CMx2... a little more to chew on...

[patboy]

Thanks for this very awaited reply BATTLEFRONT.

I wish you more success with CMX2 than you have with the trilogy and I expect quickly to see graphically what looks like CMX2.

Pat.

[McAuliffe]

...decorative bushes...

Oh no!...now we will have garden grogs jump in too

On a more serious note, it sounds all very promising, for me especially, the attention that will be paid to the environment and map design possibilities.

However, no word on more efficient game play. I hope they think about al those people that are suffering from wrist tendinitis as they have plotted complete armies from Normandy beaches to Moscow and from Tobruk to Monte Cassino...

[McAuliffe]

sorry, double post because I always mess up the quoting and editing icon

[Philippe]

I'm a little saddened to see the Space Lobsters receding into the distance, because that may mean field kitchens are too complicated to code. I was so looking forward to new and exciting culinary choices, like extra-terrestrial bisque.

And I really do have images of mobile bread and meat kitchens from about 1940. The germans seem to be wearing their practise uniforms -- appropriate, since chefs are supposed to wear white. The bread, however, looks very uninteresting (probably tasted that way, too.

[Philippe]

On a more serious note, that weird logistical tail that never gets modeled (except as an obscenely abstracted Supply Line) is what make the difference between play war chess and the real thing, especially if it is particularly cumbersome to move. In the time frame of a one hour action it shouldn't be movable, but in the context of a larger operation it is the kind of thing that can really show the difference between a platoon and a company. If playing at the battalion level all kinds of strange issues start coming into play that have a profound effect on a unit's ability to move around the map (where exactly did we leave the food/clothing/shelter/ammunition/wounded/gasoline/comfort girls?). Imagine if units ran out of ammunition in a really long battle as quickly as they do in CMx1, and you were dependant on horse-drawn wagons for resupply. The location of your supply dumps would suddenly become critical, and you would start paying very close attention to the road net.

I hate the fact that in most board games you can send your division spinning off in a different direction on a moment's notice. Real logistical tails are meaningless in a skirmish, but come into their own at battalion level.

[Midnight Warrior]

I understand how 1:1 LOS checks are not possible. For even if you just had 10 good guys vs 10 bad guys that would be 100 LOS checks. However, I would think that there is a good compromise solution that may be possible that might in fact turn this "bug" into a feature. The reason being is the thing that makes 1:1 LOS checks so time consuming in the game is very similar to the thing in the real world that keeps people from seeing everything instantly (i.e a person can't look everywhere at once).

For example, you could implement this by randomly picking a guy or two from each unit to do the LOS checks fro that unit. Say if you picked 2 of the 10 good guys cited above and two from the bad guys this would only require 4 LOS checks instead of the 100.

Before going to the next step let me digress on the subject of human vision for a moment. The human eye has high resolution fovial vision that only subtends a few degrees. The angular resolution of the eye really falls off outside this fovial vision. The eye scans this fovial vision over a scene much like a modern radar scans it's pencil radar beam over a search area and as such it takes time to search a wide area with the very small instantaneous fovial field of view. The eye scan is partly semi-random and partly conciously directed. Now in theory this phenomenology could be modeled as part of the LOS check mechanization. For example you could specify where a units is focusing its attention which would cause more LOS checks to be applied in that area. Some sort of a formula could be devised that would pick more good guys and more bad guys to do LOS checks (say 4 of each which would require 16 LOS checks) for targets that are near the designated search points (reflecting the fact that more Mark I eyeball search time would be being directed at that region). For other targets outside this region perhaps only one or two guys from each unit are selected which would require only 1-4 LOS checks. Since these are randomly picked over time if something is there to be seen it will eventually be be seen by somebody given it stays visable long enough. On the other hand if a target only stays visable for a short period of time it may never be seen by anybody unless the target is in a directed seach region.

Once anyone in a units spots somebody in another unit the formula might apply some form of hysterisis to model target lock on (or even fixation) so that he always does a LOS update on the guy he sees but also does a few additional LOS checks at random where the probabilities may be determined by some method that would reflect how a person might search the battlefield for targets. For example factors such as being shot at might also increase the LOS checks in there direction or if the range is short the units makes lots of LOS checks in that he can now see targets outside his fovial vision (since they subtend larger angles and thus require less resolution to see). Thus each guy in the battlefield would have a targetarea or a found target (or a few targets) that he is looking and hence gets a LOS check each periodically while at the same time he might get a few extra LOS checks to find other targets.

