Jeff Duquette

"Cheating" in the sense that the AI is given some advantage in final hit probability over the same hit probability for the Human player...

I think this question is well worth a bump, particularly considering the complaints about the games antitank guns and some folk’s perception of ATGs capability against infantry (see the sniper ATG thread). I don’t know how the game details this sort of thing, but would enjoy reading about how this sort of thing is modeled by the game. Final Hit probability is of course a function of a number potential error sources. Error budget; or dispersion; or round-to-round variations (whatever phrase a person might like using when they are referring to shot spread) is typically broken down into systemic errors, constant errors and human errors. Systemic dispersion or inherent dispersion or systemic error: This is a random error source that is always gonna be there. Moreover it doesn’t matter how much training you give a guy or a crew, you will always have inherent\systemic shot dispersion. This is caused by vibration or minor contrasts in how the gun or gun carriage distributes shock or stresses from firing; and/or minor contrasts in projectile weights or projectile shape; minor contrasts in propellant burn rates or contrasts in the arrangement of propellant grains between different cartridges, etc. Most of the hit probabilities one sees in books by Jentz and the like, or in firing tables or schusstafeln is based solely upon systemic dispersion. Constant error sources: This includes things like a steady cross wind; or propellant temperature; or air temperature; barrel wear, minor variations in barrel rifling and barrel alignment, etc. These things will result in variation between the location of the predicted\desired MPI and actual MPI -- i.e. difference between where you are aiming relative to the centroid of a shot grouping. But unlike systemic error sources, constant error sources can be compensated for. For example ballistic tables for artillery will invariably include correction tables for differences in air temperature, or propellant temperature, barrel wear, etc. Zeroing a weapon can compensate for minor variations in barrel "straightness". Bore sighting is a form of compensation for a constant error source...etc.. Human error. This can include aiming and sighting related errors; incorrect range estimates, lead errors, blah, blah blah... Any one of the above contributors to the error budget can be further subdivided or detailed. For example a human's aiming error may be the result of poor visibility, or it may be a function of a poorly trained soldier that doesn’t understand how to use his front and rear sights; or it may be nervous energy\pucker factor that affects a soldier’s aim. Etc etc etc Final hit probability is than a direct function of the error budget combined with target size and target movement.

Perhaps this was the tactic you should have used in the first place rather than the extremely creative frontal assault. Out of curiosity, how many casualties do you personally think is a reasonable number when conducting a screaming stand-up and sprint frontal assault against an AT gun blasting away with HE? Zero, one, two, three, four, six, eight? Or are any one of these casualty rates a statistical possibility? However, to be fair to the tactically impaired, I have only been testing this from the perspective of me playing the ATG\ATGs rather than the AI playing the ATG\ATGs. I suppose the AI may be "cheating" with respect to ATG fire. The question would than be is the AI cheating more so with ATGs accuracy than with other weapons. All the more reason to look under the hood and see what the games engine is doing with respect to shot dispersion and hit probability. Regards JD

Thanks for bringing my attention to these Finnish Army ballistic tests I am still hoping someone has more detailed information for the AP trials -- possibly impact velocity or range at which the trials were conducted. But this more detailed information may be rather difficult to come by. Best Regards Jeff

I was perusing my dusty memory cells and I seem to remember Jentz going on about the roof of the Panther being vulnerable to heavy caliber Soviet HE shells. Or maybe it was Guderian made some comment about this issue after Kursk. Anyone remember coming across this and where it's located? I know the British tested 75mm and 25-pdr HE against captured Panthers and neither shell had much effect on the heavier vertical armor – to include tests against the thin vertical armor on the rear of the turret. However, tests with the same caliber HE-shells could puncture the turret roof or hull roof armor and introduce blast and splintering into the crew compartment. By "could" I mean sometimes it did and sometimes it didn't. Conversely the British also conducted tests in which they drove a number of Churchill tanks through 25-pdr air-burst barrages, but the Churchill’s suffered very little notable damage. Of particular note was there was apparently no damage to any of the vehicles optics inspite of all the splinters flying about. I think one of the tanks received some track damage and was immobilized.

