Kill chance of gun vs. unarmoured vs armoured vehicle (serious posts only, please)

[Guest Martin Cracauer]

This is really a issue from the sdkfz 7/2 thread, however this issue (the key issue, IMHO) went under in useless discussion, so I decided to open a new thread.

Please, pretty please, keep this thread on target on this single problem. If you want to complain about the 37mm gun or small arms against vehicles, please do so on the other thread.

I want this thread to come up with concrete, implementable ideas to do a better balancing. After reading this thread, Charles should -if he follows our argumentation- be able to fix it within 5 minutes. Obviously, if we clutter up the thread with Übercrap, he will not read far enough, so please stay on target.

Surface problem (people familiar with the issue may skip this)

An unarmoured vehicle may be harder to kill than a similar armoured vehicle.

Like other games (i.e. Steel Panthers), Combat Mission has a problem with balancing the kill mechanism for guns capable of firing AT and HE rounds attacking armoured versus unarmoured targets. This is not a bug, it is a balancing issue and it is not easy to fix, so no reason to cry murder. Let us be constructive and all will be well.

When an armoured vehicle gets attacked, there is a piercing capability of the incoming round at one place on the vehicle surface, at a given angle. It is easy to compute whether there is a penetration and a further mechanism decides whether the vehicle is knocked out or not.

When an unarmoured vehicle is attacked, the armour piercing mechanism does not apply. Either the mechanism for small arms fire is used or the mechaism for explosion blast value. It seems that people generally don't have a problem with the small-arms versus unarmoured vehicle mechaisms, it seems to work fine, so lets concentrate on the incoming HE.

When one given tank (that is capable of firing AP and HE rounds) attacks an enemy unarmoured vehicle, things get messy. Would the enemy vehicle be armoured, the decision is clear and AP is fired (except in ammo shortage situations). If the enemy is unarmoured, HE is being fired.

But -and here comes the problem- the HE mechansm is very different from the AP mechanism. Near misses can do almost as much damage as hits and the damage taken from the blast must be computed. Also the HE rounds may not considered to be as precise as AT rounds.

As things stand now in the two mechanisms, it leads to situations where an unarmoured target is harder to kill than the same thing without armour.

Naive solutions for the problem do not work: if you would increasive chance of a given blast to knock out a vehicle, you would also increase the change of artillery knocking it out. However, we don't want all HE to be more effective.

The real problem

Aimed HE fire is not effective enough against unarmoured vehicles.

Existing mismodeling check #1: if CMBO models that HE rounds are not as precise at AT rounds, maybe the decrease is too big?

Existing mismodeling check #2: is the chance of a direct hit for a HE round against an unarmoured vehicle computed with the same complexity as the AT round against tank computation? If not, whatever else is in place now may be tuned to fix the game problem. May also been connected with (see below) vehicle size computation for non-tanks.

Existing mismodeling check #3: there is a mechanism to determine the blast value at a given distance from an explosion. Maybe the mechanism in place gives a small shell very near the target an inappropriate low effect on the target than a big shell exploding farther away? Even when that is not mismodeled, this effect (when done the other way round, small/near is overly effective) may be (ab)used to solve the balance problem.

Existing mismodeling check #3a (related to #3): is the surface exposed to HE computed with sufficient care? How complex is the geometry data of an unarmoured vehicle that is used to determine its share of a HE blast? Maybe the result of using the blast/distance mechanism together with the how-much-surface-exposed-to-explosion lead to small/near HE shells being less effective than they should?

Remember that the vehicle has two dimensions exposed, maybe an inappropriate simple mechanism is in place that uses one dimension instead? Example: only the length would be used, but not length x height. In that case, the impact of HE from a given distance would raise linear with the explosion coming nearer, where it should be quadratic. The result would be that small HE explosions very near would not have enough effect. This mechanismt may also be (ab)used to fix the game balance without too much code changes. Raise the effect of ever-nearer explosions with more than whatever is in place now. Adding 1/2 to whatever exponent is in use now will likely do good.