The effect of this is that now units are not quite so quick at seeing everything that goes on around them even if a unit is in theory out in the open and not concealed. Also, since LOS checks are a now a managed activity and have to be metered out by some formulaic method a player could now apply tactics such as deliberately exposing a unit to draw everyone's attention to it (i.e it starts hogging LOS checks) so that another group is less likley to be seen. This is something that people do in the real world and thus my thesis above that this no 1:1 LOS "bug" could be turned into a cool game feature if dealt with cleverly. And I am quite sure the folks at Battlefront.com are sufficently clever enough to do just that. So the question is would something like this be worth the time and effort to do (vs all the other things that could be done instead)?

[Michael Dorosh]

Originally posted by Midnight Warrior:

I understand how 1:1 LOS checks are not possible. For even if you just had 10 good guys vs 10 bad guys that would be 100 LOS checks. However, I would think that there is a good compromise solution that may be possible that might in fact turn this "bug" into a feature. The reason being is the thing that makes 1:1 LOS checks so time consuming in the game is very similar to the thing in the real world that keeps people from seeing everything instantly (i.e a person can't look everywhere at once).

For example, you could implement this by randomly picking a guy or two from each unit to do the LOS checks fro that unit. Say if you picked 2 of the 10 good guys cited above and two from the bad guys this would only require 4 LOS checks instead of the 100.

Before going to the next step let me digress on the subject of human vision for a moment. The human eye has high resolution fovial vision that only subtends a few degrees. The angular resolution of the eye really falls off outside this fovial vision. The eye scans this fovial vision over a scene much like a modern radar scans it's pencil radar beam over a search area and as such it takes time to search a wide area with the very small instantaneous fovial field of view. The eye scan is partly semi-random and partly conciously directed. Now in theory this phenomenology could be modeled as part of the LOS check mechanization. For example you could specify where a units is focusing its attention which would cause more LOS checks to be applied in that area. Some sort of a formula could be devised that would pick more good guys and more bad guys to do LOS checks (say 4 of each which would require 16 LOS checks) for targets that are near the designated search points (reflecting the fact that more Mark I eyeball search time would be being directed at that region). For other targets outside this region perhaps only one or two guys from each unit are selected which would require only 1-4 LOS checks. Since these are randomly picked over time if something is there to be seen it will eventually be be seen by somebody given it stays visable long enough. On the other hand if a target only stays visable for a short period of time it may never be seen by anybody unless the target is in a directed seach region.

Once anyone in a units spots somebody in another unit the formula might apply some form of hysterisis to model target lock on (or even fixation) so that he always does a LOS update on the guy he sees but also does a few additional LOS checks at random where the probabilities may be determined by some method that would reflect how a person might search the battlefield for targets. For example factors such as being shot at might also increase the LOS checks in there direction or if the range is short the units makes lots of LOS checks in that he can now see targets outside his fovial vision (since they subtend larger angles and thus require less resolution to see). Thus each guy in the battlefield would have a targetarea or a found target (or a few targets) that he is looking and hence gets a LOS check each periodically while at the same time he might get a few extra LOS checks to find other targets.

The effect of this is that now units are not quite so quick at seeing everything that goes on around them even if a unit is in theory out in the open and not concealed. Also, since LOS checks are a now a managed activity and have to be metered out by some formulaic method a player could now apply tactics such as deliberately exposing a unit to draw everyone's attention to it (i.e it starts hogging LOS checks) so that another group is less likley to be seen. This is something that people do in the real world and thus my thesis above that this no 1:1 LOS "bug" could be turned into a cool game feature if dealt with cleverly. And I am quite sure the folks at Battlefront.com are sufficently clever enough to do just that. So the question is would something like this be worth the time and effort to do (vs all the other things that could be done instead)?

That's all well and good and quite clever, but in the British and Commonwealth armies, at any rate, an infantry section was broken down such that each man in a squad file, or section of line, had specific "arcs" to cover. This was done to prevent the very thing you are asking to be simulated - ie only one or two men focussing attention in one specific direction at any one time. I remember very vividly my own basic training in 1987, and being told "WATCH YOUR ARCS".

