Is it too easy to spot AT guns?

[CMplayer]

Another way to approach this would be to reverse the 'action' in calculating who is spotted.

Instead of units making spotting attempts, the engine could calculate units as 'broadcasting their presence' at regular intervals. This equation would be modified by all the usuals: movement, fire, cover etc. but be non linear in terms of how much it is influenced by the number of potential spotters.

This would get past the problem articulated by BFC that units don't all try to spot at the same time, and hence subsequent spotting attempts can't have their chances reduced.

Just reverse the 'direction' in which spotting takes place. It can all happen 'under the hood' and it would still appear as if it were the way now. (except that a Pak-front would have a longer survivability against massed enemy tanks)

[ January 12, 2003, 01:31 PM: Message edited by: CMplayer ]

[kipanderson]

Hi,

as I did do a lot of testing for the spotting distances of AT guns in the beta version of CMBB it seems reasonable that I take my turn in defending them. But first, a few comments on Borg spotting, none of them very original.

Steve had already said that the new engine will have individual spotting, plus some form of networking to simulate radios in tank platoons and such. I believe this will deal will the problem just fine. Add to the above the possibility of multi-players games, team games, when you only spot what your “company or platoon” spots, and in my view the problem will be solved. Of course, the god like role of the single player means that he sees far more than any single commander in the real world. This has a bunch of implications. But once you accept that CM games are not command games, you do not play the role of only the company or battalion commander but also the role of all the squad and AFV commanders, you clearly must be able to see far more than any single real life commander. Even with individual spotting. I know some wish CM to become a command game, but I do not, and happily for me, Steve has made clear that CM is not headed that way.

When it comes to the spotting distance for AT guns I do believe they are reasonable. It may be the case that the smoke and flash free ammunition of today works quite well, but from what I have seen; WWII tank guns/AT guns produced a very unsubtle blast. Like many others here I have seen a lot of footage of WWII guns in the 75mm/88mm class firing, and the signature is considerable, even if it may be quite fleeting. The best way to describe it is to say that the blast just about covers the forward arc of the firing tank. When a Sherman fires, the blast is as large as the tank itself. You have to imagine what would happen if an AT guns fired.

If advancing units were observing a landscape, and remember they are scanning “for their lives” in an effort to spot danger at the first opportunity, a blast from a 75mm AT gun would be hard to miss. Assume the AT gun was dug in and the barrel about two feet above the ground. If it fired from a tree line there would be a sudden eruption of blast, fire, smoke, dust, foliage, twigs, snow and the like, some 6-10 feet in diameter out from the woods for twenty or so metres. Remember just how sensitive to movement the human eye is. Even very expertly camouflaged troops, once they move can and, do often, give their location away. Also consider just how much detail the eye can make out at, say, 500-1000m. When LOS discussions come up I often test my own spotting skills when out running/walking the dog in the woods. You will find that at 500m you can see even quite small branches on trees; at 1000m you can still see the individual trunks of small trees. If something were well camouflaged, and remained still, you may not spot it as even 200m, but once there was movement it is a deferent story. The blast from an AT gun would be quite extreme movement. Next time you are out, look at a bunch of vegetation between 500m-1000m away and try to imagine if you would spot a blast of smoke, fire, dust… and so on, the size of a Sherman tank. Even if only a fleeting blast.

Sadly, we cannot test this, so it has to be matter of judgement, views will differ. But in my view the current spotting distances are reasonable and realistic.

One quick last point. Do note that tanks firing HE rounds at AT guns will tend to fire ether too long, or too short far more than CMBO. In CMBB AT guns are more difficult to spot than in CMBO, and more difficult to hit with tank fire.

All the best,

Kip.

[John D Salt]

Originally posted by Leutnant Hortlund:

[snips]I think it is too easy to spot enemy AT guns.

One thing that hasn't been mentioned in the discussion so far is that, whether ATk guns are too hard or too east to spot or just right, the consequences of being spotted are usually irremediably horrid.

To my mind, the chance of destroying on-table guns is way too high. OR papers on the effectiveness of CB fire show that to destroy a gun pretty much needs a direct hit from a field arty size shell, or a near miss with a medium shell.

The problem here is that a CM gun is out of the picture when it is abandoned by its crew. In Real Life , a detachment that had been driven to cover could re-crew its gun when the incoming fire ceased, and the chances of the gun being knocked out would be relatively low. What's more, detachments from destroyed guns could and did help to serve guns still in action.

