Artifical Intelligence for World in Flames

[Shannon V. OKeets]

Thanks for the reference.

If he does show some interest, I'll make sure the NDA doesn;t fall through the cracks.

[composer99]

As I see it an Ai is indespensible.

1) It helps tutor new players.
2) An advanced played can test various strategies while making assumptions on what a human player might do. Which could be relayed to Steve for MWiF 2 .
3) If there was no Ai then no Ai would ever get better. Only by building something do we see the flaws that can then be improved upon.
4) Network down and you need a quick fix of strategy gaming.

#2 is the number one reason I want an AI on this game. It's extremely difficult to properly playtest an idea whilst playing all the powers solo. Plus I find playing against an opponent (human or AI, no matter how bad the AI) is always more edifying than playing against myself.

[Sprocc]

ORIGINAL: Neilster

This link to research at the University of Tasmania (UTAS) on Multiple Classification Ripple Down Rules (MCRDR) may come in handy when designing the AI....

http://www.comp.utas.edu.au/users/bhkang/

I'll try to get Sprocc involved again too as his research work built on MCRDR. Something went wrong with him getting his NDA and then he was so busy with his PhD that he lost interest.

Cheers, Neilster

Im about. PhD is submitted and I am no longer lecturing. Ive sent the NDA back now.

Looking at the more recent comments.
Its unlikly the AI will be better than the really good human players - but the aim should be to atleast make the game interesting for them. This should be posible through the posible permutations in the game. Sure after taking France there is only a few grand strategic choices. But which is chosen is also based on the players primary objective not only what the CW/Russian/USA players are doing. Then again its how the objective can be carried out. Yes most games are predictable but thats because most game developers arent interested in real AI. Once cheats are used they then the players only need to know those cheats.

One major advantage of MCRDR is that you can add more knowledge to the AI during the game. During development this will make it easier to improve the AI. Testers can add knowledge (There are integration issues that would need to be worked out). Possibly one of the best advantages is that you could release the AI knowledge acquisition with the game. Then individual players that find they can beat the AI can improve the AI themselves.

However implementing MCRDR is rather tricky. And at the moment there are no commercially availble implementations. My implementation from my thesis can not be used due to IP issues.

However, owning my code does not prevent me from recoding it. Which I am currently doing in my spare time. So please yell out if your interested.

Anyway - I hope to here back about the NDA I sent in.

[PanzerMike]

Currently I'm playing HOI2/DD with the HIP mod (historical improvement project). It is one of the many mods for this game. A lot of competent people are tinkering with the game's graphics, but even more importantly with the AI. The AI can be adapted because Paradox in their wisdom decided that that AI files were accesible to anyone brave enough to try to improve on them. And improved they are. The AI is much better in the mod I'm playing than in vanilla. This keeps the game alive IMO.

Will WIF also have the possibility to alter the AI by changing AI scripts ? I hope so !

[Shannon V. OKeets]

ORIGINAL: PanzerMike
Currently I'm playing HOI2/DD with the HIP mod (historical improvement project). It is one of the many mods for this game. A lot of competent people are tinkering with the game's graphics, but even more importantly with the AI. The AI can be adapted because Paradox in their wisdom decided that that AI files were accesible to anyone brave enough to try to improve on them. And improved they are. The AI is much better in the mod I'm playing than in vanilla. This keeps the game alive IMO.

Will WIF also have the possibility to alter the AI by changing AI scripts ? I hope so !

That remains to be seen. My primary objective is to create and AIO that plays well. Designing it so the player community can modify it is secondary.

[PanzerMike]

I understand your priorities. It's just that I think by doing it the HOI/DD way, the longevity of a game ican be much improved upon. I reckon a game like
WIF will be played for a long time. I know I will [;)]
 
It would be great to see the game evolve after it's initial release because the community can improve it.
 
Ofcourse I don't know the progress you have made so far on the AI and how you designed it. I realize that changing the specs in a late stadium is a nono in a project (I'm a programmer myself). But if you have not yet decided upon the issue...

[Shannon V. OKeets]

ORIGINAL: PanzerMike

I understand your priorities. It's just that I think by doing it the HOI/DD way, the longevity of a game ican be much improved upon. I reckon a game like
WIF will be played for a long time. I know I will [;)]

It would be great to see the game evolve after it's initial release because the community can improve it.