See the article on squad tactics in ASL ANNUAL '89, specifically the article THE EVOLUTION OF SMALL UNIT TACTICS: A Historical Commentary on Squad Leader, p. 24. John Hill talks about British infantry training; the training I received in 1987 was from an infantry manual unchanged in many ways from its 1944 counterpart. The article I just mentioned talks about how if you were "Man 2" in a British squad file in Burma, you would look "left" - just the same as Man 2 in a squad file in Normandy. The training, and battle drills, were that rigid.

Rather like the fighter pilot who "checks six" or dies, the infantry section trained to keep eyes on a 180 or 360 degree arc - their lives depended on it. And their ability to do so - or not - was indicative of much training and experience the section had, both individually and collectively.

[ February 14, 2005, 08:49 PM: Message edited by: Michael Dorosh ]

[mchlstrt]

That sounds like a useful concept.

strt

[junk2drive]

How is 10 and 10 100 then 2 and 2 is 4?

[Midnight Warrior]

"That's all well and good and quite clever, but in the British and Commonwealth armies, at any rate, an infantry section was broken down such that each man in a squad file, or section of line, had specific "arcs" to cover. This was done to prevent the very thing you are asking to be simulated - ie only one or two men focussing attention in one specific direction at any one time. I remember very vividly my own basic training in 1987, and being told "WATCH YOUR ARCS"."

Your right in that the way I described this approach and the example I used would not produce the behavior you described (I just picked that one example for illustation purposes, there are many different ways to implement this concept). However, the method I described actually would enable you to model the very behavior you described. The formula could be for each guy to watch his sector as you have described (perhaps these would be to the right and left of the noninal search point). Thus each guy would do a lot of LOS checks in his sector and not many outside it. The key here is that each guy can only do so many LOS checks (both to speed up processing and to simulate his real world limitations). The formula that these are handed out can vary from doing what you describe (modeling a well trained unit) to everyone looking at the same target (green horns). That's all in the hands of the game designer. Another way to implement this is the the search sector could be tied to the firing arcs. I'll leave the details to others in that I am not trying to design the game (that BFC's job) but rather just trying to toss out some general ideas that might trigger further discussion.

Onre other thing, this would fix another problem. It always bothered me that once say a poor bazooka team gets spotted the whole German army would target him as a priority target. With this method only a small percentage of the possible shooters would even see him initially in that many of the shooters eyes would alight on other targets which would then somewhat curtail their search for jucier targets. Thus this approach would further help demolish Borg spotting.

One more thought before I leave this topic. The human eye outside its fovial region also has mnotion detectors. Thus target motion could be another factor in devising the formulas that draw one's attention (i.e up the odds of a a guy getting to make LOS check on a given target).

[Rick]

Now this is a CMx2 thread that is rolling with the good stuff.

Very excited about the story, and reduction of the "god" phenomena. I think the story is one of the most important things you're working on for increasing the interest over the long haul, and the "god" issues for improving realism.

Another thing I would like to see improved is methods to convey to the player things that he would probably know before attacking than what is in CMx1. I currently feel that we know way too little going into an assault, however meeting engagements would presumably be with very little intel.

[Jasper 2x]

It all sounds good

[Wisbech_lad]

Decorative bushes. Hm. Hope CMX2 will correctly model the speed of spring travelling northwards through Europe (about 2 mph) so that scenarios at different lattitudes will have correct blossom at correct time.

[TigerTiger]

Is it possible to the make the self-learning, so as to improve itself with everygame. You know it can be done. But are you going to do it?

[TigerTiger]

Is it possible to the make the AI self-learning, so as to improve itself with everygame. You know it can be done. But are you going to do it?

[aka_tom_w]

Originally posted by Midnight Warrior:

"That's all well and good and quite clever, but in the British and Commonwealth armies, at any rate, an infantry section was broken down such that each man in a squad file, or section of line, had specific "arcs" to cover. This was done to prevent the very thing you are asking to be simulated - ie only one or two men focussing attention in one specific direction at any one time. I remember very vividly my own basic training in 1987, and being told "WATCH YOUR ARCS"."

