Tank crew AI under fire

[Hpt. Lisse]

Hello all –

I’ve been having a wonderful time playing RT – truly the pinnacle of the CM engine – a big shout out to Aris and Kieme for their respective mods.

But it’s time, for the sake of realism, to re-visit the whole ‘shell impact on crew’ mechanism.

I’m playing “The Passage” battle from the German side right now, which is CM at it’s finest. I don’t think this is ‘spoiler’ material, but there’s a lot of Panzer MkIV vs. T-34 85 ‘at range’ (1500 meters in some cases) firefights going on. And it’s perfectly balanced – the more accurate gunnery of the German tanks vs. the less-accurate but harder-hitting 85mm Russian guns, the superior armor of the T-34’s vs. the more vulnerable Mark 4’s…. beautiful. I doff my hat to whomever designed this scenario.

My issue lies with AP tank penetrations, and their effect on the tankers. I’ve seen numerous front turret/gun mantle full or partial penetrations that were NOT knockouts, only to be followed by a world-class shot 1 or 2 seconds later. This is not sour grapes – I’ve seen it happen from my side (the Axis) as well, despite injury/casualty to crew members. I cannot imagine the ear-busting chaos that an AP shot smashing through my tanks’ turret would create. But I don’t think that CM currently models this well, given its realistic morale system.

It’s bad form to not offer a solution to a perceived problem… I’m thinking that there should be a delay of firing in line with the crew’s experience level, slightly randomized of course. I would also like to see a penalty for successive penetrations on a crew within a certain time limit (except fanatical ones) as well. (There’s one T-34 85 in my battle that has no less that 4 front turret penetrations, and is still blasting away…)

Comments, suggestions?

[A co]

Seems like a good suggestion. A delay in firing, or a penalty to accuracy, when a tank takes a serious hit.

[Amizaur]

I have identical observations and also feels it's not right. The number of casualties after front hull penetrations is low, the number of wounded crewmembers - ridiculously low.

Higher energy guns like 88L71 or 122mm AP produce average number of casualties and have high chance for vehicle knock-out, lower-energy shells like 75L48 very often don't cause casualties at all, and often need many (2,3,5,7) penetrations to destroy a tank.

One of things that needs improvement is also lack of any "shock" state for crew of penetraded vehicle. As Hpt. Lisse wrote - very often after penetration of enemy tank, it quickly returns fire just like nothing happened. Only if the penetration causes a "panic" state for the enemy crew, they do not react effectively, but panic state after penetration is not happening often. In other cases the fact that vehicle was just penetraded and some crewmembers killed/wounded, damage done, don't slow down crew reaction time and effectivenes.

A high chance for crew to be "shocked" for few seconds after penetration (not responding for a moment, untill they recover) would be a good addition, IMO.

I'm currently in the process of testing the casuality rate in various combinations of guns and target tanks, I have lots of interesting observations, but don't have enough data yet for various vehicles to write a raport. I'll post it here when it's finished.

But I can already say, that many combinations of shell vs tank produce much lower casuality rate than both common sense and historical data would suggest.

The tank in CMx2 seem to be divided into hull and turret sections, each one has it's crew and list of damage-able systems assigned to it. When shell penetrates then casualities and damage are randomly-chosed mainly from the hit part (hull or turret) and there is some chance for casualities and damage from the other part.

For example, if the hull is penetrated, then there is a chance for killing some crewmembers from hull battlestations and damaging some hull systems - but who gets killed and what get's damaged is random, it's not based on "where the shell hit".

If the shell hits the driver's hatch and goes trough his seat, there is still random chance that either driver or hull machinegunner would get killed. Same for damage - random - shell may pass under the turret or trough the front part, and damage the engine. If the shell hit from the side, right under the turret, trough the lower bodies of turret crew, it usually won't hurt the commander, loader and gunner and instead it will kill the driver or MG-radio-man (or both).

There is also a chance (but much smaller) than someone from the other part of the tank (turret) gets killed, and that some turret system get's damaged.

The tracks - although assigned to the hull - may get lightly damaged (and usually gets damaged) during almost any penetration. Same for optics.

Wounded crewmembers seem to be chosen from the whole crew, not based on where the shell hit (hull or turret). But there is VERY low average number of wounded, several times lower than number of casualties.

Sometimes the shell can cause ammo or fuel to detonate, penetration causes a great explosion and everyone is dead (but not much tanks systems are damaged) .

How much crewmembers become casualities is random, but it's somehow depends on the energy of the shell remaining after penetration and maybe shell calibre.

High energy guns like 122mm AP or 88L71 penetrating thin side armor of tanks cause on average high number of casualties, lower energy guns (like 75L48) cause low or very low number of casualties.

High energy guns that penetrate strong frontal armor also may result in low numer of casualties - seems than the remaining shell energy was low and the game calculated low chance for someone being killed/wounded.

