Why infantry combat in CMx2 is so different

[LongLeftFlank]

What was that line (likely borrowed) in the old SL rulebook? The difference between a hero and a fool is all too frequently determined by a well-aimed shot.

[Dietrich]

I'm not making any conclusions, nor do i talk about any "hero code", im just talking about individuality. And how much immersion adds the fact that soldiers aren't robots and can act differently under same circumstances.

It's not at all unusual to see most of a squad cowering in their corners while one guy continues to engage the enemy. And conversely it is not unusual to see a few guys in the squad who are more prone to hiding in combat.

What was that line (likely borrowed) in the old SL rulebook? The difference between a hero and a fool is all too frequently determined by a well-aimed shot.

As per a post by Meach at little under two years ago:

Just had a battle and a US infantry marksman with a scoped M4 stood up, totally ignored incoming AK and SVD rounds, aimed for about 5 seconds and put a round right thru the throat of the enemy sniper. The rest of his squad were in the fetal position praying and wishing they joined the admin section.

This guy must have had ice-water in his veins. I honestly thought 'he's so gonna get lead poisoning any second now" but no. I hasten to add the shot was at 319m. After he took the shot he calmly took a knee and reloaded. His comment to his buddies would probably been along the lines of "Fk'n, pussies" I was totally immersed and was watching zoomed in. I swear I was holding my breath as he squeezed off his shot.

So yeah, there's already provision (in CMx2) for hero-ish moments/actions.

[Battlefront.com]

As others have said, there's plenty of variety. Individuals can also be branded Fanatical, similar to CMx1. These guys get a Morale bonus which is, like CMx1, hidden from the player so you don't misuse/abuse the unit the particular soldier is in.

I think it would be kinda cool to somehow track individual soldier feats of heroism and indicate that, somehow, in the user interface so the player could notice and remember who did something extreme and out of the ordinary. Perhaps we will someday add this to the game as it isn't in conflict with anything CM stands for. However, it will require a bit of work to get this sort of thing to function since, currently, there is no sense of tracking individual accomplishments nor evaluating them.

Steve

[Michael Emrys]

Its not just that there is inherent randomness, there is so much inherent complexity it results in the appearance of randomness. Which is much more in line with how the real world works.

Sounds really good. What you are describing—as perhaps you know—is what is included in what used to be called Chaos Theory and now goes by the name of Complexity Theory, a fascinating subject in its own right. I've been hoping to see this ever since CM was first being bruited about. I definitely think it is the way to go in wargames.

Michael

[dan/california]

My basic definition is any system where the tiniest difference in initial conditions can give completely divergent outputs over time.

[Michael Emrys]

That's a pretty good definition of it.

Michael

[Battlefront.com]

Chaos Theory is at the heart of the concept of an engineered system. The basic underlying theory is if you can get all the little elements moving around on their own, within realistic parameters, then you get more natural outcomes that are both richer in variety as well as character. Games which are based on the designed for effect method tend to purposefully suppress Chaos Theory because unpredicted results in one area, no matter how correct, may corrupt the effects in another part of the simulation.

Not to say that an engineered sim doesn't also have to suppress Chaos, because no simulation can't perfectly simulate the basic elements and their effects. But like so many things in life, the closer you are to the idea the better the results when compared to things which are further from the ideal.

Steve

[MarkEzra]

Chaos Theory is at the heart of the concept of an engineered system. The basic underlying theory is if you can get all the little elements moving around on their own, within realistic parameters, then you get more natural outcomes that are both richer in variety as well as character.

Steve

Question: Isn't this a good example of why "Triggers" in CMx2 should be approached with caution?

[Battlefront.com]

Triggers, for those who don't know, are a concept of tying the AI to do some sort of action when the Human Player does something specific. The most common example is a unit entering a specific spot, but it doesn't have to be limited to that sort of action.