OK

BUT

This sure sounds like a GREAT idea and should satisfy Dorosh et.al. when this idea is combined with the Cover Arc command/limitation for LOS checks. This could also mean that if a unit has not been assigned a cover arc the game code would limit their LOS check to some random narrow area, maybe not every unit is attentive and maybe some guys are otherwise distracted cleaning or loading their weapon or whatever. Maybe if no cover arc is assigned then 1 in 10 men would NOT do ANY LOS checks, and maybe most of the other guys might look around in random area's and be not very effective, (green, POOR training) and out of 10 guys maybe one guy has the real EAGLE eye and is granted by the game code a Full Blown wide and effective LOS check. (BUT then that whole idea could be ABSTRACTED in about the exact same way it takes place in CMx1 now I guess )

"The effect of this is that now units are not quite so quick at seeing everything that goes on around them even if a unit is in theory out in the open and not concealed. Also, since LOS checks are a now a managed activity and have to be metered out by some formulaic method a player could now apply tactics such as deliberately exposing a unit to draw everyone's attention to it (i.e it starts hogging LOS checks) so that another group is less likley to be seen. This is something that people do in the real world and thus my thesis above that this no 1:1 LOS "bug" could be turned into a cool game feature if dealt with cleverly."

This idea should REALLY be incorporated into the game, combined with Relative Spotting, Unit Memory AND cover arcs, the simulation "should" be JUST about perfect with regard to spotting!!!!

I think that is NOT a 1:1 LOS "bug" and it SHOULD be turned into a cool NEW game feature if dealt with cleverly."

This means that some things will not be seen because NOT everybody is looking EVERYWHERE at the same time! That sounds like a good simulation of battlefield reality to me.

I think it is a Brilliant idea, but it might be a NIGHTMARE to code.

Fortunatly we all know Charles is up to the challenge!

What a GREAT idea!

-tom w

[ February 15, 2005, 04:11 AM: Message edited by: aka_tom_w ]

[melb_will]

Ok

I am sure that any opinion contra to CMx2 will be greated with howls of dismay. However, I think that it would be worthwile throwing out a few ideas.

Steve has said that the ideal for CMx2 will be an accurate replication of warfare. But last time I checked warfare wasn't that fun. Ppl die, squads are pinned down and can't move etc etc.

Now I can see that if Battlefront can make an accurate replication of warfare (and I obviously don't mean to imply that they are in anyway glorfiying it etc) with multiple theartres, command levels they will make an absolute mint from various defence departments around the world (cue Aussie DD)

However, this in my opinon won't make a game that will appeal to a wide majority of the market. I remember reading through posts when CMBB came out and people being turned off because it was to desolate, infantry was too realistic and going back to CMBO which despite Battlefront being on speed when they coloured the grass was a really fun game.

So what's the point. Well I am a firm supporter of a lot of things proposed for CMx2 Relative Spotting which for me has been one of the biggest problems of the other games. But at the same time, a word of caution, if we get to realistic then its no longer much fun.

Cheers

Eagerly awaiting to defend myself

Will

[aka_tom_w]

Don't worry

I think Steve summed it up BEST in my Signature Line.

What do you think?

-tom w

Originally posted by melb_will:

Ok

I am sure that any opinion contra to CMx2 will be greated with howls of dismay. However, I think that it would be worthwile throwing out a few ideas.

Steve has said that the ideal for CMx2 will be an accurate replication of warfare. But last time I checked warfare wasn't that fun. Ppl die, squads are pinned down and can't move etc etc.

Now I can see that if Battlefront can make an accurate replication of warfare (and I obviously don't mean to imply that they are in anyway glorfiying it etc) with multiple theartres, command levels they will make an absolute mint from various defence departments around the world (cue Aussie DD)

However, this in my opinon won't make a game that will appeal to a wide majority of the market. I remember reading through posts when CMBB came out and people being turned off because it was to desolate, infantry was too realistic and going back to CMBO which despite Battlefront being on speed when they coloured the grass was a really fun game.

So what's the point. Well I am a firm supporter of a lot of things proposed for CMx2 Relative Spotting which for me has been one of the biggest problems of the other games. But at the same time, a word of caution, if we get to realistic then its no longer much fun.

Cheers

Eagerly awaiting to defend myself

Will

[Philippe]

...decorative bushes...

Hmm...

I want a shrubbery.

(I won't do that again, promise).