I asked around a bit about the Finnish Army ballistic trials against the T54/T55 turret. A fellow from another forum was kind enough to respond to my inquiry with some additional images from these tests. He also provided a photo of the museum's placard that provides brief descriptions of the various projectile impacts. The upper turret splash is 100mm HE -- fuze type and fuze setting is not indicated on the placard. Presumably OF-412 HE-Frag. The lower splash is 57mm HE -- fuze setting is also not indicated on the museum placard. Both hits were on the rear most side of the turret. The 100mm HE hit is on about the thinnest portion of turrets "vertical" armor – I think the “horizontal” armor on the turret roof is a bit thinner. Armor is about 56mm (cast) near the 100mm HE splash location, and about 62mm at the base of the turret rear in the vicinity of the 57mm HE splash location. There is a bulge evident on the interior side of the turret from the 100mm HE hit. There is also a weld crack on the roof armor as well as what appears to be the start of an arc shaped shear crack -- probably the initial phases of rear surface plug formation. There is no indication of any backsurface scabbing or spalling. So for HE with t/d ~ 2 we are starting to see some rear face bulging and the start of plug formation, but nothing yet in the way of BAE. There is no hint of any bulging or cracking on the interior side of the 57mm HE strike. Obviously at t/d of about unity there is nothing worth mentioning. This is pretty simplistic as I am guessing impact velocity will tweak these numbers up or down. Moreover OF-412 is a pretty high velocity HE round, but perforation of about 0.2 to 0.3 calibers is about the ballpark for HE vs. steel armor. This isn't completely from the hind quarters. Having looked over a bit of ballisitic test material for CPC and semi-armor piercing projectiles it's apparant that this form of shell will do about 0.45 to 0.65-calibers of armor perforation. But of course CPC and semi-armor piercing are a bit more like AP-shell design in that these things employ a heavily thickend solid nose section and base fuze rather than no nose section and a relatively fragile PD fuze. Best Regards Jeff [ June 05, 2007, 10:26 PM: Message edited by: Jeff Duquette ]

The classic damaged Tiger armor photo. Good image to post what with the HE splash adjacent to the AP perforations. Nothing to do with the topic at hand, but I read somewheres' that this photo was actually taken at Soviet testing\proving ground rather than Kursk. I know we have all seen this image in every Kursk book ever published. I can't for the life of me recall where I read this neuvo interpretation of the photo or why the dude thought it was from proving ground trials...

Ok...thanks. I am familiar with the T-72M1 tests, and have a number of photos of the turret from this vehicle on my hard-drive somewhere -- interior and exterior shots. As I understand it, the Finnish Army had tested 125mm BM-15 APFSDS and BK-14M HEAT against the T72 Turret. Interesting trials – particularly the armor section on the front of the turret. Regards Jeff [ June 04, 2007, 04:57 PM: Message edited by: Jeff Duquette ]

I think Jippo indicated 100mm HE in his post. I assume these were trials using the D-10T or D-10TG or some other iteration of D-10? I think the early T34's were all RHA or HHS, but as I recall many marks\versions employed cast armor turrets. Perhaps one of the local T34 experts can verify this bit? [ June 04, 2007, 08:20 AM: Message edited by: Jeff Duquette ]

Jippo: Excellent image. Thanks for posting it. Can you say what the original photo source was and if there are additional photos of this turret? Any larger-res images available of this turret? What projectiles were responsible for the other complete-penetrations? 100mm BR-412 -- BR-412B, or something else? Regards Jeff

So it would seem they were looking for intermediate-uber. Super-uber makes the game crash, and low-level-uber isn’t uber enough. But seriously, I think adding the StuH 42 would be a good thing, and I wouldn’t think it would be particularly difficult to add given what is already available in the game.

yeah but sturmtiger is super-uber but it didn't make the cut.