I suspect, though I cannot prove, that solving both problems would require a programming approach that allowed CM entities to collaborate to achieve certain tasks. Such an approach would, I think, be useful in lots of other circumstances too.

All the best,

John.

[CMplayer]

woops, ignore

[ January 12, 2003, 01:31 PM: Message edited by: CMplayer ]

[Das Reich]

I had a Reg Soviet gun under command of a veteran HQ with a +2 stealth bonus.

in one 60 second turn, he came out of hiding to kill three Panzer 38's, one Panzer IIc and nine infantry.

and has yet to be spotted.

[xerxes]

Keyhole.

[StellarRat]

I think the Borg spotting can be delt with by just chaining the spotting results up the chain of command to whatever level in the chain BTS decides the player is at. During the time it takes for the spotting results to reach the player's level only the TacAI of the units that have spotted (or been informed of the location of) the target should be allowed to engage the target.

For example: Say infantry unit A spots enemy AT gun B in some bushes. The TacAI decides unit A will fire at unit B (assuming A's current orders allow that) and a message is sent to A's HQ with whatever delay is appropiate. The next turn A's HQ sends a message to next higher HQ (the player) and also sends message to all of the units under it's command (with appropiate delay again) to begin firing at unit B (again all under control of the TacAI.) The next turn the spotted message has reached the player and he sees unit B on his map and can issue orders accordingly, these orders go down the chain (again with delays) and are executed. The only downside to this scheme is that the player may only find out what has happened after some engagements are already over. Maybe not too much fun sometimes, but very realistic. Also, the general in the game confusion level will rise and the players control will fall, but again this is very realistic. This will definitely make a good battle plan very important. Once the action has started your forces will be much more committed than in the current version of the game. Radios and good CC will become much more of an advantage.

[JasonC]

You don't need to undo borg spotting. You just need to gear the sighting chances properly to account for redundancy in units looking. It is a matter of getting the right formula, and doesn't need to be any more complicated than that. There are trade offs; you have to choose where to make them to get the more important things right at the expense of unimportant ones.

Take the example of the 1 tank vs. 1 tank and infantry platoon. Should the sighting chance go up by a factor of 5 because there are 5 units with LOS to the gun? No. You don't have to model information flow - there wasn't any information flow.

If the chances are all resolved independently as some % chance, say 10%, then five units become .9^5 or a 41% chance. That is too high. The 10% only "feels right" because it is tracking an artificial situation - one spotter - which practically never happens.

It would be better to get the chance of spotting wrong when there is only one spotter. It is already low. Engagement times will be low, with only one target. The difference between 5% and 10% is not significant in absolute terms, and it rarely happens anyway.

So, even if the "right" number for one spotter is 10%, you just ignore that as irrelevant and use 5% instead. Now when there are 5 spotters, you get .95^5 = 23% chance of spotting.

"But it still goes deterministic if the number of spotters is huge". Correct. So, limit the number of spotting attempts against a given unit in a given minute. Use the 10 with the best chance, and throw the others out. Set the individual chances artificially low, as discussed above.

Say the best spotters you have, in distance and such, would right now have individual spotting chances of 12, 10x2, 9x3, 8x3, 5x10, 3x20 percent, by whatever existing LOS formula is calculating spotting chances. Right now they all check and you get 86% spotting - Borged.

Instead, throw out all but the best 10 and cut each those chances in half. The chance just plummeted to 36%. That is still 3 times better than the best individual sighting chance. But it is life for the hidden AT gun, often enough and for long enough to give them their real role on the battlefield.

It doesn't matter how often you check or what the absolute chances are right now. Throw out all beyond 10 and cut the rest in half, and sighting chances will drop significantly. Without needing to remodel Borg sighting or anything so complicated.

The one thing that will still make for high absolute spotting chances, is when individual units have high spotting chances right now. Which is where you want to keep high spotting chances.

You might worry that a high chance for one unit will become only a modest one, because of the halving. That is true and it is the trade off. But it does not matter, because occasions when there is only one potential sighter and the sighting chance is high, are both rare and unlikely to be critical sighting resolutions for the overall flow of the battle.

There is no question the existing spotting is too easy, due to "borg" effects. In terms of the original poster's examples, it is reasonable that several unbuttoned tanks saw firing high caliber AT guns less than 400 yards away, firing at them.