[Vergeltungswaffe]

Originally posted by junk2drive:

How is 10 and 10 100 then 2 and 2 is 4?

10 men each doing LOS checks to 10 other men is 100 total checks. 2 men doing LOS checks to 2 other men is 4 total checks.

10x10=100

2x2=4

[imported_mike_the_wino2]

Originally posted by TigerTiger:

Is it possible to the make the AI self-learning, so as to improve itself with everygame. You know it can be done. But are you going to do it?

YOU FOOL!!!!

Haven't you seen the Terminator movies? :mad: :mad: :mad:

[tar]

Well, perhaps it would be possible to have a more flexible terrain model. The current model has elevation and terrain type separate from each other.

Increasing the separation a bit more would allow more flexibility and the addition of custom terrain for scenarios.

What I am thinking of is separating more of the terrain effects from being quite so directly linked to the underlying representation. One could have

The visual representation (what it looks like)

Trafficability (how fast you can move, separate for men, wheels, tracks)

Slope

Cover

Concealment

Weight bearing (i.e., for ice or bridges)

By varying these parameters, one could generate a lot more varied designs. The AI would need to react to the parameters rather than a more abstract tile, but I would expect you would want that to happen anyway.

===========

A lot more flexibility for victory conditions. Some ideas for this can be found in TacOps, where there are areas that must be passed through. Areas to be occuppied for certain times.

Ideally there could be objects that need to be located, or specific units/people to be captured. Time-varying victory conditions would be nice additions. This is one area where board games still have a lot more flexibility than computer games...

[kipanderson]

Hi,

All sounds good to me, lots of fun to come. Good to have all the bones in one place.

It is easy for me as I was not one of those who were unhappy about any of the bones and hints thrown by Steve before he started this thread. Once it was clear, on about the second day Steve started posting, that the focus and scale of CMX2 was going to be the same as CMX1 I was a happy chap .

As for big ideas… one day a full feature, fully functional operational game welded together with a game of the same scale and scope as CMX1/CMX2. Each of which can be played and resolved separately or one can move from one to the other. Each optimised for its own scale.

But I have already ranted on about this enough on other threads… so will not bore people any more now .

All the best,

Kip.

[Tarkus]

Kip, there was a game based on the same principle as you suggest, a dual-level game, but I'll refrain from naming it as it was sooooo badly implemented.

The main problem was that this game had nothing even remotely as interesting or efficient as a CM engine and getting down to tactical combat was mostly an erratic click fest.

Since BFC, on the other hand, have this part all figured out, maybe they could code two games, one operationnal and one tactical, and you could merge the two and exchange files to feed one another. Say you'd command a division, a brigade group or a combat command and export various tactical problems as they happens to play them in CMx2. That could be interesting for metacampaign, among other things. Sort of a wargame chest in a vertical perspective.

Sorry, I'm getting OT here.

Nonetheless, I wonder if the new engine will be usable for larger engagements up to operationnal level.

Cheers.

[Michael Dorosh]

Originally posted by Battlefront.com:

...enough people seemed to not understand what CMx2 is supposed to be, I figured I should make a single, clear statement so there will be no more confusion.

...Regardless, Charles and I aren't changing our philosophy. CMx2 is being designed from the ground up to be an improved simulation of warfare, not an improved version of Combat Mission.

Excellent! So let's sum up here what the first CMX2 release will be.

1. Not CMx1

2. A game system that is flexible enough to handle many different theaters, timeframes, and game genres without doing years of rewriting for each. This does not mean that the CMx2 codebase will automatically support things like horse cavalry or Space Lobsters, rather that it will not automatically preclude them from being added in later. The CMx1 codebase was absolutely not written this way, which is why we told you almost 2 years ago that CMAK would be the last game made using this code.

3. A graphics engine that takes advantage of today's hardware possibilities. The idea is that a more realistic looking game is a more realistic feeling game, all else being equal. When combined with #2 above it also means that internally it is a lot easier for us to make cool graphical representations than it was in CMx1 (which was a nightmare ). Therefore, the development distraction to us should be less than it was in CMx1 even though there will be a big leap forward in terms of quality. A win-win situation for everybody.

4. A game that can be played by more than 2 people, with a heavy emphasis on cooperative play (CoPlay). While we can not do this feature for the first release (not practical), the game engine itself is being coded to work with many players as soon as we can code the rather difficult technological foundation to allow such play. Think of it like CMBO not having TCP/IP functionality even though the game was written to work using this protocol.