Quick and dirty overall fix: when HE is fired on an unarmoued vehicle, do the shot calculation twice. First, the existing one for HE, but second, compute what the knockout change would be if the attacker fired an AT round and the armour was minimum thickness above zero. When the AT shot would have knocked out the vehicle, there should be a very good chance that the HE shot did this as well. Choose an appropriate probablity of -say- 70%. So, if the AT shot would have knocked out the vehicle, choose a random number and make a 70% probability of it that the vehicle is knocked out from the shot (which really was HE). Do not use this mechanism when armoured targets are attacked by HE rounds. Not sure what to do with non-vehicle unarmoured targets like towed guns.

Please, people, stay on target here. I would prefer not to see "ok, the problem is clear, now do something, BTS" postings that usually come up with bug reports. This is not a bug. It is a complex modeling and balancing issue that you can hardly blame a programmer for. Exspecially not those who has been as responsive to us as the BTS team is.

The issue is relevant to CM2 as well, otherwise I would not open this can of worms while we have to save BTS time for giving us CM2.

Thank you for your attention and I hope we can show that we do our best to add to our game's value.

Martin

[This message has been edited by Martin Cracauer (edited 04-02-2001).]

[This message has been edited by Martin Cracauer (edited 04-02-2001).]

[Guest Martin Cracauer]

Test scenario:

http://www2.cons.org:8000/history/cmbo/unarmoured.cmb

A group of Ostwind and a group of sdkfz 7/2 fight shermans in seperate arenas.

Load as Axis, issue no commands, just watch. The 7/2 is more robust.

[Olle Petersson]

Originally posted by Martin Cracauer:

Existing mismodeling check #2: is the change of a direct hit for a HE round against an unarmoured vehicle computed

IMO this is the key issue.

Are direct hits calculated at all?

Any decent size HE (50mm and up) scoring a direct hit anywhere in the engine/driver region would most likely disable/immobilise the vehicle at least, by causing damage to the engine or driver.

Since most smaller calibre HE is counted for in terms of salvoes, it should probably apply to these as well.

As for effects of being in the blast area of near misses;

Perhaps some sort of "blast armour" could be applied to each vehicle, based on general ruggedness or something similar.

Then the blast value at current distance is compared to the blast armour and some percentage value of what can happen is calculated.

Some randomness ("dice roll") is applied and the outcome decided.

Cheers

Olle

[Mark IV]

I just ran through that scenario about 6 times (nice setup). However, on the last 3, I issued a group Move order to the Shermies. Just a steady Napoleonic-era advance to the enemy lines.

The difference was pronounced. Sherms won every time (at great cost, to be sure). I think this was interesting and I'm not sure what it means (Shermans should shoot better while stationary, and AA guns are designed to hit moving targets... the range didn't close all that much by the end). The Sdkfz 7/2 should be easier to hit, since it has a higher profile than a Sherman. But it can certainly be taken out.

If we accept that flat terrain tiles are abstractions and are not really billiard table flat, and hypothesize that AA crews are more adept and better equipped at making use of camo and fixed positions, we can rationalize some of this away, but then there is their ability to shoot on the move from a truck platform, where these advantages are negated.

I guess I would agree that they are too hard to hit, but certainly not to the point of "cheating". Something to keep in mind for CM2, but not enough to stop the presses and fix. I would still use combined arms to defeat them in a real scenario. When I have time it would be interesting to compare them in the same scenario with other soft vehicles.

[Guest Martin Cracauer]

The point is not the ratio of Shermans/flak-halftrack kills, but of Ostwind/flak-halftrack kills.

As for 37mm killing the Shermans, in my test scenario the German vehicles are by default positioned about at that range that the 37mm may penetrate the front of the Sherman. Move them back, they cannot penetrate anymore. Move them forward, they kill better.

Has nothing to do with that problem that the armoured attacker is easier to take out than the unarmoured.

[Ben Galanti]

I thought (from previous threads) that the problem was the 'mismodelling #2'. Specifically, that against unarmored targets, HE shots (including infantry anti-tank weapons) don't check whether they hit the target or not like AP rounds. It's purely a matter of where the round impacts the ground. If it impacts the ground close enough that the blast damage value can knock out the vehicle, then the vehicle dies.

So, instead of trying to hit the vertical profile of the vehicle, the gunners are trying to hit the horizontal patch of ground under the vehicle, a more difficult shot, I would assume... Though, this kind of gets into the fact that CM decides hits and misses against other AFVs at the instand of firing, instead of tracing flight paths (a necesarry comprimise to computer power) but that's a different can of worms.