Chances for vehicle becoming "destroyed" by a penetration and chances for catastrophic ammo explosion also seem depend on same factors as above (remaining energy of the shell that penetrated?). 122mm AP hitting PzIV side hull cause guaranteed vehicle kill and high chances for explosion. 75L48 in same situation cause low chances for both.

If the penetrations sets the tank on fire, the (surviving) crew obviously gets off instantly.

In the other case they can choose to stay in the (destroyed) tank for random amount of seconds (from 1 second to something like 20 sec), especially if they are not panicked. The tank is listed as destroyed and they do not fight, they do nothing, but stay inside.

I wonder what makes this tank so "destroyed" if all the tank systems are operational and it's crew is not panicked .

[Oddball_E8]

I recall in CMBN, having a sherman being bombarded by 20mm fire.

It was basically rocking back every time a shell hit (and those were quads, so plenty of hits).

And yet, it didn't seem to have a lesser accuracy when firing.

I think that not only penetrations, but even riccochets from weapons should affect the accuracy of a tank.

[Hpt. Lisse]

Yup, had one of those too (that was obvious)... now, this particular T-34 85 had spent two rounds 'bracketing' the intended Pz MkIV, but got slapped (ricochet) by a 75L48 round a moment before firing (rocking on its suspension) but nonetheless scoring a direct hit. Keep in mind the long range of the engagement (this one - 1100m) where even slight gun movements would be telling...

Strange thing is, I have played an MG battle where a short-range deflection clearly caused the return shot to miss (hit the ground not far from the barrel)... wonder if it's only modeled past ' a certain point'... Phil?

[Amizaur]

I've seen a shot hitting the ground near the barrel without a reason (nothing disturbed the firing tank). So in your case it _could be_ just a coincidence.

At 1000m after 1-3 initial missed shots (range estimation), for nearly any gun every next shot are all on target. It's ok, the technical accuracy of almost any gun permits that, and the range is set correctly now.

But even in such situation (3th or 4th shot to the same target) SOMETIMES the shell my miss - my test tank was putting round after round accurately on test target and suddenly next shot went into the ground 100m ahead of the muzzle. **** happens ;-) and this is simulated too ;-).

[sburke]

Yup, had one of those too (that was obvious)... now, this particular T-34 85 had spent two rounds 'bracketing' the intended Pz MkIV, but got slapped (ricochet) by a 75L48 round a moment before firing (rocking on its suspension) but nonetheless scoring a direct hit. Keep in mind the long range of the engagement (this one - 1100m) where even slight gun movements would be telling...

Strange thing is, I have played an MG battle where a short-range deflection clearly caused the return shot to miss (hit the ground not far from the barrel)... wonder if it's only modeled past ' a certain point'... Phil?

good thing you got hit by that 75 or the third shot would likely have missed too.

[Hpt. Lisse]

Another option for this would be to institute 'tanker realism' at the warrior difficulty level (similar to realistic artillery response times, etc.)

[Michael Emrys]

Is "Crew Shocked" still modeled in the game? I haven't noticed it, but then I haven't been checking for it either. This was a feature in CMx1. If a tank was struck by an AP round that was capable of penetrating but did not completely penetrate, the crew would be inactive for several seconds and when it returned to action would have its performance somewhat degraded.

Michael

[James Crowley]

I wonder what makes this tank so "destroyed" if all the tank systems are operational and it's crew is not panicked .

I've seen this happen a few times recently and have wondered the same.

The most recent case was in a H2H when a near-miss hit a tree and the targeted HT Stummel became 'destroyed'. There was no recorded system damage and none of the crew were injured. Even if panicked and the crew jumped out this should only be 'dismounted' vehicle. Seems rather strange.

[A Canadian Cat]

Stummel became 'destroyed'. There was no recorded system damage and none of the crew were injured.

This is just the timing of the updating of the game UI. Destroyed vehicles do not show detailed system damage. So if you are watching the turn with the vehicle selected and it gets hit, one moment it will be all systems normal and the next it will be destroyed. There is no "live vehicle with massive system damage" moment before the "destroyed" vehicle state.

So, bottom line that near miss totally trashed the vehicle in some way but the game does not tell you what was wrecked.

[Amizaur]

I'm sorry but I'm just doing massive amount of tank penetration/damage/destroying tests and I see the damage done by the killing shell.

One moment the tank is 100% operational and undamaged, second moment (after penetration) the tank is "destroyed" and has various amounts of damage to tracks, engine, optics or something else. So the game UI definitely shows the damage done by the killing shell. If it doesnt' show all, this should be corrected.

The cases where there is no damage and tank is destroyed are of course rare, but they happen too. There are also cases where there is almost everything damaged on tank (engine, tracks, optics, main gun, weapon controls ect) and it is still NOT destroyed, so the "destroyed" state seem not connected with systems damage state. I wonder what else in tank can be damaged, that is not displayed, but decides about destroying. That steel frame that mounts everything ?

