking for US/CW land attacks, and then using HQ for reorg (making FF a 'Little Italy' - doing air to help the land powers)

_____________________________

Regards
Nikolaj

(in reply to composer99)
Post #: 103
RE: AI for MWiF - France - 5/8/2007 7:28:38 PM   
Froonp


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

An alternative idea (just an idea - haven't really thought it through in detail ): build almost exslusively FF LND's (they get some nice bombers from '43, + LL US LNDs) and HQ(s), and then have FF doing air impulses, ground striking for US/CW land attacks, and then using HQ for reorg (making FF a 'Little Italy' - doing air to help the land powers)

This IMO is the far better use of FF I can think of.
Plus using the existing fleet remnants for convoy escort.
Forget about building MAR of TRS, a lone MAR is usualy useless, a TRS can be good, but it's lots of BP for something the CW & USA will have plenty of already.
A Free French Air Force is the best FF can do.

(in reply to npilgaard)
Post #: 104
RE: AI for MWiF - France - 5/8/2007 8:13:52 PM   
npilgaard

 

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

ORIGINAL: Froonp
a lone MAR is usualy useless,

I agree, that it is not worth much for 'real' invasions - but it is nice to be able to invade for free on French partisans. Also, it can be used with an division (+shore bombardment) to invade hexes with only 1 notional factor (of course CW mar can do that as well, but CW don't always have the land moves)

quote:

ORIGINAL: Froonp
a TRS can be good, but it's lots of BP for something the CW & USA will have plenty of already.

True that there are a lot of TRS - but a TRS is only slightly more expensive than a LND, and it allows for some flexibility when US/CW is doing land.

But pros and cons for both, I guess (and good thing it is so - afterall, what makes a game fun is to do tough decisions )

_____________________________

Regards
Nikolaj

(in reply to Froonp)
Post #: 105
RE: AI for MWiF - France - 5/10/2007 10:40:01 PM   
coregames


Posts: 470
Joined: 8/12/2004
Status: offline
quote:

ORIGINAL: Froonp
A Free French Air Force is the best FF can do.

There is something to be said for having at least one group of FF units that can utlilize a land action in a key theatre when the US or CW are forced to take naval or combined actions.

_____________________________

"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson

(in reply to Froonp)
Post #: 106
RE: AI for MWiF - France - 5/11/2007 5:12:06 AM   
wfzimmerman


Posts: 660
Joined: 10/22/2003
Status: offline 		Sounds like this should be one of those situations where if the FF AI has x bps, there is an 80% chance he will invest in air first priority and 20% he will invest in land first priority. the AI's builds should not be completely predictable (or, perhaps, completey rational?)  I would like to see a small chance (1-3%?) that the AI will do something loopy like build a BB, AMPH, or CV -- just to keep things fresh and keep the human player wondering -- what is the AI planning to do with that?

_____________________________

Contribute to the Steve H. thank you book! http://www.nimblebooks.com/wordpress/2009/04/contribute-to-the-wargamers-wwii-quiz-book/

(in reply to coregames)
Post #: 107
RE: AI for MWiF - France - 5/11/2007 5:50:00 AM   
Shannon V. OKeets

 

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

ORIGINAL: wfzimmerman

Sounds like this should be one of those situations where if the FF AI has x bps, there is an 80% chance he will invest in air first priority and 20% he will invest in land first priority. the AI's builds should not be completely predictable (or, perhaps, completey rational?)  I would like to see a small chance (1-3%?) that the AI will do something loopy like build a BB, AMPH, or CV -- just to keep things fresh and keep the human player wondering -- what is the AI planning to do with that?

Yes, I mostly agree. I see things as Obviously Good, Obviously Bad, and somewhere in between. The AIO will do obviously good things, without randomness distracting it. It will avoid obviously bad decisions , all the time. When things are in between, it will invoke probabilites as you described.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to wfzimmerman)
Post #: 108
RE: AI for MWiF - France - 5/11/2007 7:12:07 AM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline 		Related to the Free French AI and what it should build are the decisions that the CW and American AIs need to make vis-à-vis funding the Free French to allow them to build all these wonderful airplanes and land units.