Ofcourse I don't know the progress you have made so far on the AI and how you designed it. I realize that changing the specs in a late stadium is a nono in a project (I'm a programmer myself). But if you have not yet decided upon the issue...

The complexity of WIF is much higher than any other war game I know of - and I own and have played over 100 board games. The sequence of play alone is much more detailed than most games where you mostly just move units and then attack. The AIO will have hundreds of unique rules. Designing a system so they function well becomes more complicated if the requirement to make them available for modification by a player is added.

[Zorachus99]

ORIGINAL: Shannon V. OKeets

ORIGINAL: PanzerMike

I understand your priorities. It's just that I think by doing it the HOI/DD way, the longevity of a game ican be much improved upon. I reckon a game like
WIF will be played for a long time. I know I will [;)]

It would be great to see the game evolve after it's initial release because the community can improve it.

Ofcourse I don't know the progress you have made so far on the AI and how you designed it. I realize that changing the specs in a late stadium is a nono in a project (I'm a programmer myself). But if you have not yet decided upon the issue...

The complexity of WIF is much higher than any other war game I know of - and I own and have played over 100 board games. The sequence of play alone is much more detailed than most games where you mostly just move units and then attack. The AIO will have hundreds of unique rules. Designing a system so they function well becomes more complicated if the requirement to make them available for modification by a player is added.

I know this may be the wrong time, but getting back to how the play testers are having to re-setup every time they start the campaign.

Are you interested in writing the AI routines for setup of the major and minor powers at war early?

Three reasons:

1) Players can accept default setups, modify as needed, and move on.
2) An easier part of the AI will be partially complete
3) The default setups that you define will be subject to much curiosity and a LOT of feedback concerning strategy. This by itself may be very productive for you.

Perhaps a bit too early or late?

[pad152]

Speaking of WITP, two things that help.

Randomness

One thing that helps make the AI playable is the randomness of combat, sometimes very strange things happen like a (PT boat sinks a battleship). Yes strange things did and do happen in war, sure some people complain this is impossible and should never happen. Yet it's things like this that adds that uncertianty to war not found in other war games that helps keep it fresh. You would never find this level randomness in a board or John Tiller game.


Fog of War and Combat reports

Combat reports in WITP are notorious inaccurate, this seems to drive new players crazy [:D](how come those 6 Zero's shot down all of my B-17's, this must be a bug). Sometimes it's weeks in the game before you really know if you sunk that battleship or carrier. Just like in real war, combat reports are almost always incorrect. This really adds that fog of war that's missing in all board war games and most computer war games.

These are two things that only can be done with a computer war game. To many ports of board games to computer make the mistake to only give the player the same old dice results for combat, and limit fog of war to what they can or can't see on the map. This makes a lot of combat in war games boring/dry and limits the buyers to only those true die hards of the board game.

[Shannon V. OKeets]

ORIGINAL: pad152
Speaking of WITP, two things that help.

Randomness

One thing that helps make the AI playable is the randomness of combat, sometimes very strange things happen like a (PT boat sinks a battleship). Yes strange things did and do happen in war, sure some people complain this is impossible and should never happen. Yet it's things like this that adds that uncertianty to war not found in other war games that helps keep it fresh. You would never find this level randomness in a board or John Tiller game.


Fog of War and Combat reports

Combat reports in WITP are notorious inaccurate, this seems to drive new players crazy [:D](how come those 6 Zero's shot down all of my B-17's, this must be a bug). Sometimes it's weeks in the game before you really know if you sunk that battleship or carrier. Just like in real war, combat reports are almost always incorrect. This really adds that fog of war that's missing in all board war games and most computer war games.

These are two things that only can be done with a computer war game. To many ports of board games to computer make the mistake to only give the player the same old dice results for combat, and limit fog of war to what they can or can't see on the map. This makes a lot of combat in war games boring/dry and limits the buyers to only those true die hards of the board game.

I do not want to change the CRTs from WIF FE. For a whole lot of reasons but one that stands out in my mind is my contract with Matrix Games.

Inaccurate reports to the players of combat results when the fog of war option is being used, is quite interesting though - a possibility.