It is much less reasonable that a few unbuttoned StuGs saw a small AT gun firing from over a kilometer away so easily. The cover was only brush, the StuGs have good optics, fine. It is still unlikely. The best tank aces say finding the ATGs was hard, in exactly such situations.

As for buttoned tanks fired on from good cover at 800 meters, locating every shooter and KOing them within 20 seconds, it is obviously just flat wrong. It is not an acceptable level of realism. It is broken and needs fixing, without an undoable major engine rewrite.

Tweak the formula by throwing out chances past ten "eyes" against the same unit, and cut all of the remaining chances in half. The problems that result will be minor compared to the problems that exist now. You will minimize the practical innaccuracies in the modeling.

You can't get that by obsessing over the lone-pair chance and its accuracy and ignoring the scaling up problem. It is faulty critical design variable analysis - the critical issue is not one on one low chance spotting being "just so", it is many on many spotting not being automatic.

Small absolute model inaccuracies about low absolute chances in rare and unimportant situations are perfectly tolerable. Large absolute model innaccuracies about sizable chances in common and critical situations are not.

If you are resolving them sequentially, it is possible to apply a "best 10" type rule. You just put the chance used in a buffer for that unit, using the full value only until the buffer is filled. After that, use the difference between the new sighting attempt and the lowest number in the buffer, if the new one is higher, otherwise no additional sighting attempt. When a difference is used, the higher value goes in the buffer in place of the old value.

E.g. if the buffer currently has 10 "5s" and an "8" comes along, you give a 3% chance and replace one 5 with an 8 in the buffer.

[ January 13, 2003, 03:12 AM: Message edited by: JasonC ]

[Hortlund]

I'm not sure I want to drop this subject. I still feel the AT guns have gotten the short end of the stick here for some reason. They are much more vunerable/easy to spot than they actually were if you go by reading combat reports.

SO I did a little test. Flat map, 19 fire lanes (tall pines used to block off the lanes). In each lane one German 50mm AT gun, and one T-34/76 M43.

At the end of each line I placed one terrain tile of every kind, that means: tall pines, woods, scattered trees, grain, steppe, rough, open ground, rocky, hedge, stone wall, concrete road.

To get a somewhat better average result, I also added 2 extra lines of woods, scattered trees, grain, and open ground.

I removed all AP and T ammo from the guns (for the v1.0 version of the test I forgot that, wich lead to 15 destroyed T-34s in one turn...sounds a bit much actually, but that is for someone else to test if they want).

I placed all T-34s 800 meters from the terrain tile where the AT gun was set up. During the orders phase, the AT guns were given an area target somewhere in the middle of the lane.

All AT gun and T-34 crews were set to regular.

The results:

ALL guns were spotted within 16 seconds from the beginning of the turn. All AT guns fired their first shot 2-4 seconds into the turn. That means that some guns managed to fire 3 shots, but the majority fired 2.

The AT gun cleverly hiding on the concrete road tile was spotted immideately.

Of the rest, worst off were the AT guns in wood-tiles, they were spotted after 11, 12 and 14 seconds.

Tall pines (15 seconds)

Scattered trees (12, 12 and 13 seconds)

Then came the rocky and rough guns (14 seconds each)

Steppe and open ground (12-16 seconds)

This is kinda surprising, because the last AT gun to get spotted was one hiding in open ground.

Behind hedge (13 sec) Behind stone wall (15 sec)

Grain seems to be best (15, 15, 16 sec)

All guns were dead at the end of the turn.

If I can be bothered, I'll run the tests again tonight with buttoned tanks to see if the numbers change.

[ColumbusOHGamer]

A gun that is placed 25m into the trees/woods is far less likely to be spotted than a gun that is only 5m into the trees/woods. None of the testers have said how far back the guns were in thier tests.

I suggest 2 45mm AT Guns, one at 5m depth and one at 15m-20m depth, (in foxholes) in woods vs. 4 PzIIIN's @ 500m distance.

I bet the AT guns win.

COG

[CMplayer]

Originally posted by ColumbusOHGamer:

A gun that is placed 25m into the trees/woods is far less likely to be spotted than a gun that is only 5m into the trees/woods. None of the testers have said how far back the guns were in thier tests.

This is, as you say, a very important variable, but if it's like CMBO you'll notice that AT guns firing from deeper in the forest also have their 'to hit' chance reduced, so the outcome is far from certainly in their favor.

