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: Artifical Intelligence for World in Flames

 
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: Artifical Intelligence for World in Flames Page: <<   < prev  6 7 8 9 [10]
Login
Message << Older Topic   Newer Topic >>
RE: Artifical Intelligence for World in Flames - 5/25/2007 4:40:10 AM   
Mziln


Posts: 1107
Joined: 2/9/2004
From: Tulsa Oklahoma
Status: offline
I assume that the cities under French control are Lyons, Paris, Tolouse, and Vichy. 

(1) Start over with a better French player.
(2) Don't commit all your units from North Africa to France.
(3) Wait until you have a CW leader who can support an invasion of France (18.2 who co-operates/foreign troop commitment). This will probably cause the Axis to trigger Vichy.
(4) If/when the USA enters the war this will probably cause the Axis to trigger Vichy.


In the case you have given:

For the army of France organized resistance has ended it is crushed and humbled. The Axis refuses to recognize any French goverment to discuss surrender terms (or draws out the negotiations). While the French distrust the Commonwealth who in their eyes failed to come to ther assistance.

Does that sound about right?

(in reply to lomyrin)
Post #: 271
RE: Artifical Intelligence for World in Flames - 5/25/2007 12:22:17 PM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: lomyrin

About Major Power Surrenders there occasionally occurs a gamey situation in France. Germany can knock out all the Fench units but not declare Vichy nor take more than half of the factory stacks or Paris so that France cannot surrender. This can leave France in a limbo and the US entry chits for Vichy or conquests are not rolled for. It also means that the CW is limited by foreign troop commitment rules from invervening in France.

Present rules have no mechanism to deal with this situation. 

Lars

So, I guess one consequence of this is that: if France can surrender, it should, once its build points are reduced to zero and its units in Metropolitan France are, say, 3 or less? [I realize that France might not be able to surrender if they hold too many factories.]

_____________________________

Steve

Perfection is an elusive goal.

(in reply to lomyrin)
Post #: 272
RE: Artifical Intelligence for World in Flames - 5/25/2007 12:27:25 PM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: ptey

I agree that the stuka and the 2 fighters should almost always be LLs to Italy.
I think however that there are a few more details to the LND3 that a worth examining. It should be taken into consideration what LND3's got at setup. The Italian LND3 forcepool in 39 is split pretty much 50/50 between good and bad plains, so if you get 2 good ones, you should imo not build any more LND3's before you can scrap all the bad ones, which isnt untill 42. At the same time, if Germany get some bad LND3's at startup, its propably a good idea to build some LND3's since the chance of getting a good one will be quite big.
This is imo especially the case if playing with the bounce combat option, since most of the stukas have a lower A2A rating than the LND3's, the LND3 have a somewhat better chance of surviving a bounce than the stukas. So Germany should imo focus more on LND3's than LND2's when playing with bounce conbat. Even though the stukas have slightly better tac factor than the LND3's, this is imo easily offset by the better range of of the LND3. Germany really lacks airmissions, and getting the stukas in the right positions all the time can be a real pain. This is somewhat easier with the LND3's since they require fewer rebases to get into position and dont have to be repositioned all the time at the front.
However, if also playing with the oil and variable reorganisation options, LND2's only costs half the oil and half the reorg points that a LND3 costs. This means that LND2's once again become more attractive for Germany. When playing with oil the euroaxis also have less production and it will be even harder for Italy to afford any LND3's.

If however the situation is the opposite of the above (Italy gets bad LND3's from the start while Germany gets good ones). Italy should perhaps consider building a couple of LND3's, while Germany shouldnt build any before it can scrap all the bad ones in the forcepool, meaning that the JU-88A1 should be lended. However most of the time i would prefer just to build the 3 cost Navs with Italy, even if they only have a tac value of 2 while most of the LND3's have 3, they are also much better at killing british and american amphibs.

When all that is said, the most likely thing is that you get a mix of good and bad LND3's with both Germany and Italy at setup. In that situation some of the things i wrote above still holds. So depending upon options i would most likely build some German LND3's in the first half of the game, while building none with Italy. This means that i would like to have the Ju-88A1 in the German forcepool, since its a medium good plain and thus increases the chance of getting a good LND3 when building one.
Then again, if you are a gambler you can also LL to Italy and let it rot in their forcepool, since it would both increase the chance of getting a very good and a bad one. If i feel im "ahead" i wouldnt try these kind of gambles, but if you are falling behind, seeking some of these gambles (and be so lucky that they go through) may help you back on track.



I believe most of the pros and cons of this are filtered by the 2 requirements: Italy wants the air unit and is likely to spend build points on the unit type. Without the intent to build, Italy will not ask for the air unit.

Or are you suggesting that by lend leasing the unit to Italy, Germany removes a specific air unit it doesn't want from a group of air units that it wants to build; thereby making its force pool richer?