Ben

[Mark IV]

Originally posted by Martin Cracauer:

The point is not the ratio of Shermans/flak-halftrack kills, but of Ostwind/flak-halftrack kills.

Understood. It does not surprise me that Ostwinds are easier to kill, as they are large mass but thinly armored, and as discussed, being engaged by the AI with more accurate AP.

FWIW, most of the HE misses on the flak trucks were unders. Perhaps they are trying to skip the rounds?

As for 37mm killing the Shermans, in my test scenario the German vehicles are by default positioned about at that range that the 37mm may penetrate the front of the Sherman. Move them back, they cannot penetrate anymore. Move them forward, they kill better.

But, as the Shermans move forward (slowly closing the range at 5 mph, in effect making the 37mm MORE effective), the flaks are LESS effective. They did NOT kill better. Perhaps they were intimidated.

At any rate, those were my observations. I was still in the cost/benefit mode, wondering whether this is really worth fixing, whereas you are well into the "how" of the fix, a pursuit into which I have no useful insight.

[Steve McClaire]

But, as the Shermans move forward (slowly closing the range at 5 mph, in effect making the 37mm MORE effective), the flaks are LESS effective. They did NOT kill better. Perhaps they were intimidated.

Probably because the Shermans are in motion, rather than stationary -- the Flak guns have a slightly harder time achieving a hit (or they hit with less rounds per volley), and don't benefit from 'ranging in' on a stationary target. I would expect this sort of slow movement to favor the Shermans with their gyro-stabilized guns (i.e. it would be less of a handicap for the Shermans than it would for the Flak trucks).

But we're getting off topic.

[pritzl]

I think this problem has nothing to do with HE vs AP rounds. Go ahead and run a test with Shermans vs SdKfz 7/2s with all of the HE rounds removed fromt the Shermans. They are foced to use AP and it is even less effective than the HE.

I'm starting to wonder if it might be a bug with the "size" of the SdKfz 7. Perhaps the data in the game code is missing a 0 at the end and it is therefore an incredibly small target or something of that nature.

It would then be a case of not so much being hard to kill as nearly impossible to hit. The end result would be what we're seeing now. Who knows? BTS?

[Wilhammer]

I just played this out, and I am dismayed at the results.

Essentially, unarmoured vehicles, guns and infantry are of the same target classification. This is borne out of the fact that the Ostwinds as targets have hit and kill chances, while the unarmoured do not.

The Ostwinds, after 3 trials, were all destroyed in under 30 seconds, while at the end of the turn the 7/2s had only 2 or 3 of their number killed. Since they had better survivability, they inflicted more damage on the Shermans than the Ostwinds did, which is ridiculous.

Why not just give unarmoured vehicles very light armor, like 1 or 2 mm?

In practical CM terms, unarmoured vehicles are NOT classified as vehicle targets by the CM engine.

Therefor, because SOFT targets like Infantry, Guns and unarmoured vehicles are subject to "spotted target tight area fire", the blast radius of the HE round is very important. You, in effect, never hit a soft target, you get firepower damage against it.

The question now is how do you deal with it "game" tactically?

A gun armed unarmoured vehicle can thus act like a very mobile ATG or Field Gun. So, instead of buying the immobile 50mm PAKs, buy these armed trucks.

[This message has been edited by Wilhammer (edited 04-02-2001).]

[Dschugaschwili]

Originally posted by Wilhammer:

Why not just give unarmoured vehicles very light armor, like 1 or 2 mm?

Sounds like a viable way of dealing with this problem to me. Infantry should still be able to knock out such a vehicle because even now infantry weapons (at least MGs) have an armor penetration value, although it's not shown to the player (could this be changed in CM2?). And guns could finally hit those unarmored vehicles.

Dschugaschwili

[Olle Petersson]

Originally posted by Wilhammer:

Why not just give unarmoured vehicles very light armor, like 1 or 2 mm?

I think this is the most complex option available, and I'll explain why below.

A better option is to make all vehicles valid targets for direct fire.

Why adding light armour is complex:

* With armour, all guns that can fire AP will fire AP, by preference.