If the tester could find an optimal depth to place the guns, that would be very useful info for ladder weenies.

[ColumbusOHGamer]

Originally posted by CMplayer:

This is, as you say, a very important variable, but if it's like CMBO you'll notice that AT guns firing from deeper in the forest also have their 'to hit' chance reduced, so the outcome is far from certainly in their favor.

You are correct. But isn't it better to take 3-4 shots to hit a target and still live to hit another target rather than get a kill on the 1-2 shot and immediately get pummeled by HE rounds?

Originally posted by CMplayer:

If the tester could find an optimal depth to place the guns, that would be very useful info for ladder weenies.

I was just thinking the same thing. Someone should post a chart of maximum (or optimal) firing depth for any gun (does it apply for MG/ATR/Sniper/Screck/etc. too?) in all the terrain types. Would make setup go WAY faster on Probe/Attack/Assult games.

COG

[JasonC]

What I am noticing from the tests and reports is that the caliber of the ATG seems to be the single largest factor in whether it is spotted. 45mm and smaller guns, I hear repeated stories of difficulty spotting them - though others say they are spotted quickly at short range (400 meters - as I'd expect) and sometimes at long range (1100 meters - the 3 unbuttoned StuG vs. 45mm in brush report). Meanwhile, every report so far in this thread where the ATG caliber was 50mm or more, the gun was -always- spotted, in about 15 seconds, at ranges out to about 1 km.

This suggests a different set of variables for the test. Not terrain type with gun caliber and range held constant, but varying by gun caliber and range. 88s at 3 km, 75 PAK at 2 km, 50 PAK at 1 km - that kind of thing. Expecting that the bigger guns need to be much farther away to avoid rapid detection - if indeed they can avoid detection at all. Also it should be tried with tanks buttoned. I think it is clear there are -some- cases where spotting is too easy, but it would be good to pin down the actual current behavior.

[Redwolf]

The tests above, where they done with foxholes for the guns?

Probably doesn't matter given how few effect CM foxholes have.

[YankeeDog]

Hortlund -

Thank you for setting up a test, but in order to get any kind of meaningful results, you really need to have a larger sample size than 1-3 runs.

There's a big random factor for spotting in CM. Statistically speaking, your results are identical for all terrain types because the sample size is so small. You are really going to need at least ten, and preferably more like 20-30 runs per terrain type to get the standard deviation small enough to have the results mean anything.

However, I have some questions about your test as set up. First of all, you forgot to mention whether or not the AT guns were dug in - that should make a BIG difference. Second, as noted, it makes a BIG difference how far back the gun is placed from a treeline - were your guns basically in the middle of the terrain tiles, and therefore all ~10m from the open ground? Finally, what were weather conditions? Are we talking clear mid-day or overcast at dusk?

Assuming good weather and lighting conditions, I acutally don't think your results are very far off. My gut feeling is that maybe the ATG should maybe on average be able to get off one more shot or so before being spotted, but that's about it.

Why? Well, If I understand your setup correctly, all of the tanks were facing directly at the Guns from the start of the test. CMBB accurately models the fact that tanks spot VERY well in a relatively narrow band to the front of the tank, ESPECIALLY when standing still with the turret facing straight forward and unbuttoned. In this configuration, a T-34/76 M43 has four pairs of eyes scanning forward. In fact, two of those pairs of eyes - the driver and asst. driver/MG gunner can only really see forward, so that' s where all of their attention is going to be focused. More importantly, at least one of the pairs of eyes, the TC, is probably using binoculars. Furthermore, the engine is at low idle, so at least the TC will be able to hear reasonably well.

Now, ATGs are REALLY loud. Even over the idle of a diesel engine with your tankers' helmet covering your ears, you are going to hear the gun report. The 50mm ATG fires a supersonic round. therefore, the TC is also going to hear the 'crack' of the sonic boom as the round passes by him, or he's probably going to see the round hit the dirt in front of him, so he's going to know that the gun he just heard was aimed at him (and he's DEFINITELY going to know the shot was targeted at him if it hits!).

The human ear is pretty darn good at figuring out the direction a sound came from, so I'm wagering that TC figures out that the round came from somwhere in front of the tank (probably about a 20-30 degree arc) on the first shot, even if no one actually sees smoke or muzzle flash.