_____________________________

Steve

Perfection is an elusive goal.

(in reply to ptey)
Post #: 273
RE: Artifical Intelligence for World in Flames - 5/25/2007 4:05:34 PM   
ptey

 

Posts: 41
Joined: 9/25/2006
From: Copenhagen, Denmark
Status: offline
Most of the thoughts that i tried to convey on the subject, was naturally about if Italy wanted to build the plain or not. Apart from what i wrote, this is ofcourse also dependent on overall strategy by both the euroaxis and the allies, the losses you take, etc.
Most of the time, i dont think Italy should build that many LND3's - if any at all. But as i wrote, the 2 you get from the start of the global war scenario matters quite alot in this decision, and also the ones Germany gets.

Now, the plain in question will most likely be an average unit in the German Forcepool, so even if Italy doesnt want to build any LND3's and Germany have build a LND3 or two, but have gotten bad ones, the Ju-88A1 may now be below average. This means Germany can consider removing it by lend leasing it to Italy to improve the average quality in the LND3 pool.

(in reply to Shannon V. OKeets)
Post #: 274
RE: Artifical Intelligence for World in Flames - 6/5/2007 1:58:42 PM   
npilgaard

 

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

quote:

ORIGINAL: ptey

Most of the thoughts that i tried to convey on the subject, was naturally about if Italy wanted to build the plain or not. Apart from what i wrote, this is ofcourse also dependent on overall strategy by both the euroaxis and the allies, the losses you take, etc.
Most of the time, i dont think Italy should build that many LND3's - if any at all.


A few of factors play a role here, I think:
- if not playing with PiF there are much fewer planes, and (almost) all planes have value
- if using Italy mainly as a 'German Luftwaffe extension', using German resources to buy It ac and It doing lots of air actions (especially useful if the overall euro-axis strategy is to go after Russia in force), then lots of ac should be built (if It has a fair amount of BP and the 'primary' ac have been built already, then even the not-so-good LND3 could be built, imho - of course the poor ones will be scrapped asap).
- if not playing with pilots the each ac is more expensive, and there is not a way to choose which ac to pilot and put on map. In that case the LND3 becomes less useful.

(Again, one of the things that makes WiF a really great game is the large number of strategies and different views on what is useful to do - and most of those views actually turn out to be valid and useful, even when focusing on entirely different things )

_____________________________

Regards
Nikolaj

(in reply to ptey)
Post #: 275
RE: Artifical Intelligence for World in Flames - 6/5/2007 2:58:07 PM   
npilgaard

 

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

quote:

ORIGINAL: Shannon V. OKeets
I believe I will change from average/mean value to median value. That is, whether there are more or fewer air units in the force pool that are betterthan the lend lease air unit. When there are more, then the chances of randomly drawing an air unit that is better than the lend lease air unit is over 50%.


You have probably considered this already, but with the median value you lose info about the quality of the units.
E.g: a force pool contains a number of very poor ac, and the rest (a larger number) is of average strength. With the median value you cannot 'see' whether the poorer ac are 'very poor' or just 'a little below average'.
When playing with PiF/pilots that is not to much of a problem, since there are lots of ac anyway, and the poor ac can just be kept off map if necessary, but if playing without PiF then each ac becomes more important, and drawing a very poor ac can actually mean something.
So with median value the draw becomes more 'chancy' it seems (?).

Same problem probably arises when a major power is to determine whichs units to build. I.e. it is worth building a LND3 compared to the quality of the ac that is likely to be built.

Don't know what solution to present - maybe include both mean value and median in the considerations.

_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 276
RE: Artifical Intelligence for World in Flames - 6/5/2007 9:00:17 PM   
Shannon V. OKeets

 

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

ORIGINAL: npilgaard
quote:

ORIGINAL: Shannon V. OKeets
I believe I will change from average/mean value to median value. That is, whether there are more or fewer air units in the force pool that are betterthan the lend lease air unit. When there are more, then the chances of randomly drawing an air unit that is better than the lend lease air unit is over 50%.


You have probably considered this already, but with the median value you lose info about the quality of the units.
E.g: a force pool contains a number of very poor ac, and the rest (a larger number) is of average strength. With the median value you cannot 'see' whether the poorer ac are 'very poor' or just 'a little below average'.
When playing with PiF/pilots that is not to much of a problem, since there are lots of ac anyway, and the poor ac can just be kept off map if necessary, but if playing without PiF then each ac becomes more important, and drawing a very poor ac can actually mean something.
So with median value the draw becomes more 'chancy' it seems (?).

Same problem probably arises when a major power is to determine whichs units to build. I.e. it is worth building a LND3 compared to the quality of the ac that is likely to be built.

Don't know what solution to present - maybe include both mean value and median in the considerations.