Generally, the Free French probably can't expect much (or anything at all) in 1940-41, but the other Allies should start trying to fund them in '42, and by '43 they should be willing to fork over 5-6 bps a turn + resources (if the Free French have a factory) and oil (they need it, too). Once France is liberated, the Allies should lend bps only for rebuilding factories, because the French will have at least 6 red factories producing at 1.5 per turn (or more if Germany is still in France). But resource and oil lending must remain (even step up - if all three Western Allies can have their factories filled with resources, they will easily be able to churn out the force they need in 44-45 to race the USSR to Berlin).

Moral of the story - make the effort to build up the Free French. It will be worth it in the end.

_____________________________

~ Composer99

(in reply to Shannon V. OKeets)
Post #: 109
RE: AI for MWiF - France - 7/19/2007 10:32:11 PM   
Shannon V. OKeets

 

Posts: 22095
Joined: 5/19/2005
From: Honolulu, Hawaii
Status: offline 		Here is another attempt by me to get your help in thinking of the probability for the AIO taking each of several alternatives during play. [I tried this a couple of days ago with the timiing of the Italian DOW on France and CW.]

I am not looking for additional alternatives (though those are always welcome). Nor am I particularly interested in which you think is best ("I always ...").

What I am looking for is how to get the AIO to select which to do randomly. If it always makes the same choice, it becomes quite predictable and human opponents can do silly things, confident that the AIO will continue on its given path regardless.

I would like to favor the better choices yet still give some probability to less attractive choices. The sum of the probabilities should equal 100%. However, if you are more comfortable saying that the best choice is twice as good as the second best and the remaining choices are really bad, then I can translate that into 67%, 33% and 0% for the rest.

The 2 topics I would like your opinion on are French Convoys and French BPs to the CW. The following text is from the AIO Strategic Plan for France.
=============
1.1.6.1 Resources and Convoys
France starts with 14 factories, 11 resources, and 10 convoys. The factories are all in Metropolitan France. The resources are located:
∙ 6 in Metropolitan France.
∙ 1 Algerian that can be brought in by rail through Gibraltar and Spain.
∙ 1 Senegalese that uses 2 convoys to get to Gibraltar, then by rail into France.
∙ 1 Iraqi Oil that uses 2 convoys to get to Marseilles.
∙ 1 Indo-Chinese that uses 6 convoys to get to Marseilles.
∙ 1 New Caledonian that is not used.

This uses all 10 convoys and generates 9 PP, which become 5 BP with France’s 1939 production multiple of 0.5. In 1940, France produces 7 BP and gets to save 1 oil per turn.

Note that this depends on France either drawing the Territorial for Indo-China, or Indo-China not coming up on the Partisan roll. Also, using French convoys in the Mediterranean. leaves them vulnerable to attack by Italy.

Convoy Plan A: French Convoys for French Resources
∙ Normal placement of convoys as described in 1.1.6.1.

∙ Using the French convoys as losses takers is not necessarily good.

Convoy Plan B: French Convoys in the Atlantic
∙ Put half of the French convoys in the Bay of Biscay and the rest in the North Atlantic as the front-line of the CW convoy pipeline, letting them take losses if the Axis make a naval sortie in the first winter of the war.

∙ If the CW uses CW convoys to bring back the French RP east of the Mediterranean, then 4 French convoys can be used as losses takers in the Bay of Biscay or the Mediterranean.

Convoy Plan C: French Convoys in Mediterranean
∙ If the CW wants some convoys in the Mediterranean, they should be French.

∙ If the CW uses CW convoys to bring back the French RP east of the Mediterranean, then 4 French convoys can be used as losses takers in the Bay of Biscay or the Mediterranean.

Convoy Plan D: French Convoys in High Risk Areas
∙ Place 1 or 2 French convoys as loss takers in sea areas at risk. That usually means 2 in the Mediterranean and 5 or more in the Atlantic. This helps the survivability of British convoys and makes both the CW and France less vulnerable to an Italian DOW on just one of them.

∙ If the CW uses CW convoys to bring back the French RP east of the Mediterranean, then 4 French convoys can be used as losses takers in the Bay of Biscay or the Mediterranean.

1.1.6.2 Build Points and Trade Agreements
∙ France can't receive any loans during the first turn because they are not an active Major Power in the Lending phase.

∙ France has no trade agreements in effect.

∙ Send CW RPs to France instead of BPs, since France has unused factories. This reduces the number of convoys required, since otherwise separate convoys are needed for delivering the RP to the United Kingdom and the BP to France.