So now on the second shot, all eyes are forward and actively looking for the origin of the shot. Maybe the TC is buttoned now, but if so, he's looking at the suspected area through the gunsight, which also gives him the advantage of optics. It seems likely to me that someone in the crew is going to spot a bit of muzzle flash, some kicked up dust or some gunsmoke on the second shot - they may not spot the gun itself, but I think it's very likely they're going to spot enough evidence to peg the location of the gun to within 20-30m or so. Keep in mind that they're not scanning the whole arc in front of the tank, just likely hiding places for a gun within that arc.

At this point the TC may well open fire. Why? Well, he doesn't need to have actually spotted the gun barrel amd shield to open fire on an AT gun. If he's caught a glimpse of muzzle flash, and he's pretty sure it's not a tank (which he probably would be given the fact that he's not seeing a big vehicle through the gunsight), my bet is that he's just going to shout "Round Up" and start firing. "Spotting" in CM (at least at the level of getting a "Light Gun?" contact) does not necessarily mean that the unit has actually seen the outline of the gun, it means that the unit has seen enough evidence of the unit's location and type to open fire.

So I can believe that sometimes the crew spots the ATG on the second shot. In fact, I think it would be realistic for one or two of your 19 tanks to get lucky and spot the gun on the first shot (a crewmember looking in the right place at the right time). Personally, under the conditions you set up I would say spotting on the third shot should be about the average with a very occasional tank taking until the fourth shot to spot the gun, but that's really a gut feeling based on inference and certainly not something I can provide concrete evidence for. In any event CMBB's spotting model isn't that far off my gut instinct.

So how do I reconcile this with all of the real-life stories both you and I have read about ATGs firing undetected for round after round, even when in LOS of multiple tanks? Simple. Very few of the stories I have read involve ATGs opening fire on stationary, unbuttoned tanks who are facing directly at the ATG. If a moving, buttoned T-34/76 facing at a 45 degree angle to the gun spots a PaK38 at 800m on the second shot most of the time, then something is wrong. In fact, if CMBB is working correctly (haven't tested, but based on play experience this seems to be the case), even a facing change for the tank of 25-30 degrees should make a BIG difference in how long it takes to spot the gun, especially if the tank is buttoned.

To me the problem with CM is more the fact that if a moving, buttoned platoon of T-34/76s facing at a 45 degree angle to a PaK38 is fired upon by the gun at a range of 800m, a Platoon in a patch of trees 400m away can spot the gun on the second shot (as it probably should) , and INSTANTLY the entire platoon of tanks knows where the gun is and opens fire. IRL, information like this would probably take several minutes to reach the tank platoon if both the T-34s and the infantry platoon commander had radios, and longer if they didn't.

Well, anyway, that's my .02.

Cheers,

YD

[Redwolf]

What I noticed in gun spotting tests is that the small FlaK guns, 20mm, 25mm and Axis 37mm were incredibly more difficult to spot than AT guns, where the Soviet 37mm was as easy to spot as the AT guns. This was done at constant range, I think 600m or so. The difference was not like twice as hard to spot, it was total - the AT guns always reliably spotted, the FlaK guns never.

My testing indicates that Jason is right on target.

The model in effect seems to be a simple table with ranges - practially ignoring terrain, foxholes, HQ stealth bonus and so on. Not sure I like that. If this is correct, a player figuring out this table can always have his guns fire just above the spotting range and gain a large and totally gamey advantage.

Experience of spotting units, optics/binocolars and buttoning status need to be investigated, but my gut feeling from playing is they are mostly ignored, too, when it comes to spotting of firing guns.

[White Phosphorus]

But even if one unit spots the gun the player does too, and on the next turn every unit he has will fire at the gun.

[YankeeDog]

Originally posted by redwolf:

What I noticed in gun spotting tests is that the small FlaK guns, 20mm, 25mm and Axis 37mm were incredibly more difficult to spot than AT guns, where the Soviet 37mm was as easy to spot as the AT guns. This was done at constant range, I think 600m or so. The difference was not like twice as hard to spot, it was total - the AT guns always reliably spotted, the FlaK guns never.

Hmmm. If so, then CM would definitely seem to have it bass ackwards in this respect. Assuming similar caliber, I would think a Flak gun should be much easier to spot than an AT gun. Flak guns usually have a higher profile than AT guns of the same caliber because they're designed to track rapidly moving targets, have a higher ROF, etc. I'd have to go look at pictures and ideally even film footage of the guns firing to develop an opinion as to whether, say, a 20mm AA should be easier or harder to spot than a 50mm AT, but a 37mm AA should definitely be easier to spot than a 37mm AT.