[Neilster]

ORIGINAL: pad152

Speaking of WITP, two things that help.

Randomness

One thing that helps make the AI playable is the randomness of combat, sometimes very strange things happen like a (PT boat sinks a battleship). Yes strange things did and do happen in war, sure some people complain this is impossible and should never happen. Yet it's things like this that adds that uncertianty to war not found in other war games that helps keep it fresh. You would never find this level randomness in a board or John Tiller game.


Fog of War and Combat reports

Combat reports in WITP are notorious inaccurate, this seems to drive new players crazy [:D](how come those 6 Zero's shot down all of my B-17's, this must be a bug). Sometimes it's weeks in the game before you really know if you sunk that battleship or carrier. Just like in real war, combat reports are almost always incorrect. This really adds that fog of war that's missing in all board war games and most computer war games.

These are two things that only can be done with a computer war game. To many ports of board games to computer make the mistake to only give the player the same old dice results for combat, and limit fog of war to what they can or can't see on the map. This makes a lot of combat in war games boring/dry and limits the buyers to only those true die hards of the board game.

Welcome pad152!

Unusual combat results do occur in WiF. Polish fighters can defeat or destroy German ones for example and naval combat is especially unpredictable. Also, in MWiF, combat results tables with more or less randomness can be selected at the start of the game. It has to be remembered though that MWiF is a game on a very large scale and that an air counter, for example, typically represents hundreds of aircraft (not all of them the same as that depicted) and an air combat models several days (or more) of activity. Those factors will tend to smooth out results in favour of the stronger side.

For the same reason, IMHO, deliberately inaccurate combat reports don't seem appropriate. An air commander is hardly going to miss that many of his bombers got destroyed or damaged in a week long offensive against the Ruhr. Nor will a 1941 Soviet armoured corps, shattered by days of combat with German panzer forces, be mistaken for one in good condition, ready to fight.

Having said that, MWiF's scale makes FOW (Fog of War) of production and units even more interesting. There can be some very nasty surprises in store for those too adventurous because losses are often on a large scale. I think FOW will be one of the things that makes MWiF much better than cardboard WiF.

Cheers, Neilster

[amwild]

ORIGINAL: Neilster
...

For the same reason, IMHO, deliberately inaccurate combat reports don't seem appropriate. An air commander is hardly going to miss that many of his bombers got destroyed or damaged in a week long offensive against the Ruhr. Nor will a 1941 Soviet armoured corps, shattered by days of combat with German panzer forces, be mistaken for one in good condition, ready to fight.

Having said that, MWiF's scale makes FOW (Fog of War) of production and units even more interesting. There can be some very nasty surprises in store for those too adventurous because losses are often on a large scale. I think FOW will be one of the things that makes MWiF much better than cardboard WiF.

Cheers, Neilster

I agree that a player probably should not be given inaccurate information as to the status of his own forces, but Fog of War could give inaccurate information about enemy forces. If all a player's unit(s) in a single hex are destroyed by an enemy unit that was not in contact at the beginning of the phase, and there are no friendly units "in sight" (i.e. no surviving friendly units in adjacent hexes, or the surviving friendly units in adjacent hexes are in terrain with limited visibility such as forest or mountain hexes), the player may not be given totally accurate information (or even any information if the defeat was sufficiently overwhelming) as to the enemy unit(s) that did the deed. It is far more likely that this sort of inaccurate or absent data will occur in air or naval combat, but it is not impossible in land combat.

To use your examples, An allied air commander is hardly going to miss that many of his bombers got destroyed or damaged in a week long offensive against the Ruhr, but he may not know that the reason for the greater-than-expected losses are due to the introduction by the Germans of the Me-262 as a front-line unit. All that the surviving pilots may report in such a situation is that they were attacked by "Something new which is really fast and has really big guns". A 1941 Soviet armoured corps, shattered by days of combat with German panzer forces may not be able to report just which German panzer units attacked them if they were caught by surprise.

There is a situation in which a player might be given inaccurate information as to the status of his own forces - a land unit may be disrupted (but not destroyed) by combat, but may be mistakenly (until the next phase) be reported as destroyed if there are no other friendly units in direct contact. This could be accounted for as a disruption in communications leading to the presumption of the loss of the unit, until communications are restored. In WWII conditions, this sort of thing is not impossible. This isn't modern times where a soldier may be able to dig his mobile/satellite phone out of his pocket...