* AP against regular armour is effective, but large calibre AP against soft or thinly armoured targets is ineffective.

The effect of AP against armoured targets is that the interior is exposed to ricochets of the AP plus fragments of torn armour, and in addition to that there's a loud bang and sudden rise in air pressure (the latter not applicable to open topped vehicles).

The effect of AP against soft targets is that a single, fairly small, projectile pass through in a straight line. If this projectile doesn't hit anything vital (as is the most likely outcome) it doesn't do any noticeable damage.

As is large calibre AP is too effective against light armoured targets (in CMBO), and that's something I can live with. If the same is supposed to be applied to soft vehicles as well, then there's a need to remodel the effect of AP hits, almost from bottom up.

An example of current mismatch:

* Target is a SPW 250/1, without passengers or cargo, and fired at straight from the side. Then a hit will most likely be at the passenger/cargo compartement. (90% chance of hit >1m above ground, most of that area is the rear compartment.

- If the shot a 17pdr AP, then that shot will pass straight through and most likely not hit the driver, gunner or engine.

- If the shot is a single .50 cal FMJ, it will penetrate the armour and the bounce around inside, with a higher chance of hitting something vital.

* In CM(BO) it's a much higher chance that a single 17pdr AP hit will destroy the vehicle than that a .50 hit will cause any damage at all...

A fairly easy (but still complex) way to counter this would be to have tables like;

IF armour plate thickness <X mm AND (hitting round is AP shot with penetration power >Y mm {>>X mm}, THEN treat target as if soft.

... apply similar to AP shell and so on...

Cheers

Olle

[massimorocca]

Question, what is the kill ratio against the 7/2 with tank's mg? Anyone tested if they works better than main gun?

Massimo

[Wilhammer]

As I posted;

Why not just give unarmoured vehicles very light armor, like 1 or 2 mm?

This would be a quick fix, perhaps.

This depends on the logice of ammo selection. If it was something like IF ShellPenValue is 5X TargetArmourValue THEN select GreatestAmmoLoad. In most cases, an HE round is fired. The question is, what is the penetration value of HE?

As Olle posts,

"A better option is to make all vehicles valid targets for direct fire."

That is a better fix, but if HE does not have an inherent penetration value, AP would be used.

So, does HE have a penetration value?

Yes, it must have one, for how else does a mortar shell penetrate the top of a tank in CM?

Olle's method and my method are not all that different. One can rightly assume that unarmoured vehicles have very thin armor. They at least have cheap sheet metal.

The same can be said for ATGs, etc, what with armoured gun shields (or the lack of), but with the accuracy of direct AP fire, guns would have even shorter lives than they do now.

Has anyone seen a gun get destroyed? Or a gun hit on a gun?

Not me.

The way a gun is knocked out is when the crew is routed away from it; it becomes de-crewed.

Is the same thing happening to 7/2s?

If so, an elite crewed 7/2 must be an awesome thing.

[This message has been edited by Wilhammer (edited 04-03-2001).]

[Wilhammer]

More testing with unarmoured.cmb.

Pardon me if this is "old hat" to some of you.

Based on my interpretation of what I observed in CM;

1. Two target classifications exist

a. Armoured vehicles. They are directly targeted by guns, RLs, etc. They take direct hits and they can suffer damage from HE blasts.

b. Soft targets. These are NOT directly targeted. Instead, I must define two type of area fire. Area Fire one is "Terrain Area Fire", the kind we do when we wish to blow up a building. I will call this TAF. The second type is "Target Specific Area Fire", which I will call TSAF.

The possible damage by either is the same, it is the tightness of the shell fall (or bullet fall) that is different.

Soft targets are damaged by a firepower value that is a function of the blast value of the round and its proximity to a soft target.

Before I get too far, a soft target, ST, is any unarmoured target; infantry, guns and UA vehicles.

Hope I have not lost you.

So, when I test this theory, I find that UAVs are NOT direct targets but they are TSAFs. The shell always misses the vehicle. Always. The damage done is from the blast. The success of the blast is from its power and proximity. The blast ALWAYS leaves a crater. The shell hits near the target, then the damage is calculated. Sometimes the blast blows up the vehicle or KOs it (has anyone seen a mobility kill vs a soft vehicle?). Most of the time, the crew fails a morale check and abandons its mount.