I would be very interested to hear any hard evidence about this.

Cheers,

YD

[Redwolf]

Originally posted by YankeeDog:

I would be very interested to hear any hard evidence about this.

Not needed.

It is outright totally obvious. Just run any test comparing the guns I indicated. The difference is so big you won't have any doubts. I'm just not quite sure which range exactly I had, but 600m should do fine.

[JasonC]

OK, I wanted to try a test that is a situation I think actually matters. Running with my "gun caliber" idea, I set up the following fight.

Time - September 1943. Southern Russia. Day, clear, cool, dry. Flat, open, 1.25m changes, rural. Steppe country. The auto gen map still had significant patches of trees on it - these were all painted over with steppe. Brush, rocky, occasional rough patches, one house, a bit of dirt road angling on the map and back off - you get the picture.

The Germans have 2 veteran 88mm FLAK, dug in, in large patches of brush. Each is slightly over 20 meters from the edge of the brush (the patches are at least 40m deep). The map is 480m wide, with a single large flag behind the gun positions, which are about 200 yards apart.

The map is 3.2 km long on the east-west axis. Initial ranges, with set up zones, are about 2.8 km. The Russians have a company of T-34/76, M43 late model, at the far end of the map. Meaning, 3 platoons, plus 1 independent tank. Regular quality.

If the Russian player knows what he is doing, the 88s are toast and the objective is his. He will probably lose 3-6 tanks accomplishing this, which may give a score result (with the large flag) of "draw" or "tactical victory". But he will only lose to the 88s if he plays his hand incorrectly. I submit this is rather generous to the T-34s.

How does it work? The 88s have 30 odd AP each. Their hit chances, once the T-34s start moving, are very low at 2.5km and upward ranges - a few percent. If they fire off all of their ammo at such ranges, they will get a couple of T-34s then run out and be overrun.

At 2.5km, the first few shots will not lead to any actual spotting. But within a minute or so, the T-34s will have "sound contacts" on at least the nearest gun, even at that range. If the T-34s attempt to fire on the move to suppress the guns by area fire, that is about the only way they can lose. Their shots will miss comically long or short, because of firing while moving. And the noise and explosions of their own rounds will prevent complete ID of the gun positions.

The other way that can lose for the T-34s is trying to charge all the way to 750-1000 meters before engaging. The last parts of that, the 88s hit chances will soar high enough to take out the onrushing T-34s, and run the table.

The correct solution is for the T-34s to fast move, ignoring the incoming and the sound contacts, until they reach about 1.5km. Then hunt. Resolutely unbutton every turn. The 88s will be picked up rapidly once under 1.5km. And once located, they will be suppressed within a minute by 76mm HE, and soon KOed. Their own reply chances are good enough that you should expect to lose 2-3 T-34s at this range. But the 88s are toast once you get the bulk of the company to 1500m.

It is also possible to play is slightly differently, fast moving to 2km or so, and then firing at the sound contacts with platoon volleys of HE, from "short halts", while other platoons continue to hunt. At that range, providing you aren't firing on the move, you can pin an 88 with near misses. In addition, their reply hit chances are still low. They will use up their ammo rapidly.

But this is not as certain as the previous method, because you do not have full IDs of the guns. When they pin, they will cease fire, and you will lose the sound contact in a minute or so. Meanwhile you will lose some tanks to their replies, even at such range. Firing from halts, you are a better target than moving. The match up is actually superior at 1500 yards, where you have full ID and can overwhelm each gun rapidly.

As for the defender's side, with the 88s, none of the fire discipline options available to you are satisfactory. Opening at 2.8 km, your hit chances against moving tanks will start around 3% and climb only to about 7% at 2km. You can easily blow your entire load of AP at extreme range this way, and KO only 2-3 T-34s.

Waiting for good hit chances is not a solution. At 1500 yards, your initial chances against moving targets are only around 15%. They will rise to more like 25% as targets on "hunt" halt to return fire. But you will be picked up so rapidly at these ranges, that the first gun "gang tackled" by the tanks will only get a couple of shots off. You will be lucky to bag one tank. The problem is compounded by fog of war and the death clock - you are likely to overkill the first tank, not hit a second, if you live slightly longer.

If you want much longer, for very close shots, initial hit chances go on climbing. But your life expectancy after opening up drops to less than 30 seconds. The overkill problem gets worse. You are not making progress, therefore.