Trade Agreement Plan A
∙ Send a few RPs from CW to France in Nov/Dec 1939 and Jan/Feb 1940.

Trade Agreement Plan B
∙ Same as Plan A plus always send BPs back to CW in the following turns. This often nets a transfer of 10 to 15 BPs to CW before France is conquered or Vichy declared .

Trade Agreement Plan C∙ France may have excess BPs from Mar/Apr until conquered. This comes about if Germany attacks Belgium in Mar/Apr 1940. The excess BPs can be sent to the CW. Sending 5 BPs per turn will still leave enough BPs for France to build/rebuild MIL and INF and ensure reinforcements are in production in case Germany attempts an incomplete conquest of France.

Trade Agreement Plan D
∙ Don't lend French BP to CW, nor vice-a-versa. Both major powers need to build at full production, all the time. In particular, the CW needs to lay down its AMPH from turn one (unless a France First happens), to have offensive capability as early as possible in 1941 against Italy.
=====================

_____________________________

Steve

Perfection is an elusive goal.

(in reply to composer99)
Post #: 110
RE: AI for MWiF - France - 7/19/2007 11:56:32 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline 		My own personal assessment of the percentage chance for these two options is:

Convoy Deployment Options
Option A ("Standard") - 48%
Option B ("Atlantic") - 15%
Option C ("Med") - 15%
Option D ("High Risk") - 12%

Resource Options
Option A (CW lends resources to France) - 40%
Option B (A + France lends bps to CW from ND39) - 10%
Option C (A + France lends bps to CW from MA40) - 40%
Option D (no lending in either direction) - 10%

Edit: And if I remember to come back to this tonight or tomorrow, I'll come up with some kind of reasoning (however harebrained it is after tonight's WiF session) for these numbers.

< Message edited by composer99 -- 7/19/2007 11:58:45 PM >

_____________________________

~ Composer99

(in reply to Shannon V. OKeets)
Post #: 111
RE: AI for MWiF - France - 7/20/2007 12:44:42 AM   
Froonp


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

Convoy Deployment Options
Option A ("Standard") - 70%
Option B ("Atlantic") - 2%
Option C ("Med") - 20%
Option D ("High Risk") - 8%

Resource Options
Option A (CW lends resources to France) - 20%
Option B (A + France lends bps to CW from ND39) - 2%
Option C (A + France lends bps to CW from MA40) - 8%
Option D (no lending in either direction) - 70%

And I only put 70% to have numbers other than ridiculous for the other options, because I'd rather have put 90% for those options where I put 70%.

I too am harebrained after today's / tonight's WiF session (Tripoli is Gooooooonne !!!!! whooooowooooooooo ). Cheers !

(in reply to composer99)
Post #: 112
RE: AI for MWiF - France - 7/20/2007 5:44:43 AM   
Zorachus99


Posts: 1066
Joined: 9/15/2000
From: Palo Alto, CA
Status: offline
quote:

ORIGINAL: Shannon V. OKeets

Trade Agreement Plan A
∙ Send a few RPs from CW to France in Nov/Dec 1939 and Jan/Feb 1940.

Trade Agreement Plan B
∙ Same as Plan A plus always send BPs back to CW in the following turns. This often nets a transfer of 10 to 15 BPs to CW before France is conquered or Vichy declared .

Trade Agreement Plan C∙ France may have excess BPs from Mar/Apr until conquered. This comes about if Germany attacks Belgium in Mar/Apr 1940. The excess BPs can be sent to the CW. Sending 5 BPs per turn will still leave enough BPs for France to build/rebuild MIL and INF and ensure reinforcements are in production in case Germany attempts an incomplete conquest of France.

Trade Agreement Plan D
∙ Don't lend French BP to CW, nor vice-a-versa. Both major powers need to build at full production, all the time. In particular, the CW needs to lay down its AMPH from turn one (unless a France First happens), to have offensive capability as early as possible in 1941 against Italy.
=====================


This does miss the context of the CW strategy.

In one particular case, it misses the attempt of CW to avoid losing many convoys in the Med due to a surprize Italian convoy attack as well as lucky italian naval actions.

Sometimes britain would like to build the convoy line around africa instead of through the Med to avoid losing many in the Med. That coupled with british ASW protects convoys with less effort.