These sort of things couldn't happen in cardboard WiF, but a computer would make this pretty easy to achieve in MWiF.

[Neilster]

ORIGINAL: amwild

ORIGINAL: Neilster
...

For the same reason, IMHO, deliberately inaccurate combat reports don't seem appropriate. An air commander is hardly going to miss that many of his bombers got destroyed or damaged in a week long offensive against the Ruhr. Nor will a 1941 Soviet armoured corps, shattered by days of combat with German panzer forces, be mistaken for one in good condition, ready to fight.

Having said that, MWiF's scale makes FOW (Fog of War) of production and units even more interesting. There can be some very nasty surprises in store for those too adventurous because losses are often on a large scale. I think FOW will be one of the things that makes MWiF much better than cardboard WiF.

Cheers, Neilster

I agree that a player probably should not be given inaccurate information as to the status of his own forces, but Fog of War could give inaccurate information about enemy forces.

Yes, I think this should be explored.

If all a player's unit(s) in a single hex are destroyed by an enemy unit that was not in contact at the beginning of the phase, and there are no friendly units "in sight" (i.e. no surviving friendly units in adjacent hexes, or the surviving friendly units in adjacent hexes are in terrain with limited visibility such as forest or mountain hexes), the player may not be given totally accurate information (or even any information if the defeat was sufficiently overwhelming) as to the enemy unit(s) that did the deed. It is far more likely that this sort of inaccurate or absent data will occur in air or naval combat, but it is not impossible in land combat.

We have to remember that a MWiF hex represents (from memory) about 1600 square km of terrain. An infantry unit represents 3-5 divisions or perhaps 50 000 troops, with all of their command and control apparatus. I don't think the terrain of, or absence of, adjacent units is really the vital thing. All of the components of the defeated divisions are going to be sending reasonably accurate reports of the strength of enemy forces up the chain of command. Whilst these might become somewhat garbled, the high command is going to get a pretty clear idea of total enemy strength. For this reason I don't think it would ever be appropriate to have no information of enemy strength in this situation. However, perhaps only an approximation of the combat and movement factors of the enemy might be gleaned.

This ties in with an idea I've had for Fog of War, in that it might be an idea to have a level of FOW between only knowing the unit type/size and having the unit completely revealed.

To use your examples, An allied air commander is hardly going to miss that many of his bombers got destroyed or damaged in a week long offensive against the Ruhr, but he may not know that the reason for the greater-than-expected losses are due to the introduction by the Germans of the Me-262 as a front-line unit. All that the surviving pilots may report in such a situation is that they were attacked by "Something new which is really fast and has really big guns".

Well, historically the pilots said that and "They didn't have props!!!". As jets had been flying around in England for over 2 years it was pretty obvious what they were. Again though, we're talking about maybe a week of intensive air operations in which the sum total of the debriefings is going to give the brass a good approximation of enemy strength.

A 1941 Soviet armoured corps, shattered by days of combat with German panzer forces may not be able to report just which German panzer units attacked them if they were caught by surprise.

I think they would. They've been fighting them for days, they've got radios, they'll take prisoners and can read unit insignia on destroyed and captured enemy equipment.

There is a situation in which a player might be given inaccurate information as to the status of his own forces - a land unit may be disrupted (but not destroyed) by combat, but may be mistakenly (until the next phase) be reported as destroyed if there are no other friendly units in direct contact. This could be accounted for as a disruption in communications leading to the presumption of the loss of the unit, until communications are restored. In WWII conditions, this sort of thing is not impossible. This isn't modern times where a soldier may be able to dig his mobile/satellite phone out of his pocket...

No, but they did have radios and landlines. I've nothing against the principle of friendly FOW but I can't see how your example would work in practice given the game mechanics of disrupted/shattered/destroyed units. What happens to the supposedly destroyed unit? Is it removed from the map? If so who by? Can the enemy exploit into it's hex? To where does it return if it's hex is occupied?

These sort of things couldn't happen in cardboard WiF, but a computer would make this pretty easy to achieve in MWiF.

I agree that these sorts of ideas should be investigated.