HE has never been particularly effective at destroying tanks -- with a few notable exceptions -- like British 25-pdr crews talking about direct hits on panzer-IIs and III's being able to knock a turret a-skew. But HE's lack of effectiveness against armor plate is kinda' why armor piercing shot and armor piercing shell were invented It's kinda why 25-pdrs also started including armour piercing shot in their ammunition load outs. It's kinda why most artillery units by wars end typically included HEAT or Armour piercing projectiles in their ammo load outs. Just in case the battery were about to be over-run by tanks. Nose fuzes used on HE are relatively fragile. They are crushed when PD-delay is used on armor and the shell happens to hit a tank. The shell casing is than just a thin fragile cylinder of mild steel. Hardly an optimal shape or optimal casing hardness for perforating armor plate. If the fuze is destroyed by crushing on armor plate – well you might get some detonation of the bursting charge in the form of low order deflagration -- and you might get some minor penetration from the kinetic energy of the thin shell casing -- assuming it doesnt shatter on impact. PD-super quick will give you high order blast and fragmentation but not much real penetration -- the shell explodes too quickly for it to penetrate. As to PD-quick or superquick scabbing or spalling armor plate ala HESH or HEP, point detonating fuzes employed on HE don’t create the appropriate conditions for such an effect. Moreover HESH and HEP employ a base detonating graze fuze. The nose of HESH crushes up against the armor, than the base fuze initiates detonation of the bursting charge. The blast wave\blast pressure transmits from the rear of the bursting charge toward the crushed nose of the HESH projectile. The high pressure detonation wave propagating toward the armor and into the armor plate creates one of the key conditions for back-surface scabbing or spalling. Conversely, PD-fuzed HE shells result in the complete opposite direction of blast wave propogation – i.e. the blast wave is initiated at the nose fuze and propagates backward toward the base of the shell. Common HE shell creates little or no real potential for back surface spalling from blast pressure\blast wave propogation into an armor plate. I suppose with poor quality – thin armor and PD-dely – there might be sufficient kinetic energy from the shell impact itself to knock off a back surface scab. But again an HE shell consists of a rather thin mild steel cylinder with a relatively fragile nose fuze – fragile with respect to impacting armor steel. Artillery delivered HE vs. tanks is traditionally not considered a target destruction mission. It’s more of a target suppression mission. Fragmentation, and to a lesser extent blast, can destroy vehicle optics – it can get crews to button up – it can also generate F-Kills or M-Kills from direct hits or near misses by larger caliber HE shells, but rarely does it generate K-Kills against tanks. It’s sort of why modern artillery turned to Copperhead, EFP and similiar warheads when engaging tank formations in which the target effect required is destruction rather than suppression. WWII statistics for tanks being K-killed by HE shell fire are relatively low and typically run around 5% to 10% of all tank kills studied for a typical BDA operational report -- or a BDA report for a specific operational stomping ground. The vast majority of tank K-Kills were generated by either other tanks or antitank guns firing AP-shot or AP-shell. Second highest source of kills is invariably mines. Air to ground, infantry antitank weapons (bazooka, panzerfaust, magnetic hand placed mines etc etc) and artillery make up the remaining tank kills with each weighing in at around 5 to 10% of total tank losses. A notable exception regards very late war statistics in which we occasionally see German infantry hollow charge AT weapons accounting for up to 20 or 40%(+) of all tank kills for a specific operational study. For those looking for references, these sorts of statistics can be found by anyone willing to put a little bit of elbow grease into studying the subject – and by elbow grease I don’t mean a quick goggle followed by perusal of the latest Wikipedia disaster. Best Regards Jeff [ June 02, 2007, 11:05 AM: Message edited by: Jeff Duquette ]

Hi Oudy: I have never been able to get the CMMods.com web master to register me. If you have changed anything in your scenario files from the second version of your scenario you emailed me, would you mind emailing me the newer version of the scenario? Thanks a bunch. Jeff

Sorry -- just an aside -- but while futzing around with another test scenario I determined that tanks can run over and kill infantry. It would seem the designers actually intended for this sort of thing to happen as the soldier graphics actually look like a dude is being run over. Not that I am morbidly dwelling on this sort of thing, but it is one of those additional touches to a game that is rather interesting. I also noted that I could change positions of crewmen within a crew by a simple drag and drop method during game play. So if my loader has better gunnery skills I can simply drag and drop him into the gunnery position and he will take over the shooting.