5. The plan is for unit focus (scale) to be flexible, though the tactical focus for the first two CMx2 games is the Squad/Team just as it was for CMx1. This may or may not vary from title to title afterwards, we simply aren't planning that far ahead. Just know that in theory the CMx2 code allows us to keep things a bit flexible.

6. The command level is, like the unit focus, somewhat flexible. However, like CMx1 the first two planned games for CMx2 are Battalion/Company centric.

7. Each soldier has its own 3D representation in the game. For the first two CMx2 games there will not be 1:1 control over these soldiers, but if the scale is lowered for another game 1:1 control is possible (eg. we make a Platoon level game where you only have 30 soldiers, obviously more control is desirable). 1:1 simulation is also desired, but hardware limitations will mean some carefully implemented compromises (i.e. 1:1 LOS checks are impossible). Overall the control should be roughly the same as CMx1, but the abstractions far less.

8. WeGo is not being abandoned. In fact, there will be more options to make this system work even better.

9. Relative Spotting, as described in depth over the past couple of years, will be a huge part of the CMx2 experience. This feature can not be turned off.

10. Overcoming CMx1 "Borg" problems is a top priority to us since it is one of the biggest distractions from a realistic combat simulator. However, there is only so much we can do with this, so it isn't like we can eliminate the Borg problem. It will, however, have far less influence over games than it has in the past. Some of these features can be toggled off for those who really want more unrealistic game experiences.

11. The "God" problem, which is related to #10 but is not the same, is also something we are trying hard to knock down. The player will be able to choose how much he wants to be like a real Human commander and how much he wants to be a God.

12. A tighter focus on "story" than in past CMx1 games. A lot of the previous mentioned features will add to this, but we are putting in specific features to draw them together into a more clean message for the player. CMx1 games were sometimes described as "soulless" because of how little influence we (Battlefront) and scenario designers had over the "big picture" setting. We agree with this and therefore are putting in more tools for the scenario designers as well as us the game designers. Again, these sorts of things will vary from game release to game release, being either a more or less important part of the game.

13. Much finer detailing of terrain. This means a ton more flexibility in how maps look and how units interact with them.

14. Coupled with the above, we are including a lot more stuff for making more realistic looking scenarios. In a CMBO setting this might be stuff like telephone poles, previously wrecked tanks, far more rubble options, decorative bushes, haystacks, etc. These things may or may not have much value to the game play (wrecked tanks would, decorative bushes not so much), but the atmospheric affects will help out in a major way.

So what does this mean to us?

1. Given. Irrelevant.

2. Nice touch, but irrelevant. Doesn't give us a clue what the game will actually BE.

3. Good news indeed - but pretty is as pretty does, as they say. Does hint that at the very least scenario design will be aided by more realistic terrain modelling. But what about effects. ie - Can we have muddy terrain at lower elevations, but dry terrain higher up on a hilltop?

4. CoPlay is not anticipated for the first release, so this tells us nothing.

5. Unit focus is unchanged from CM for the initial release.

6. Command level is unchanged for the first two releases.

7. 1:1 representation, but not 1:1 control (in other words, control is unchanged from CM)

"Overall the control should be roughly the same as CMx1, but the abstractions far less."

8. "WeGo is not being abandoned." Status quo again, though with "more options to make this system work even better." No options have been discussed, so this is not really new.

9. Relative spotting is a major issue. No real idea how this will work, but certainly a major new change.

10. Borg problems will not be eliminated. "Less influence" isn't really defined here, but certainly something to be excited about.

11. The "God" problem is being investigated, but no hard news here.

12. The abandonment of Operations in favour of Campaigns. "More scenario design tools" discussed, but no hard idea of what this means.

13. More flexible maps - excellent news, coupled with

14. Pre-destroyed vehicles, telephone pole doodads, variable rubble effects. Eye candy which may or may not affect gameplay.

So, what does that give us?

Better looking maps, more flexible, with more features. Unit focus, command level, level of unit control, WEGO system all unchanged from CM. Improvements (but not abolition of) God and Borg problems. Relative spotting in. Operations out, story-driven Campaigns in.

Revolution, or evolution?