Triggers do, in fact, direct the AI in a way that is extremely predictable based on whatever the parameters and reactions are. So yes, Triggers are very anti-chaos if you like. The problem is there isn't a viable alternative to giving the AI fairly strict, inflexible instructions. To do otherwise would require several AI programmers, probably with PhDs and 20 years of practical experience, coding for several years to get even moderately acceptable results. Since that's not going to happen we have to go with more traditional game AI programming strategies of laying down "breadcrumbs" for the AI to follow through the woods.

Steve

[Bertram]

Triggers, for those who don't know, are a concept of tying the AI to do some sort of action when the Human Player does something specific. The most common example is a unit entering a specific spot, but it doesn't have to be limited to that sort of action.

Triggers do, in fact, direct the AI in a way that is extremely predictable based on whatever the parameters and reactions are. So yes, Triggers are very anti-chaos if you like. The problem is there isn't a viable alternative to giving the AI fairly strict, inflexible instructions. To do otherwise would require several AI programmers, probably with PhDs and 20 years of practical experience, coding for several years to get even moderately acceptable results. Since that's not going to happen we have to go with more traditional game AI programming strategies of laying down "breadcrumbs" for the AI to follow through the woods.

Steve

More general, every AI uses triggers, as do humans (and animals) . The trouble with the AI is describing the triggers in such a way that they are general usefull. An animal will let another anmimal come within a certain distance, before fighting or fleeing. The reaction depends on the animal in question, the size and kind of anaimal approaching and the circumstances as terrain, etc. A human, playing CM, will let the enemy come within a certain distance before retreating (or opening fire). The distance being dependend on the size of the enemy and the kind of units, as well as the units the player has, the terrain, victory conditions etc.

A good AI would "recognize" (that is: would be programmed to take in consideration) all the relevant variables for a decision like that - and would be able to gather data on those variables (maybe an even more complex problem). For CM even the first would be rather complex to do.

A shortcut is to make a system in which the designer of the scenario defines the triggers, and orders the reaction. You sidestep quite a bit of the AI complexity needed, by letting the scenario designer decide which triggers will be usefull in the scenario - the designer knows the terrain, the forces involved, and can guess the things a human will try. Of course you (as developer) wil have to decide twhich trigger conditions to offer, and how to implement them so that the system recognizes the values for those triggers. The more elaborate the available trigger system, the more choices the scenario designer has. And the more chance he has to make the computer controlled side act realistic.

The flip side is that the more elaborate the system is, the more complex the setting up of a computer controlled side is - the scenario designer has to have more knowledge of the system, and he has to put more time in the the computer contolled opponent.

The Decisive Battles system of the Strategic Studies Group is a nice example of this. Their first engine was relative simple, and the AI was easily tricked by a human. Fighting against the AI got stale very quick. On the other hand, lots of player made scenario's were available, and there were lots of options for a PBEM game. Thier latest iteration of the engine revamped the AI. It works much better -though the AI is still no fun when you have mastered the system (in my opinion). The negative side is that there are almost no player made scenario's anymore - the AI is just to complex deveop one. So PBEM for this system is down to the developer released scenario's - and in this case that seems not to be enough - their site is almost dead.

[Alchenar]

As others have said, there's plenty of variety. Individuals can also be branded Fanatical, similar to CMx1. These guys get a Morale bonus which is, like CMx1, hidden from the player so you don't misuse/abuse the unit the particular soldier is in.

I think it would be kinda cool to somehow track individual soldier feats of heroism and indicate that, somehow, in the user interface so the player could notice and remember who did something extreme and out of the ordinary. Perhaps we will someday add this to the game as it isn't in conflict with anything CM stands for. However, it will require a bit of work to get this sort of thing to function since, currently, there is no sense of tracking individual accomplishments nor evaluating them.

Steve

I'm sure this has been asked before but: why doesn't CMx2 track individual kills?

It wasn't just fun, it was quite useful to come to the end of a scenario in CMx1 and discover that one unit in a particular place or used in a particular way was far more effective that you realised when playing.

[LongLeftFlank]

The triggers in CMx1 were predictable because the flags were readily visible and known to both sides; take a flag and you could expect the AI to arrive promptly.