Also a reserve of convoys should always be available to the CW specifically for:
1) restoring supply
2) restoring resource routes that are not in contest

That's my quick thoughts on this...

Ciao

_____________________________

Most men can survive adversity, the true test of a man's character is power. -Abraham Lincoln

(in reply to Shannon V. OKeets)
Post #: 113
RE: AI for MWiF - France - 7/20/2007 11:34:44 AM   
fallgelb

 

Posts: 28
Joined: 5/8/2007
Status: offline 		I agree to Zorachus: global decision is necessary for CW and FR:
1. building the line through the Med (and I would suggest if then all) or (!)
2. building around africa
Mixing 1. and 2. I would use either FR or CW CONV and place only one or two per Sea Zone. If you defend the Zones then full scale.

Ressorces to France are at best delivered via Marseilles (MED Route) or via Spain (Atlantic Route). Therefore it can be very efficient to produce full with France (TA possible from ND39) and to deliver BP to CW from about JF40 or MA40.

I think the Indochina Res should go to china asap (US Entry Option). Why not taking Malaya Res for FR? And the Cypern Res is very effective for FR also. The Oil should go to Syria or egypt (TA), but can also be delivered to FR via one CONV in the North Sea (the CW often has one there to supply the BEF).

I have difficulties to answer in your system because even if we play FR/CW with two players, we coordinate our CONV lines (CW player in charge). I will try to compute an alternative system ...

(in reply to Zorachus99)
Post #: 114
RE: AI for MWiF - France - 7/20/2007 12:55:35 PM   
Shannon V. OKeets

 

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

ORIGINAL: fallgelb
I agree to Zorachus: global decision is necessary for CW and FR:
1. building the line through the Med (and I would suggest if then all) or (!)
2. building around africa
Mixing 1. and 2. I would use either FR or CW CONV and place only one or two per Sea Zone. If you defend the Zones then full scale.

Ressorces to France are at best delivered via Marseilles (MED Route) or via Spain (Atlantic Route). Therefore it can be very efficient to produce full with France (TA possible from ND39) and to deliver BP to CW from about JF40 or MA40.

I think the Indochina Res should go to china asap (US Entry Option). Why not taking Malaya Res for FR? And the Cypern Res is very effective for FR also. The Oil should go to Syria or egypt (TA), but can also be delivered to FR via one CONV in the North Sea (the CW often has one there to supply the BEF).

I have difficulties to answer in your system because even if we play FR/CW with two players, we coordinate our CONV lines (CW player in charge). I will try to compute an alternative system ...

Rather than expand the problem to include all the CW resources (and China) for maximizing production, please just consider the much smaller question I am asking.

1 - Should the French convoys be set up to take losses instead of CW convoys? If so, how vulnerable should they be? Should they always take the hits? Or only some of the hits?

2 - Should the CW send resources to the French? [If you prefer a quid-pro-quo, where, say, 2 French RP go to CW and 2 CW RP go to France, that could be an addition to the current list of alternatives.]

3 - Should the French send BPs to the English? If so, how many? Should they only be "excess BPs" ? Or a fixed quantity, even if that cuts into France's ability to build units?

Expanding the problem to optimize CW convoys and production is much more complex.

I would prefer to get a sense of how this should be played by the French AIO. Again, I am not looking for a 'best' solution. Given any 'best' solution the human opponent is likely to be able to exploit the predictability of the AIO and come up with actions that defeat it. My answer to that problem is to have variability embedded in the AIO decision making process. Whle it might look like a small change here, once I have 7 or 8 places where alternative plans are selected from randomly, the AIO will look very unpredictable to the human players.

Some decision points where alternative plans are possible (even for the French which is quite restricted in its choices) are:
1 - French convoys taking losses instead of CW convoys
2 - Trade agreements with the CW
3 - Armor defense versus massive infantry defense
4 - Leaving a INF in Africa to defend agains the Italians attacking west from Libya
5 - Sending some of the French fleet to the Pacific instead of Central West Africa
6 - Choosing Syria or French Indo-China as the new home country
7 - Placing MIL units on the Italian front instead of in the front lines against the Germans
8 - Sending part of the navy into the Baltic to attack German convoys

Many of these are questionable moves and should have a low probability attached to them. But in the aggregate, they force the human player to keep on his toes, and not be completely confident that he knows what the AIO is going to do. That is crucial in my opinion.