[Skwabie]

I'm sorry but I'm just doing massive amount of tank penetration/damage/destroying tests and I see the damage done by the killing shell.

One moment the tank is 100% operational and undamaged, second moment (after penetration) the tank is "destroyed" and has various amounts of damage to tracks, engine, optics or something else. So the game UI definitely shows the damage done by the killing shell. If it doesnt' show all, this should be corrected.

The cases where there is no damage and tank is destroyed are of course rare, but they happen too. There are also cases where there is almost everything damaged on tank (engine, tracks, optics, main gun, weapon controls ect) and it is still NOT destroyed, so the "destroyed" state seem not connected with systems damage state. I wonder what else in tank can be damaged, that is not displayed, but decides about destroying. That steel frame that mounts everything ?

developed 2 theories on this

1. energy past armor reaches a certain level.

2. energy past armor accumulates to a certain level.

and there must be many random factor in play in these... otoh i don't think subsys dmg contribute to tanks KO based on same evidence in your post.

all just theories ofc, tho no need to think the system cm uses is very complex. games are just games...

[A Canadian Cat]

I'm sorry but I'm just doing massive amount of tank penetration/damage/destroying tests and I see the damage done by the killing shell.

OK interesting - I stand corrected. I will watch more often. So far I have never seen it. But then again I was not looking for it either.

[Oddball_E8]

Well there are damages that are not included in the UI of the game.

Things like dislodging the turret or a penetration that severs the driveshaft might not be reported on the UI and thus the vehicle is effectively destroyed, but still has no reported damage on the UI.

(ok, driveshaft damage might go under "engine" but what would a dislodged turret be? After all, the systems in the turret such as the main gun and coax are still operational)

[Hpt. Lisse]

While we're all working the tank crew AI stuff....

Same battle, one of my Pz Mk IV's has taken a couple hits - still operational, but down to 3 crew members. Nearby, a Crack Mark 4 crew of five is smoking cigs after losing their ride... job opportunity!

I bail the tank, and order the Crack crew to man it. Instead, they mount the hull as 'riders'...

[Michael Emrys]

While we're all working the tank crew AI stuff....

Same battle, one of my Pz Mk IV's has taken a couple hits - still operational, but down to 3 crew members. Nearby, a Crack Mark 4 crew of five is smoking cigs after losing their ride... job opportunity!

I bail the tank, and order the Crack crew to man it. Instead, they mount the hull as 'riders'...

War is hell.

Michael

[slysniper]

While we're all working the tank crew AI stuff....

Same battle, one of my Pz Mk IV's has taken a couple hits - still operational, but down to 3 crew members. Nearby, a Crack Mark 4 crew of five is smoking cigs after losing their ride... job opportunity!

I bail the tank, and order the Crack crew to man it. Instead, they mount the hull as 'riders'...

Bf has stated, that they do not want this in the game, claiming it gamey or would be used incorrectly. I think it is a excuse.

Almost every tank commander story I have read tells a account of losing their mount and they go and take command of another tank in their unit.

It would be a nice feature, But they should limit it to HQ units and then only to other tanks within their command. But the feature will likely not come.

As to the other post in this thread, I agree with most. What basically is lacking is crew down time after a major penetrating hit, there should generally be a period for the crew to have to recover from it, and only once in a great while a crew that somehow is not phased by the blow. Right now, responce time and fire is just not affected. The game does surely lack there.

[Andrew H.]

As to the other post in this thread, I agree with most. What basically is lacking is crew down time after a major penetrating hit, there should generally be a period for the crew to have to recover from it, and only once in a great while a crew that somehow is not phased by the blow. Right now, responce time and fire is just not affected. The game does surely lack there.

While I agree that this is *plausible*, I would still like more evidence that making such a change would be realistic. I.e., gunner has gun aimed at enemy tank; enemy shell hits tank and kills radio operator...and the gunner *doesn't* fire? Why not? What is he doing? It's not like being a rifleman where you can significantly enhance your (short term) survival by keeping your head down or by cowering in the bottom of the foxhole. And I haven't read any accounts of tank commanders having to order gunners or loaders to get back to aiming or loading the guns because they are cowering at the bottom of the turret.

(And there is a lot of "cowering" in tank behavior already, but mostly caused by tanks reversing away instead of engaging like I ordered them to...).

[Michael Emrys]

And I haven't read any accounts of tank commanders having to order gunners or loaders to get back to aiming or loading the guns because they are cowering at the bottom of the turret.