If a AFV is fired upon, it is subject to direct fire damage. If it is hit by close proximity area fire, it can also be damaged. The crew is not going to be paniced out of the vehicle. It could abandon the vehicle, almost always as a result of 2 crew members getting KOed.

On the subject of MGs, it still seems that the for UAs, the result is not against the vehicle, but against the crew. If it routes, the vehicle is abandoned. Sometimes the vehicle does take damage, but it does seem that this is akin to collateral blast damage and not any form of penetration damage.

Now, I took unarmoured.cmb and gave the Germans all elite crews, and the 7/2 is a decidedly awesome weapon. The Ostwinds, having smoke dischargers and penetrable armour alwas withdrew behind smoke. The 7/2s did not withdraw, the Tac AI actually moved them to seek better target aspect. However, since they lack smoke dischargers, I cannot be sure of why they did this. Was it because they had nowhere to hide or was it because the TacAI knew they stood a very good chance of winning the battle?

The result on the Ostwinds was that their survivability went up, mostly because of smoke. Because of this, the Sherman Regulars vs Ostwind Elites resulted in fairly even kill ratios. The 7/2 elites vs Sherman regulars resulted in a high proportion of 7/2 survival and the destruction of the Shermans.

Clarity Note:

Infantry or team equipment can be damaged by blast or gun fire. Its vulnerability to such varies. Hence, a Jeep is like tissue paper, while the 7/2 is like a metal drum. This is the meaning of Knocked Out for crew served weapons, the weapon itself was rendered useless. This also further suggest that a soft vehicle is modeled as an ATG or Gun; it is modeled as a crew served weapon.

[This message has been edited by Wilhammer (edited 04-03-2001).]

[This message has been edited by Wilhammer (edited 04-03-2001).]

[Wilhammer]

This gets wierder every time I do unarmoured.cmb.

I just saw a 7/2 (I modified the scenario for all elites and Axis starts hidden) move and STAY hidden (I was Axis player).

Also, I noticed something I have been blind to; you cannot button an unarmoured vehicle!

BTW, now that I changed the 7/2s back to rear facing the shermans, the TacAI does try to withdraw to the board edge.

They again managed to kill all the Shermans and lose just 2 of their number.

The Ostwinds were crushed.

Another test; due to this peculiar behaviour, what German AFV is a 7/2 roughly equally effective as?

I traded in the Ostwinds for Panthers, then for PzKw IVHs.

After several trail runs, a 7/2 is about as effective vs M4A3s as an equal number of IV H's when the engagement range is roughly 500 meters or less.

You just got to remember to reverse them into battle. A Regular IVH cost 118 points. A Regular 7/2 will cost you 48. Elite cost is 205 vs 84.

[This message has been edited by Wilhammer (edited 04-03-2001).]

[This message has been edited by Wilhammer (edited 04-03-2001).]

[Guest Martin Cracauer]

Originally posted by Wilhammer:

Why not just give unarmoured vehicles very light armor, like 1 or 2 mm?

In addition to what Olle said, I think this is a bad idea because of ammunition usage from the attacker. The round attacking a soft vehicle should be HE. And HE has side effects, for example nearby other units will suffer from a HE shot at the soft vehicle, but not from a similar AT shot.

I am pretty sure that there is a problem in CMBO mechanics that lead to a miscalculation of how much surface is exposed to a blast from a given explosion. Either oversimplification of non-armoured vehicle geometry. Or neglect to calculate the correct exponential rise of blast effect for nearby explosions.

Your suggestion of minimal armour would push unarmoured vehicles into the same geometry mechanism as tanks and would therefore solve problem #1 (if it exists), but full armoured mechanics do not apply.

[Wilhammer]

Martin, that is why I further suggest something like this from an earlier post;

This depends on the logic of ammo selection. If it was something like IF ShellPenValue is 5X TargetArmourValue THEN select GreatestAmmoLoad. In most cases, an HE round is fired. The question is, what is the penetration value of HE?

If we want to model the real world, then giving very thin armor to a 7/2 is realistic. As I said before, it is covered in sheet metal. Of course, though, the weapon system is completely exposed.