Cheers, Neilster

[amwild]

My point is that there are circumstances in which complete information on enemy forces may not be available. Maybe my examples weren't the best, but the amount of information available may depend upon the degree of the victory/defeat. A swift, crushing defeat as the result of a surprise attack tends to leave the surviving people in the unit less able to report their circumstances than a prolonged marginal defeat where both sides know where the enemy is.

Opponents could be aware of a unit's type and even its name under this sort of FoW, but may not be informed of any or all of its combat values (replaced by question marks or something so the players know this isn't a glitch) - or vice versa... This may not be common, but it would be interesting on occasion.


As to my idea about a player's disrupted units "dissapearing", the computer would hide the disrupted counter (from either its player, the enemy player/AI, or possibly both) as if it was destroyed, until- say - the owning player's next phase. (I'm not exactly a WiF expert here, I've only played a few samll scenarios a decade or more ago, so I'm guessing a bit here.) The unit is still there, and if any player attempted to occupy the hex, the flipped unit would reappear immediately. The enemy would be as able to occupy the hex as they would a hex that is occupied by the flipped unit if it wasn't hidden - except that they may not be certain that it was there until they tried the move and ended up in combat. IIRC, in WWII there were occasions when units blundered into enemy units, one or neither of whom realised that the other was there until the shooting started.

Think of it like this: an infantry unit is attacked by and retreats from enemy armour, and is disrupted, losing its communications in the process. If communications cannot be restored immediately, HQ may think the unit has been destroyed, until such time that communications are restored, or another unit happens upon the unit in question. When the unit retreats, the disorder caused by loss of communications may give the enemy the illusion that they are destroying the cohesion of the unit, which may regoup unnoticed by their enemy a little while later, remaining combat-effective to some a greater degree than the enemy believes.
Naturally, a disrupted unit vanishing like this would be more likely for units that are short on heavy equipment - like infantry units or irregulars. Its harder to hide/miss armour or artillery.

This sort of missing unit FoW option would work especially well with ships - combat damage could wreck radios that might take days to repair, or the damaged ship might even just show up at the nearest friendly port. Bad weather would be a necessary factor if the ship was part of a convoy or taskforce, but even this wouldn't be impossible. The sea is a big, empty place, after all.

I don't see how aircraft could dissapear and reappear - their short fuel supply makes it far more likely that if a plane doesn't show up more or less when expected, it isn't ever going to. Pilots could be (and were) easily presumed lost and returned later.

[Froonp]

We have to remember that a MWiF hex represents (from memory) about 1600 square km of terrain. An infantry unit represents 3-5 divisions or perhaps 50 000 troops, with all of their command and control apparatus. I don't think the terrain of, or absence of, adjacent units is really the vital thing. All of the components of the defeated divisions are going to be sending reasonably accurate reports of the strength of enemy forces up the chain of command. Whilst these might become somewhat garbled, the high command is going to get a pretty clear idea of total enemy strength. For this reason I don't think it would ever be appropriate to have no information of enemy strength in this situation. However, perhaps only an approximation of the combat and movement factors of the enemy might be gleaned.

I wholeheartly agree, and that's the reason why I think FOW has nothing to do in a game of the scale of WiF.

As a side note, a WiF hex is 70-80 km across on the European Map, and about 220-240 km across on the Pacific map (from measurements I made on Google Earth and the map).

[Froonp]

Such FoW that amwild describes is good for a tactical or operationnal wargame, not for a grand strategic wargame.
The scale here is 80 km per hex, 2 months per turn (about 1-2 weeks per impulse), 300-500 planes per aircraft counter, 50000 mens per land unit.
At such a scale, FoW cannot exist.
At such a scale, all Army commanders know where the enemy armies are and who they are.
Unknow units doesn't exist, at such a scale.

[Neilster]

ORIGINAL: amwild

My point is that there are circumstances in which complete information on enemy forces may not be available. Maybe my examples weren't the best, but the amount of information available may depend upon the degree of the victory/defeat. A swift, crushing defeat as the result of a surprise attack tends to leave the surviving people in the unit less able to report their circumstances than a prolonged marginal defeat where both sides know where the enemy is.