WIF is a very complex game and I personally do not believe there is any perfect plan for how to play any of the major powers. Each strategy or tactic has a counter strategy or tactic. Predictability is usually a fatal weakness.

To keep the AIO unpredictable, I need to assign probabilties to each of the alternative 'plans'. Now I could just do that all by my lonesome, but I think that as a group, we can come up with better numbers than I can by myself.


_____________________________

Steve

Perfection is an elusive goal.

(in reply to fallgelb)
Post #: 115
RE: AI for MWiF - France - 7/20/2007 1:04:13 PM   
Froonp


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

The Oil should go to Syria or egypt (TA), but can also be delivered to FR via one CONV in the North Sea (the CW often has one there to supply the BEF).

This needs crossing soviet territory which I believe is verboten with RAW7aug04.

The French oil to Syria, or the CW one in Egypt is a good deal though, as it can then be used for oil reorg, using infinite supply line that goes through Egypt, Sudan and all Africa to Senegal and then through secure CW convoys to France.

(in reply to fallgelb)
Post #: 116
RE: AI for MWiF - France - 7/20/2007 3:56:43 PM   
wosung

 

Posts: 692
Joined: 7/18/2005
Status: offline 		Well perhaps here again the "main contradiction" is between team-play and historical probability. To be more precisely, it's about how much taking into account that France 1940 mostly is lost cause. I've humbly choosen my side.

quote:

1 - French convoys taking losses instead of CW convoys

10%:
Proud France won't loose aginst the German Boche, so it's not very likely that proud French Merchant Marine will take losses for the Islanders.

quote:

2 - Trade agreements with the CW

10%:
1939-40 most Wally emergency trade was between US-FR and US-CW. FR and CW both badly needed strategic materials and over all had no surplus of them to trade. Instead they even then relied on the Arsenal of the Democracy.

quote:

3 - Armor defense versus massive infantry defense

20%:
French military culture over all was defensive minded and infantry/artillery/fortress based. How persuasive a tank enthusiast like De Gaulle could have been? For this Great French and for game variety at a critical point: 20% for Armor defense.

quote:

4 - Leaving a INF in Africa to defend agains the Italians attacking west from Libya

10% leaving them:
I'm not really informed about French planning with colonial troops and French perception of the Duces' plans. But probably the Duce himself wasn't either. And colonial troops are needed against the Boche. So 10% for them staying in Algerie.

quote:

5 - Sending some of the French fleet to the Pacific instead of Central West Africa

5% in the Pacific:
Highly unlikely for a variety of reasons: Distance from the Motherland, logistics, greedy Japs.

quote:

6 - Choosing Syria or French Indo-China as the new home country

5% French-Indochina:
This is far too far away from the Motherland and remember the greedy Japs already active in Southern China.

quote:

7 - Placing MIL units on the Italian front instead of in the front lines against the Germans

perhaps 30% or even more?
Main foe still is Germany. Benito is not so predictable, even was a peace advovate at Munich 38. So the MIL is not so unlikely to guard the Italian border.

quote:

8 - Sending part of the navy into the Baltic to attack German convoys

5%
It's the mighty French Navy which is to be preserved for the future (A thought quite common in the military culture of smaller navies, even like the Italian one). Besides: The British do have a navy, well, a little bit larger than the French one. They should do the job, non? Besides: Suicide naval missions, wouldn't be exactly French Savoir vivre. It sounds more like Japanese or German Savoir mourir.

Regards

(in reply to Froonp)
Post #: 117
RE: AI for MWiF - France - 7/20/2007 5:51:53 PM   
fallgelb

 

Posts: 28
Joined: 5/8/2007
Status: offline 		To Froonp: Oops, always played the wrong way. "Verboten", you're right. But it was so tricky...