Ok.

I'll add the condition that before asking for an air unit to be lend leased, it has to pass a certain threshold of goodness (most likely an ACV number - Air Combat Value). If it doesn't, we don't want it. It also has to be at or above the median value (ACV) of the air units of that type already in the force pool, or else it will decrease the probability of drawing the better air units. Remember, we can always ask for the air unit next turn, at which time the force pool may have changed due to builds, losses, and new arrivals.

There will also have to be some logic to handle when multiple air units of the type are being built. I.e., if we expect to build 4 F2s, then maybe asking for another F2 (or more) is warranted.

As for whether to build a unit type or not, that is based on the average/mean CV of the units of that unit type in the force pool. So if there is one really great SS armor unit sitting there plus 2 terrible armor units from Axis allies, the value of the SS unit will increase the mean/average above the value of the median unit, and therefore increase the desirability of building armor. Of course, we might also have instructions from the strategic plan to build all units of a given unit type (HQs come to mind).

_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 277
RE: Artifical Intelligence for World in Flames - 6/5/2007 10:51:48 PM   
Froonp


Posts: 7995
Joined: 10/21/2003
From: Marseilles, France
Status: offline
I saw comments about playing with or without PiF, but I seem to remember that MWiF will play with PiF all the time. Playing without PiF (or SiF) is not possible.

(in reply to Shannon V. OKeets)
Post #: 278
RE: Artifical Intelligence for World in Flames - 6/6/2007 12:03:18 AM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: Froonp

I saw comments about playing with or without PiF, but I seem to remember that MWiF will play with PiF all the time. Playing without PiF (or SiF) is not possible.

Yes. Pilots are optional, but SiF and PiF are always in effect.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Froonp)
Post #: 279
RE: Artifical Intelligence for World in Flames - 6/6/2007 2:50:18 AM   
ptey

 

Posts: 41
Joined: 9/25/2006
From: Copenhagen, Denmark
Status: offline
Here's another idea (that you probably already thought about). If fx. Germany have 1 great SS armor and 2 bad minor country armors in its forcepool, it should be considered if any of those 2 minor country armors will arrive when the minors are at the brink of being conquered by the allies. 4 turns build time can be long. The same consideration should ofcourse also be done for some majors (Italy and France mostly).

(in reply to Shannon V. OKeets)
Post #: 280
RE: Artifical Intelligence for World in Flames - 6/6/2007 3:41:04 AM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: ptey

Here's another idea (that you probably already thought about). If fx. Germany have 1 great SS armor and 2 bad minor country armors in its forcepool, it should be considered if any of those 2 minor country armors will arrive when the minors are at the brink of being conquered by the allies. 4 turns build time can be long. The same consideration should ofcourse also be done for some majors (Italy and France mostly).

A good point. I have a system in place about this for France and Itlay. I'll extend it to minor countries as well.

Another thing I have noticed I need to do in more detail is how to set up each minor country. And the set up may be different depending on the come in for the Axis or Allies.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to ptey)
Post #: 281
RE: Artifical Intelligence for World in Flames - 6/6/2007 9:31:27 AM   
Zorachus99


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

quote:

ORIGINAL: Shannon V. OKeets


quote:

ORIGINAL: ptey

Here's another idea (that you probably already thought about). If fx. Germany have 1 great SS armor and 2 bad minor country armors in its forcepool, it should be considered if any of those 2 minor country armors will arrive when the minors are at the brink of being conquered by the allies. 4 turns build time can be long. The same consideration should ofcourse also be done for some majors (Italy and France mostly).

A good point. I have a system in place about this for France and Itlay. I'll extend it to minor countries as well.

Another thing I have noticed I need to do in more detail is how to set up each minor country. And the set up may be different depending on the come in for the Axis or Allies.


A rule of thumb I often use is to not allow a minor into my forcepool as Germany unless I have plans to builds its (almost always) inferior units. Finland excepted due to the winterized value.

Also I tend to wait for Jan/Feb of a year to decare them as aligned if possible to get that free unit that comes up for most years. But that doesn't change my mind about allowing them into my forcepool and cluttering it with poor units.


_____________________________

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

(in reply to Shannon V. OKeets)
Post #: 282
RE: Artifical Intelligence for World in Flames - 6/6/2007 12:12:18 PM   
npilgaard

 

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

quote:

ORIGINAL: Shannon V. OKeets
Yes. Pilots are optional, but SiF and PiF are always in effect.

Ah, ok - thanks - that removes two of the (many, many...) variables to be taken into consideration.