I don't think anybody has claimed that. I do know that what I had in mind was more along the lines that in a tank that has been hit hard enough to at least rock it and maybe also do some damage, everyone inside might take a moment to take a deep breath and check to see if anybody is wounded or any systems are damaged. And in the case that you mention of the gun already aimed at a threat, yes, I too might expect the gunner not to hesitate to pull the trigger. But that is near to being an outlier. A more common case might be that whoever has rung their bell has not even been spotted yet. So a few seconds to figure out what has just happened and then to start looking for the cause does not seem amiss to me.

Michael

[James Crowley]

Another element to this is the superfast replacement of a dead or badly injured driver.

I have seen it happen perhaps only a handful of times, when the crew hasn't bailed.

It happened recently in a H2H when my oppo declared that the side hull penetrating hit on his JS-2 led to the driver being KO'ed and replaced. This happened within the space of a turn; something which I would have thought impossible, particularly given the driver's location, in almost any tank.

[Amizaur]

I read accounts where gunner or some other crewmember was so concentraded and pumped with adrenaline that he didn't NOTICE a penetration that wounded other crewmebers - and continued to perform his duty, but I suppose it was not a "major" penetration (like someone called it).

I also read accounts where after penetration everybody or part of the crew were so shocked that they FORGET they are in tank and fighting, that there is enemy outside that continues shooting - only they wanted to escape, checked if they are ok or if their teammates are ok, and only after some time (several seconds) they were able to either evacuate from tank or get back to their stations and continue fighting, someone had to rally them and get them back to work.

Yes, gunner can "forget to shot". Suddenly all hell breaks lose and the survival is most important thing, the only thing. Remember the Panther aiming at the Pershing in Cologne, with shell in the barrel ? The gunner had his foot on the firing pedak, and only waited for the American tank to stop. But Americans fired on the move, penetrated the Panther and Germans never shot. IIR the gunner survived to tell this, or maybe some other crewmember. And we are not talking about situations where the gun is already aimed and shell loaded. OK, I believe in good percent of cases the gunner of penetrated tank WOULD still fire the shell, if he survived.

But there is lot of other combinations - gun is aimed but the shell is only being loaded, the penetration occurs that causes loader to drop and shell and check if he is ok and if tank is not burning.

The shell is loaded but the gun is not aimed yet, and the gunner momentarili loses his consciousness from the penetration effects, even for 2 seconds. It would take time for him to recover and remind what he was doing, and possibly first to check if he is ok and if tank is not burning, ask the commander what to do ect.

In CMx2 often one tank has the advantage of the first shot and penetrates other tank who didn't see the attacker - and only then the penetrated one spots (instantly) the shooter, quickly rotates turret and engages the shooter before he manages to reload and fire second time. And this way the ambushed tank often wins -with a bit of luck destroying the ambusher with first shell, before ambusher reloades for second shot.

Do we really think that after major penetration of the crew compartment (flash, bang, shrapnel and sparks flying everywhere, overpresure, flame, smoke) the crew is able to quickly locate the shooting tank, rotate the turret, load the shell and aim precisely in time as quick as fully operational crew ? Everyone (commander, gunner, loader) are ok and not shocked and operate like cyborgs without any moment for recovering, understanding what happened and if it's still safe inside ? How they communicate with possibly broken eardrums and eyes blinded for few seconds by the flash of the penetration ?

I can tell from reading many, many tanker diares (all I could find, and still looking for more) that in 80-90% of cases the first thought of the surviving crewmembers of penetrated tank was "...I live! should I escape?!?" "Is the tank burning yet and exploding in 5 seconds ??? I do not want to burn alive!" Second thought could be about his state (am I ok?) and state of his mates, if they are ok - maybe they need help. Sometimes - a reflexion about a decapitated friend laying on the floor and blood everywhere. Only then they would think about if the tank is ok and maybe they should return to their duties.

Sometimes there is no thinking at all - there is the hit, and then people just suddenly find themselves outside of the tank, looking at the smoldering tank and not remembering how did they get out from it ! And suddenly notice those burns or this severed leg or lack or an arm...

There is 10% cases where people acted differently (even didn't notice, or noticed but heroicly continued the fight) but the big majority would act about that like I described.

So I really think that a penetrated tank crew (some or most of them, randomly) should get a random 0-5s or 0-10s time of "inability" that would froze their actions (commander does not spot, gunner not aims, loader not loads ect). Only after that time they would recover and continue shooting.

The probability of the "shock", it's strength (length), how many crewmembers are affected, could be based on some factors like energy of the penetration, energy of the burster charge (if it went off), and especially on number of killed and wounded crewmebers - and of course highly randomised then.

Stronger the penetration, greater chance for stunning the crew, and also I think that if for example 2 crewmembers were killed (maybe in a gore way) and 1 wounded there should be greater chance for strong "shock" effect than if there were no serious casualties.

[Michael Emrys]

Amizaur has it exactly right IMO.

Michael

[James]

Amizaur has it exactly right IMO.

Michael

I totally agree with you.