Very interesting SlapHappy. Thanks for posting your findings. What data file is this stored in? Have you been able to determine where hit probability is calculated and how it is calculated? I'm curious to see if or how systemic dispersion is handled by the game engine (gun/projectile inherent shot dispersion -- or streuung -- or whatever other term you might like using).

In an attempt to move from anecdotal toward pseudo-scientific, I setup up a simple scenario to test antitank guns snipping capability. Two German 37mm PAKs vs. two Squads of Soviet infantry. And since it was indicated above that ATGs were pretty much getting direct hits on infantry dudes, I also edited the ammo load out for the PAK’s. I eliminated all the HE ammo and gave the guns only armor piercing ammunition. I setup the infantry about 150meters from the PAKs. The gun crew ratings were regular – one gun had a gunner with a +5 gunnery rating. The other crew – well I didn’t have a chance to see the gunner’s skill rating as the crew from the second gun was mowed down before the gun could do any shooting. The remaining crew got whittled down to two guys relatively quickly -- but did manage to do a fair bit of shooting before getting completely eliminated as well. As to the pin-point accuracy of the PAK vs. the Infantry dudes, they fired 58 APCBC rounds and killed three Russian infantrymen. Typical ranges were between about 100 to 200meters.

Not that this needs to be done in your scenario, but you can place units outside of deployment zones within the editor. I did this by accident in a couple of scenarios I modified. The unit is basically frozen in place during the deployment sequence. I think you may be able to rotate its facing, but I'd have to check that again. As to the scripting thing -- I have not played with it at all in the editor. I'm sure there are alot of folks here that would enjoy reading your experiances with scripting the scenarios and how to go about scripting things. Or is this already covered in one of those editor videos?

Wounding and there effects within the game are a different issue. Not having seen the coding, I have no idea what a wound means in real game terms. I know my little computer dudes seem to get about reasonably well after a wound -- and they can still shoot and such. I assume there is some effect for being wounded, but what it is I don't know. Probably worth opening a seperate thread on wound effects.

I did -- see above. Sorry -- I disagree with your assessment. I don't think there is an issue. But as I said above, if you do think there is an issue, than tweak the gunner skills in the editor.

That seems to be what my issue was. I was adjusting their starting positions and the guns would start the scenario unlimbered. I played it again -- the new version -- and didnt adjust the start position of the guns or their transports and the things stayed limbered. The game however drug on till I killed all the attacking Germans. I had to kill 46-Germans, than the game finally ended. I suffered pretty heavy -- 44-KIA. I wonder if that is resulting in the extended play -- i.e. if the Soviet loose alot of men?