(Any particular reason for this? - I can easily see that it helps regarding designing the AI, but otherwise it doesn't appear (at least on the surface ) to be much of a problem to include/exclude the PiF units, since all the non-PiF ac are in the game already.
This is not to re-open the discussion - I assume that it has been discussed at an earlier point. The reason I ask is that in the games I have played we have usually (like most others I guess) used PiF as standard, but in the latest game we are playing without PiF (we grew somewhat tired of the great fighter race, the large amount of BP spent on ac (meaning that some unit types were rarily seen in play), the large number of ac slowing things down in '44/'45, and the 'over-importance' of lba compared to CVs combined with the Air-to-Air combat systems bad design when doing large air battles with lots of ac (eg tons of Axis FTRs/NAVs in the 1/2-box in the Med during invasion of Italy, or huge stacks of CV and lba in 1-2 important sea areas near Japan in '44/'45) ), and it actually works very well with the somewhat fewer ac - matches the action limits, and each ac tend to mean more).

_____________________________

Regards
Nikolaj

(in reply to Shannon V. OKeets)
Post #: 283
RE: Artifical Intelligence for World in Flames - 6/6/2007 8:18:04 PM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: npilgaard


quote:

ORIGINAL: Shannon V. OKeets
Yes. Pilots are optional, but SiF and PiF are always in effect.

Ah, ok - thanks - that removes two of the (many, many...) variables to be taken into consideration.

(Any particular reason for this? - I can easily see that it helps regarding designing the AI, but otherwise it doesn't appear (at least on the surface ) to be much of a problem to include/exclude the PiF units, since all the non-PiF ac are in the game already.
This is not to re-open the discussion - I assume that it has been discussed at an earlier point. The reason I ask is that in the games I have played we have usually (like most others I guess) used PiF as standard, but in the latest game we are playing without PiF (we grew somewhat tired of the great fighter race, the large amount of BP spent on ac (meaning that some unit types were rarily seen in play), the large number of ac slowing things down in '44/'45, and the 'over-importance' of lba compared to CVs combined with the Air-to-Air combat systems bad design when doing large air battles with lots of ac (eg tons of Axis FTRs/NAVs in the 1/2-box in the Med during invasion of Italy, or huge stacks of CV and lba in 1-2 important sea areas near Japan in '44/'45) ), and it actually works very well with the somewhat fewer ac - matches the action limits, and each ac tend to mean more).

No particular reason. It is how CWIF was set up and I just inherited it when I started MWIF. Putting in PiF and SiF would make processing the sceanrio data more complex, and if you look at the thread on Scenario Data, you will see that it is quite complex enough already.


_____________________________

Steve

Perfection is an elusive goal.

(in reply to npilgaard)
Post #: 284
RE: Artifical Intelligence for World in Flames - 6/19/2007 9:13:31 PM   
composer99


Posts: 2923
Joined: 6/6/2005
From: Ottawa, Canada
Status: offline
quote:

ORIGINAL: Zorachus99

A rule of thumb I often use is to not allow a minor into my forcepool as Germany unless I have plans to builds its (almost always) inferior units. Finland excepted due to the winterized value.

Also I tend to wait for Jan/Feb of a year to decare them as aligned if possible to get that free unit that comes up for most years. But that doesn't change my mind about allowing them into my forcepool and cluttering it with poor units.



Germany can easily find uses for those inferior land units - anti-partisan garrison in France, Yugoslavia, Poland, USSR, and coastal defence in France, Yugoslavia, the Low Countries, maybe even Spain. Better to use these during 41-43 instead of good German units that need to be used in the USSR or whatever the primary theatre is.

Edit: And to boot, the Axis have a substantial disadvantage in land unit numbers compared to the Allies - they need to make that up any way they can.

< Message edited by composer99 -- 6/19/2007 9:14:48 PM >


_____________________________

~ Composer99

(in reply to Zorachus99)
Post #: 285
RE: Artifical Intelligence for World in Flames - 4/5/2020 2:41:05 AM   
Shannon V. OKeets

 

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

_____________________________

Steve

Perfection is an elusive goal.

(in reply to composer99)
Post #: 286
RE: Artifical Intelligence for World in Flames - 4/5/2020 2:54:13 AM   
Lothos

 

Posts: 16
Joined: 5/23/2006
Status: offline
Why the bump to such an old thread in regards to AI?

(in reply to Shannon V. OKeets)
Post #: 287
RE: Artifical Intelligence for World in Flames - 4/5/2020 3:02:01 AM   
Shannon V. OKeets

 

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

quote:

ORIGINAL: Lothos

Why the bump to such an old thread in regards to AI?

I want them to appear near or at the top in the AI Opponent Discussion forum.

_____________________________

Steve

Perfection is an elusive goal.

(in reply to Lothos)
Post #: 288
Page:   <<   < prev  6 7 8 9 [10]
All Forums >> [New Releases from Matrix Games] >> World in Flames >> AI Opponent Discussion >> RE: Artifical Intelligence for World in Flames Page: <<   < prev  6 7 8 9 [10]
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

4.421