To Shannon:

I think (incorpotrating the problem of unpredictable AI) it is in any case right to decide, if there should be
1. A CW/FR convoy chain around africa without a single CONV in the Med (60%)
2. A CW/FR convoy chain around africa with 1 or 2 CONV in WMS, EMS (20%)
3. A CW/FR convoy chain through the Med with only necessary CONVs for South Africa and Senegal outside and a good air defense in the Med (20%)

Your Questions considering FR CONV:
A (Standard): 50%
B (Atlantik): 5%;
C (Med): 5%
D (High Risk): 40% (Cape St. Vincent)

Your Questions considering FR/CW Trade agreement:
A: (CW lends ressources to FR) 100% (because of lesser CONV needed to transport the ressources to spain)
B: (FR lends BP to CW from N/D39) 50%
C: (FR lends BP to CW from M/A40) 50%
D: no lending 0%

1. "Should the French CONV take hits?": yes, all that cannot rebase and about halve of that which can rebase to FR Equ Africa because even FR Equ Africa can get Vichy and if you have 10 Vichy CONV there because you've lost 10 CW CONV you have gained nothing (for example: losing 6, FR Equ Africa 4, Neucaledonia 1); each vichyfied CONV can result in one additional BP for germany/turn or in the ability to reoil vichified Ships overseas
2. "Should the CW send ressources to FR?": yes, via Spain or Marseilles, up to 6 (unpredictability = quantitative) 
3. "Should the FR send BP to CW?": yes (25% from M/A40 all, 25% from M/A almost all, 25% from N/D excess, 25% from N/D fixed)

For excample Option D  above (no lending between FR and CW) makes no sence, even if it's unpredictable. You can get more than 15 Ressources to Europe and FR needs help vs. GE. There is no gain for the Axis if it knows that there is lending underway (at least i cant't see the point).

(in reply to Froonp)
Post #: 118
RE: AI for MWiF - France - 7/20/2007 8:55:51 PM   
Shannon V. OKeets

 

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

ORIGINAL: fallgelb

To Froonp: Oops, always played the wrong way. "Verboten", you're right. But it was so tricky...

To Shannon:

I think (incorpotrating the problem of unpredictable AI) it is in any case right to decide, if there should be
1. A CW/FR convoy chain around africa without a single CONV in the Med (60%)
2. A CW/FR convoy chain around africa with 1 or 2 CONV in WMS, EMS (20%)
3. A CW/FR convoy chain through the Med with only necessary CONVs for South Africa and Senegal outside and a good air defense in the Med (20%)

Your Questions considering FR CONV:
A (Standard): 50%
B (Atlantik): 5%;
C (Med): 5%
D (High Risk): 40% (Cape St. Vincent)

Your Questions considering FR/CW Trade agreement:
A: (CW lends ressources to FR) 100% (because of lesser CONV needed to transport the ressources to spain)
B: (FR lends BP to CW from N/D39) 50%
C: (FR lends BP to CW from M/A40) 50%
D: no lending 0%

1. "Should the French CONV take hits?": yes, all that cannot rebase and about halve of that which can rebase to FR Equ Africa because even FR Equ Africa can get Vichy and if you have 10 Vichy CONV there because you've lost 10 CW CONV you have gained nothing (for example: losing 6, FR Equ Africa 4, Neucaledonia 1); each vichyfied CONV can result in one additional BP for germany/turn or in the ability to reoil vichified Ships overseas
2. "Should the CW send ressources to FR?": yes, via Spain or Marseilles, up to 6 (unpredictability = quantitative) 
3. "Should the FR send BP to CW?": yes (25% from M/A40 all, 25% from M/A almost all, 25% from N/D excess, 25% from N/D fixed)

For excample Option D  above (no lending between FR and CW) makes no sence, even if it's unpredictable. You can get more than 15 Ressources to Europe and FR needs help vs. GE. There is no gain for the Axis if it knows that there is lending underway (at least i cant't see the point).

Terrific. Thank you very much.

I will get to the question of CW convoys, but that is a ways down the road. When I do, I will use what you mentioned here (in fact I will add it to the comments on the CW strategic plan now).

I have put finishing the French Strategic Plan ahead of strategic plans for the other major powers, though I am working on them in parallel to some degree. The Italian strategic plan in second in line, for a variety of reasons, mostly due to me wanting to finish both an Allied major power and an Axis major power; and that the Italians & French strategic plans interact to some degree. It is very difficult to separate the strategic plan for each major power from what the other major powers on their side are doing, but I want to keep the AIO from being a monolithic decision maker.