The optimal solution seems to be to have around 10 AP or less remaining when the tanks pick you up around 1500 yards. You want to expend the rest of the AP loads in 3-4 minutes of firing, at ranges around 2 km to 1.5 km. If they are hunting, that means opening around 2 km; if they are fast moving, a turn or so before.

You can bag half the company this way, but you'd need to be quite lucky to actually run the table. The critical variable is to expend as much of the ammo as possible, before the tanks get close enough to correctly ID your exact locations. But to cut it close enough that you still fire some in the duel after they halt. You have to watch for the alternate "area fire" tactic, though, and reduce ammo expenditure = re-hide if the tanks slow their advance or stop. Otherwise you will run out of AP too soon.

Now, is all of this historical? I submit that it is not. That T-34 company, without supporting artillery to suppress the 88s, or anything ahead to scout out their locations, or any folds in the ground to hide in during their approach - ought to be toast. Part of the problem is definitely the accuracy of their HE replies, and the ease with which they therefore KO dug in guns at ranges well over a kilometer.

It is not as far off as it might be, though. It is not like the 88s are firmly IDed at 2.5 km. Only sound contacts are available - although sometimes ones good enough to area fire pin the 88s. While the tankers would certainly know something was firing at them, at 2 km the sound of the report at the barrel is several seconds behind the shell and any flash - a disconnect that many tankers identified as making it difficult to detect supersonic AP shooters at long (mile) ranges.

I'd like to see a similar set of tests with the 75mm PAK 40. It had a much lower profile - but is still a high velocity, big bore gun. I suspect the engagement range questions will be only marginally better - perhaps full ID out to 1250 yards, say. Penetration issues may make it useless for them to open up at fully 2 km.

The trade offs will wind up being much the same. They will hold their fire until 250-500 yards behind the range they can be picked up, fire for a minute or two of "approach", and then live long enough for one minute of dueling - or less - once in "ID range."

It also means tankers should learn their typical ID ranges to various enemy PAK at least as rigorously as their penetration ranges against various enemy tank models. Because the proper closing behavior is critical to dealing with ATGs effectively. Done right, they are no more than a minor threat, at least to superior numbers of decent medium tanks.

[Guest Mike]

Just a point - IIRC 88mm HE is perfecty capable of destroying T34's too - although it's not as accurate as the AP (lower MV).

So the limit on AP rounds isn't all THAT important.

[JasonC]

That's true, and a fine point (lol). A nice thing about 88s compared to 75 PAK, that. Actually, the MV of the HE is even higher, so the AP is almost pure "armor overkill".

What I am noticing in these tests, though, is that the area of "probability behavior" seems quite limited. You are either in the range where a full spot is possible, or you are not. If you are, somebody will see the thing very soon.

Things like "?" commanders and smaller caliber may reduce the range for spotting becoming possible. But the behavior, once over the range threshold, is pretty much 1-0, on-off. I'd guess because there are so many spotting checks being made, that once the chance of a full one goes positive, somebody "hits" in short order.

[Redwolf]

Jason,

one thing missing from your scenario is that CMBB has a more decent model of zeroing in than CMBO. In CMBO zeroing was lost on the slightest move of target or shooter, in CMBB you seem to have an area around previous shots. This may or may not make shooting at the long "safe" range an option.

Of course, in reality this kind of picking off with zeroing in would work best because you would often mine the fast moving lanes at distances where you would open fire with your AT guns, forcing the attacker to slow down. But in CM you are not allowed to do that, unless the scenario designer is very fancy.

On a related note, has anyone ever been able to observe the hit probablity benefit that defending vehicles that never moded are supposed to enjoy in CMBB?

[Scipio]

Originally posted by Das Reich:

</font><blockquote>quote:</font><hr />Originally posted by GRUMLIN:

AT guns are NOT too easy to spot.

<snip>

Likewise, have just finished a PBEM where a 150mm inf gun of mine was in woods and fired all game in sight of nme without being spotted - maybe because I had a +2 stealth section leader covering him. Try using the stealth bonus on guns and see your kill ratio go up.....

Good point, company and Battalion HQ's with good stealth would make them harder to spot. Problem is, its hard to have a HQ for each gun. They should probably be set up in batteries so that one HQ can command 3 or 4 guns or something.</font>

[Redwolf]

Stealth-bonus HQ control do not make firing guns harder to spot in CM.