[Lorak]

Wilhammer,

Thank you for bringing this thread to my attention. I haven't tried the battle yet, and I have never really played with the flak trucks either. I really need to check this out.

Thanks again for the heads up

Lorak

------------------

"Do not wait to strike till the iron is hot; but make it hot by striking."--William Butler Yeats

Cesspool

Combatmissionclub

and for Kitty's sake

=^..^=

[Wilhammer]

Another though comes into mind.

Lets say you have some squads you want to move rapidly into battle; what is more survivable, riding in a HT or a truck?

[Olle Petersson]

Originally posted by Wilhammer:

Another though comes into mind.

Lets say you have some squads you want to move rapidly into battle; what is more survivable, riding in a HT or a truck?

Trucks are pretty survivable, as long as they're able to move at a decent speed. I have a scenario with several trucks on dry open terrain, and they're a real PITA to take out. At the same time it's very simple to knock out a truck moving in snow...

On the other hand I've got a feeling that the infantry will get less beaten if transported in an armoured halftrack than if transported in a soft vehicle.

Cheers

Olle

[Guest Martin Cracauer]

Originally posted by Wilhammer:

Lets say you have some squads you want to move rapidly into battle; what is more survivable, riding in a HT or a truck?

The unarmoured transport vehicle is still difficult to knock out, the infantry that is not armour protected will suffer from small arms and HE as if it was in the open. And infantry riding in an APC will suffer only from penetration shots.

So far so good, but when a gun capable of firing AT and HE is attacking the transport, infantry riding without armour protection is still less vulnerable than those in APCs, because of the low chance of a hit for HE vs. unarmoured.

This is not a different problem, though. When the mechanism to compute the effect of HE shells on unarmoured targets is remodeled to give aimed HE shots (small shell explosion, but very near) their appropriate damage chance, both problems are fixed.

Of course, CMBO does not model the cover given by the tank turret or other vehicle parts for the unarmoured transport, but that is a different issue.

[aka_tom_w]

This issue is Still a MAJOR problem IMHO.

-tom w

[Kanonier Reichmann]

<BLOCKQUOTE>quote:</font><HR>Originally posted by aka_tom_w:

This issue is Still a MAJOR problem IMHO.

-tom w<HR></BLOCKQUOTE>

Couldn't agree more, hence this attempt to bring the matter to BTS's attention again.

Regards

Jim R.

[aka_tom_w]

<BLOCKQUOTE>quote:</font><HR>Originally posted by Kanonier Reichmann:

Couldn't agree more, hence this attempt to bring the matter to BTS's attention again.

Regards

Jim R.<HR></BLOCKQUOTE>

from Steve:

"Big Time Software

Administrator

Member # 42

posted 04-05-2001 04:52 PM

A quick response...

The main, overwhelming problem is not the vehicle itself but the whole issue of "Rarity". The fact is that they were quite deadly,

but they were also hardly ever seen in a frontline battle type situation. At least not a pitched battle type environment like CM

simulates. So putting even one of these flak weapons into a battle will produce results that will appear to be quite unrealistic when

thinking about combat in an average sort of way. This is as true for flak vehicles as it is for any other rare, but potent, type of

weapon.

Putting large numbers of these weapons into place is just silly. It tells one nothing about how good or bad the modeling is. Put in

20 Jumbos and tell me how good PzIVs are, or 20 Jagdtigers and see how good Cromwells are, etc., etc. One can not make a

totally unrealistic scenario and expect to learn anything about realism through its conclusions. Faulty logic from the get go.

As for patches to CM1... there are none planned. Certainly none for the short term. We are full bore CM2 and that is not going to

change. Constant tweaking and fixing of CM1 has already delayed us far more than we would have liked. No game can ever be

perfect, so therefore there will always be "unfinished business" remaining. At some point a game has to be considered "done",

and 1.12 is it.

Might we issue a 1.13 sometime in the future? I don't know. It hasn't been rulled out, but we also have no plans for one. Even if

we do release a 1.13 at some point, we most likely will not do any major coding changes. Addressing the basic issue of soft

skinned vehicles as targets for HE will definitely require such work, so it will almost certainly not be found in any potential future

patch. Again, no game is ever perfect so by definition that means there will be imperfections to be found. There is no way around

this bit of reality.

Steve "

testing

:eek:

-tom w