My purpose in completing a strategic plan is to then be able to complete the design of the data structure that will hold strategic plans. These puppies will be read in from a data file (since I am starting from scratch writing that code is much easier). I made a good start on designing the data structures while in Europe, but I haven't typed them in yet. My work on the AIO consists mainly of doing paper edits during breakfast and lunch, followed by typing in the changes later in the day. Though from a forum reader's perspective it might seem like it is a high priority for me, it isn't at this moment.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to fallgelb)
Post #: 119
RE: AI for MWiF - France - 7/20/2007 9:14:26 PM   
Shannon V. OKeets

 

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

ORIGINAL: wosung
Well perhaps here again the "main contradiction" is between team-play and historical probability. To be more precisely, it's about how much taking into account that France 1940 mostly is lost cause. I've humbly choosen my side.

quote:

1 - French convoys taking losses instead of CW convoys

10%:
Proud France won't loose aginst the German Boche, so it's not very likely that proud French Merchant Marine will take losses for the Islanders.

quote:

2 - Trade agreements with the CW

10%:
1939-40 most Wally emergency trade was between US-FR and US-CW. FR and CW both badly needed strategic materials and over all had no surplus of them to trade. Instead they even then relied on the Arsenal of the Democracy.

quote:

3 - Armor defense versus massive infantry defense

20%:
French military culture over all was defensive minded and infantry/artillery/fortress based. How persuasive a tank enthusiast like De Gaulle could have been? For this Great French and for game variety at a critical point: 20% for Armor defense.

quote:

4 - Leaving a INF in Africa to defend agains the Italians attacking west from Libya

10% leaving them:
I'm not really informed about French planning with colonial troops and French perception of the Duces' plans. But probably the Duce himself wasn't either. And colonial troops are needed against the Boche. So 10% for them staying in Algerie.

quote:

5 - Sending some of the French fleet to the Pacific instead of Central West Africa

5% in the Pacific:
Highly unlikely for a variety of reasons: Distance from the Motherland, logistics, greedy Japs.

quote:

6 - Choosing Syria or French Indo-China as the new home country

5% French-Indochina:
This is far too far away from the Motherland and remember the greedy Japs already active in Southern China.

quote:

7 - Placing MIL units on the Italian front instead of in the front lines against the Germans

perhaps 30% or even more?
Main foe still is Germany. Benito is not so predictable, even was a peace advovate at Munich 38. So the MIL is not so unlikely to guard the Italian border.

quote:

8 - Sending part of the navy into the Baltic to attack German convoys

5%
It's the mighty French Navy which is to be preserved for the future (A thought quite common in the military culture of smaller navies, even like the Italian one). Besides: The British do have a navy, well, a little bit larger than the French one. They should do the job, non? Besides: Suicide naval missions, wouldn't be exactly French Savoir vivre. It sounds more like Japanese or German Savoir mourir.

Regards

Most of these I have already decided on for probabilities; some of my decisions are based on lively discussions in this thread previously.

The choice of armor versus infantry will be driven by whether there is a France First strategy by Germany of not. When there is, Infantry is chosen. When there isn't a France First stratgy, then Armor is chosen, though that might be tempered to only 1 unit (AT or MECH) instead of 2 (AT and MECH). If the CW provides extra RPs, then both will be build (assuming th AT in in the force pool).

I am of the opinion that 50% is a better number for leaving the the INF in north Africa. That makes Italy's decision (played by a human) to attack west from Libya not something he can depend on succeeding (when there is no French opposition) or rule out completely (1 French infantry supported by the navy and may long range air is a tough nut for Italy, while holding off the CW to the east.). Italy has to want and see what France does, just as when playing against a human. Italy can't plan how he sets up his units with any certainty as to which way to go.

Yes 5% for the Pacific and Baltic choices is about right. I might make it lower, or choose one of the two @ 5% but not both.

I have 6% for Syria and 3% for French Indo-China. Some of the support for French Indo-China was based on the WIF FE Pacific map. With the unified scale, taking out French Indo-China is easier for the Japanese (personally I don't think it was ever very difficult). These percentages will turn out to be lower in practice, since sometimes those countries go Vichy.

The MIL on the Italian front will be driven by the strength of the MIL (weaker ones preferred versus Italy), where the MIL come in (Paris MIL is very convenient for fighting the Germans), and some timing considerations. If Paris is going to fall anyway, then placing the MIL against Italy and cycling regular army units to the north (say to the Maginot line) is good for when Vichy France is declared.

My real puzzlement concerned the French convoys and the CW-France trade agreements.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to wosung)
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 - France 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

1.031