Consider instead if you had hidden flags (or zones), buried like landmines and not known to the player. These may be, but don't have to be obvious -- that's up to the cunning scenario designer.

Perhaps each Trigger zone has its own AI Group -- that is the only Group programmed to react to the enemy entering the zone (with Vehicles / with Infantry or Vehicles?).

The AI Group's reaction doesn't have to be simple knee-jerk "recapture the flag" assault. Its actual order set is programmed separately using the CMx2 method, and could be quite elaborate, or even have nothing to do with recapturing that area (e.g. withdraw if you lose the hill)

It just won't Exit Setup until the Trigger is hit (as opposed to at a specific BEFORE/AFTER time).

Programming spaghetti I know. But well worth it IMHO.

[Battlefront.com]

CM:BN has unit level kill stats, similar to CMx1. It wasn't in CM:SF simply because we didn't have a chance to code it due limited time.

Technically, an Objective is not a Trigger. However, inherently Objectives are a goal for either a Human or the AI to attempt to secure, while a Trigger is something that evokes a certain behavior from the AI only.

We have an excellent AI improvement blueprint already in hand. We're simply waiting for the chance to implement it. Phil, our First Second Programmer, has a lot of AI coding experience to help out when we get to it.

Steve

[dan/california]

Any feel yet if we will see it before the Eastern front Game? If Normandy is half as good as the AAR indicates, either AI or better multiplayer/Coplay would seem to be nearing the top of the pile.

[Chops]

As others have said, there's plenty of variety. Individuals can also be branded Fanatical, similar to CMx1. These guys get a Morale bonus which is, like CMx1, hidden from the player so you don't misuse/abuse the unit the particular soldier is in.

I think it would be kinda cool to somehow track individual soldier feats of heroism and indicate that, somehow, in the user interface so the player could notice and remember who did something extreme and out of the ordinary. Perhaps we will someday add this to the game as it isn't in conflict with anything CM stands for. However, it will require a bit of work to get this sort of thing to function since, currently, there is no sense of tracking individual accomplishments nor evaluating them.

Steve

A recent example of "heroism" - I designed a Map/Scenario using CM:SF NATO that is based on the Squad Leader Scenario 45, Hide and Seek, 6-7-44 Normandy, France. It involves U.S.Army Infantry supported by several tanks, attacking a farm complex surrounded by hedgerows and agriculture, and defended by German Infantry.

Anyway, after about 1 hr. 30 min. of fighting I had my last remaining M1A1HC Abrams moving down a dirt road, bordered on both sides by hedgerows. The tank was conducting mop-up and looking for any remaining infantry in the hedgerows. The tank crews' commander, wrongly assumed that the Germans had expended all their anti-tank weapons/ammo.

As the tank was slowly moving down the road, all of the sudden, several explosions rang out and the tank halted in its tracks. German infantry consisting of what looked like a 5-man crack Hochgebirgsjägerzug team and at least one member of a Gebirgsjäger squad ambushed the tank. The Germans hurled several DM-51 hand grenades at the side of the tank which immobilized it, prior to the tank unloading with 12.7mm and seriously wounding two Germans. This was followed by several more hand grenades which destroyed the Abrams. The tank crew bailed out and was promptly gunned down, with two men KIA and two seriously wounded.

The Abrams had received some light to moderate damage to its tracks, targeting, IR optics, radio, and smoke launchers in previous fighting prior to this incident. However, prior to moving down the road, the tank was temporarily bogged in a muddy area, but got itself un-stuck so the tracks must not have been significantly damaged.

U.S. Army infantry in another perpendicular hedgerow were providing overwatch and had engaged some other German infantry in the hedgerows near the tank, but failed to see the German ambush team.

The German crack Hochgebirgsjägerzug team had a 0 leadership modifier, a +2 morale modifier, and was in a cautious state. The soldier in the regular Gebirgsjäger squad had a -1 leadership modifier, was in a rattled state, and had moved away from the hedgerow during the attack. The Germans had been engaged in previous fighting and had been on the battlefield for about 1 hr. 30 min.