Yes, but as I've already explained, swift, crushing defeats of the sort you describe are not what land combat in WiF models. In the several days of combat that occurs, over a front about 100km long, plenty of information about what is happening would go up the chain of command. Take the destruction of Army Group Centre in June 1944 for example. The German armies were rapidly crushed by massive and fast moving Soviet forces and although Army Group HQ was unable to effectively intervene, they were all too aware of the strength of the opposition. Overruns might be an exception but that is destruction during movement and I have already argued for possible FoW for adjacent enemy units rather than automatic revelation.

Opponents could be aware of a unit's type and even its name under this sort of FoW, but may not be informed of any or all of its combat values (replaced by question marks or something so the players know this isn't a glitch) - or vice versa... This may not be common, but it would be interesting on occasion.

I've already discussed this above.

As to my idea about a player's disrupted units "dissapearing", the computer would hide the disrupted counter (from either its player, the enemy player/AI, or possibly both) as if it was destroyed, until- say - the owning player's next phase. (I'm not exactly a WiF expert here, I've only played a few samll scenarios a decade or more ago, so I'm guessing a bit here.) The unit is still there, and if any player attempted to occupy the hex, the flipped unit would reappear immediately. The enemy would be as able to occupy the hex as they would a hex that is occupied by the flipped unit if it wasn't hidden - except that they may not be certain that it was there until they tried the move and ended up in combat. IIRC, in WWII there were occasions when units blundered into enemy units, one or neither of whom realised that the other was there until the shooting started.

The status of smaller formations can often be ambiguous or unknown but for the reasons I've stated above, the status of 50 000 troops and their associated support infrastructures is likely to be quite well known, even if it's bad news. Also, if a unit is supposedly destroyed, someone, either the phasing or non phasing player, removes it from the map. You haven't explained how this can be handled or the case of shattered units. When you talk of units blundering into one another, you again seem not to grasp the scale of WiF's land units. Corps, and even divisions don't move without dedicated recon units probing ahead and local attacks being launched to find weak points. Small units can bump into one another but not large formations.

Think of it like this: an infantry unit is attacked by and retreats from enemy armour, and is disrupted, losing its communications in the process. If communications cannot be restored immediately, HQ may think the unit has been destroyed, until such time that communications are restored, or another unit happens upon the unit in question. When the unit retreats, the disorder caused by loss of communications may give the enemy the illusion that they are destroying the cohesion of the unit, which may regoup unnoticed by their enemy a little while later, remaining combat-effective to some a greater degree than the enemy believes.
Naturally, a disrupted unit vanishing like this would be more likely for units that are short on heavy equipment - like infantry units or irregulars. Its harder to hide/miss armour or artillery.

I'm perfectly aware of what you're on about, it's just not appropriate for the massive forces represented by WiF land units. A whole corps never loses its communications. It's a massive organisation consisting of dozens of organic or attached units, each with their own radios, runners, dispatch riders, carrier pigeons and landlines. A whole corps can't regain combat effectiveness without the enemy noticing because of the multitude of sources that would be providing evidence to the contrary. Battalion after battalion would be reporting stiffening resistance or local counterattacks. Regimental HQs would be passing on these reports. Again, you're underestimating the scale of the units involved.

This sort of missing unit FoW option would work especially well with ships - combat damage could wreck radios that might take days to repair, or the damaged ship might even just show up at the nearest friendly port. Bad weather would be a necessary factor if the ship was part of a convoy or taskforce, but even this wouldn't be impossible. The sea is a big, empty place, after all.

I think there is more of a place for friendly FOW in naval combat but it still has to be remembered that the naval counters in WiF don't typically represent single ships.

I don't see how aircraft could dissapear and reappear - their short fuel supply makes it far more likely that if a plane doesn't show up more or less when expected, it isn't ever going to. Pilots could be (and were) easily presumed lost and returned later.

As I've already pointed out, air actions represent several days of action. One's own losses and serviceability would be very obvious. A handful of pilots that were lost might turn up several months later after escaping but that's not worth modeling when there are hundreds in an air unit.