Yeah I saw the same show on the history channel. Very entertaining. Weren’t the dudes Canadians recently returned from Afghanistan? I think they were firing a 105mm Howitzer. I have to wonder if these dudes had ever done much direct fire shooting before this competition. But than you don't need a direct hit between the eyes with 105mm HE to get a **** load of target effect. The 105mm Howitzer isn’t optimized for direct fire. I dunno for certain about this particular make of 105mm, but some of these howitzers require one crewmen adjusting for elevation and another crew member adjusting for deflection and lead. Sorta like two guys trying to aim the same rifle at the same time. Although elevation is gonna be whatever you need for the range estimate, so maybe not so complicated. In practice these things could fire HEP\HESH and HE over open sights – just in case the battery (or gun) were about to be overrun by tanks or infantry. You see accounts of this sort of thing – particularly during the initial phases of the Korean War. Over-running artillery batteries wasn't always a cake walk for attacking tanks or infantry. The Parrot crew were ringers. They manufactured their own ammunition, and I think they also made their own precision gun barrels. It wasn’t their first competition either -- they were old hands at that game. I’d guess these guys with their precision made Parrot and precision made ammunition would have out shot 99.9% of the gun crews in the Army of the Potomac. On the other hand, light and medium ATGs have one gunner dealing with both elevation and deflection. The system is obviously optimized for direct fire against point targets. Training was routinely conducted against moving targets – for obvious reasons (albeit it vehicle sized moving targets). I have collected several WWII antitank gunsights over the years. I’ve got sights from a 2-pdr, Canadian made 6-pdr telescope, a US-Army 57mm and a telescope for a German 37mm PAK. They are pretty much like a telescopic sight for a rifle, cept’ some of these things also included lead lines and stadia lines for different ranges and different ammunition types. As to the rest, I’m gonna have to disagree about the pinpoint accuracy stuff in the game. I’m really not routinely seeing 10 shots, 10 kills against infantry. I just finished replaying a scenario with a couple of Soviet 45mm Paks. Yes the guns can kill infantry with HE -- I'd guess that's why they gave these things HE. There real good at killin' dudes at close range when the dudes are settin’ still. But guys runnin' -- yeah I suppose there was the occasional one shot one kill thing going on -- but ten for ten... Maybe a gunner with maxed-out gunnery skills. On the other hand I did experience the joy of a German sharpshooter getting four for four kills on some of my low crawling guys. I finally killed the sniper with HE from a 45mm PAK. He took off running after my gun crew spotted him and started shooting HE in his direction. It took about six HE shots to bring the sniper down. This was pretty short range too – less than 50m initially and prolly 100m by the time the Pak did him in. It wasn’t a direct hit between the eyes. More like a near miss. But it was a gratifying kill all the same. Shell splinters must have got him as he ran for the tree line. The 45mm HE round is I suppose sort of akin to a fragmentation grenade going off. In other words it doesn’t need to be a direct hit to generate target effect. It only needs to be close. And a fellow moving upright is far more susceptible to being hit by shell splinters than a dude lying prone. But I think there might be an easy go around on this – that is for players that think the ATGs are too accurate with their abilities to fling HE. Crew skills can be tweaked in the editor. There is an accuracy skill rating and a gunlayer skill rating. [ May 30, 2007, 10:15 PM: Message edited by: Jeff Duquette ]

Still not tracking. If I'm about to be overrun, I reckon I'm gonna throw as much HE down range as I have at hand. Nothing unrealistic about it. But hey -- that's just me. Lighter AT guns usually are set up with a yoke where by the gunner simply traverses via a nudge from his shoulder toward the direction of the target. Or if the weapon isnt traversed by a simple yoke than theres a simple traversing wheel utilized by the gunner. If the target is outside the frontal arc of the gun than what I see is the "game" crew exposing itself by picking up the trails and manually traversing the entire gun. But if the attack is within the frontal arc the gunner traversing to engage individual targets isn't the result of some monumental effort required on the part of the gun crew. Personally I'm not seeing a problem with the game -- at least not for this particular issue. If infantry starts closing in on an Antitank gun what I see is the gun trying to defend itself by firing HE. The crew is typically taking casulties from small arms fire. If the gun gets flanked it is very easy to kill the crew -- i.e. once small arms fire is directed into the crew position from outside the covered arc of the gunshield the crew begin dropping like flies. If your doing a frontal assault on an antitank gun and crying cause' you're taking fire and casulties from HE -- well you need to look to your own errors in tactical judgement rather than blaiming some "bug" or foible associated with the game engine. Try some overwatch via a sniper or an LMG. Than try flanking the gun with a maneuver element. [ May 30, 2007, 10:12 PM: Message edited by: Jeff Duquette ]

I guess I’m not tracking – why is it unrealistic for an antitank gone to fire HE at troops that are 100-meters away? Seems like a rather reasonable thing to do to me. Regarding the post by Fifty-cal – yup, you took the words out of my mouth. Some antitank guns carried canister rounds in their standard ammunition load outs. Soviet 45mm PAK is a good example of a gun that should prolly have a canister round. Soviet 76mm also could fire shrapnel with the fuze set to “canister mode”.

Perhaps it is something I am doing during initial setup. However it seems like sometimes the guns start automatically limbered -- other times the trucks will take off without the guns being limbered. I'll have another look at this tonight -- your updated scenario file. Getting the guns to the village is pretty crucial to the Soviets being able to hold the village. Regarding the Maxim -- I have never been able to get it to limber onto the jeep. I always end up prolonging it to the village.