Cheers, Neilster

[wfzimmerman]

ORIGINAL: Froonp

Such FoW that amwild describes is good for a tactical or operationnal wargame, not for a grand strategic wargame.
The scale here is 80 km per hex, 2 months per turn (about 1-2 weeks per impulse), 300-500 planes per aircraft counter, 50000 mens per land unit.
At such a scale, FoW cannot exist.
At such a scale, all Army commanders know where the enemy armies are and who they are.
Unknow units doesn't exist, at such a scale.

I think having exact knowledge of enemy capabilities down to the individual combat factor is unrealistic even at a strategic scale. I deplore factor-counting. I would strongly support some form of "shrouding" of capabilities of enemy units that have not yet been engaged. I.e. the unit's presence and type (inf, art, arm ...) is known, but not the precise unit type (airborne, mtn, etc.) and not the combat strength.

One interesting variation on "shrouding" might be shrouding value of new units until first encounter with enemy. For example, the USN did not fully appreciate the combat value of Japanese cruisers until the first engagements with Long Lances. Similarly, the cv of the ME262 was somewhat unknowable until the first air to air engagements occurred.

[Neilster]

ORIGINAL: Froonp

Such FoW that amwild describes is good for a tactical or operationnal wargame, not for a grand strategic wargame.
The scale here is 80 km per hex, 2 months per turn (about 1-2 weeks per impulse), 300-500 planes per aircraft counter, 50000 mens per land unit.
At such a scale, FoW cannot exist.
At such a scale, all Army commanders know where the enemy armies are and who they are.
Unknow units doesn't exist, at such a scale.

I've provided examples in the Intelligence thread of where FoW of even adjacent enemy units seems appropriate.

"What about units that are adjacent across frontiers, say, pre-Barbarossa for example? Also, sometimes armies were very good at masking their strength. I'm thinking of the Ruskies in 43-45 and the Jerries before the Ardennes Offensive."

What about units that are deep in an enemy's hinterland? The Germans were unaware of all those Siberians that turned up in front of Moscow in late 1941. What about Operation Fortitude North that convinced them that there was a whole Allied army group in East Anglia, ready to invade the Pas de Calais? Only the vaguest information would be available of many enemy units, especially when out of photo-recon range. How would OKW know the exact strength of an infantry corps in Chelyabinsk? They'd be lucky to know it was there.

Strategic deception was very important in WW2 and I definitely think there is a place for it as an optional rule in WiF. I think it really enhances the game by preventing gamey tactics and adding realism.

Cheers, Neilster

[SamuraiProgrmmr]

I am not sure where my opinion falls on this issue. It is fairly obvious that there should be decent information about army groups in a hex even if the hex is off of the front line. Which unit may have been fairly easy to figure out from intel reports.

But Fog Of War may have had an instrumental part to play in the outcome....

Massive effort was put into a deception (coded radio messages and such) that made it look like Patton was in charge of a large force that would attack German holdings. I have always heard (correct me if I am wrong) that Hitler refused to release reserves to Rommel on D-Day because he was convinced that Normandy was a feint and the real invasion would occur somewhere else (Netherlands?). I have also heard it speculated that if Rommel had gotten the reserves, D-Day may have failed to hold the beachhead.

If all of this is true (IF IF IF), then the deception surrounding Patton's command may have been instrumental in success.

Also, I have always heard that the battle of Midway was won because we broke the code and knew what was happening. If (IF IF IF) we had not, or had not trusted the code breakers, part of the fleet might have been elsewhere (Alaska?) fending off an attack from a fleet that was not there.

If so (IF IF IF), then 'fog of war' would be an interesting addition to the game. I have seen very few board games that have been able to implement this. The only substantial game that comes to mind is Victory Games' Vietnam, In it, the VC counters had a side that had no information other than 'VC might be here'. You had to attack to find out if it were a VC or a decoy.

I have even heard that during the darkest hours of the Battle of Britain, cardboard replicas of Fighters were set up near airfields to increase the enemy's estimate of air power.

I think Fog Of War in some fashion is an interesting concept to add to a computer game if for no other reason than it is generally impossible to do with a board game but easy to do with a computer game.

I think it should be tied to air superiority in an area. For example, if a bomber gets through to a hex, then certainly, there should be more information available from the hexes along the flight path.

I also think it might be interesting to spend an O-chit on deception (ala Patton).

Just ideas... What does everyone else think? Are they tenable? Are they usable? Will they destroy balance in the game?