Gonna need explanation

[Dakpilot]

iirc the less guns with more accuracy to save fps was also the case in 1946, this solution might be carried over into this game.

 

Yep, and It must also be remembered that PC hardware and game performance optimisation has come a long way since BoS earliest release with regards this design decision/option

 

Cheers Dakpilot

[unreasonable]

I don't think you can use that argument as I have seen more than one mission maker say that they could not put in the correct number of guns as it affects frame-rate badly so they user fewer guns with more accuracy per gun.  That actually helps the attacker because if 6 guns have the firepower of 12 then each gun you destroy takes away the firepower of two.

 

 

<snip>

 

That one gun - certainly on Normal or High - is more effective than one RL gun is not an argument: it is a fact.

 

I have no objection to there being high AI levels - if mission makers wish to use a 1 gun = 2 or 4 or whatever I have no problem with that: I would simply like to have the option to use 1=1 and get a vaguely realistic result.  Sometimes I get the impression in these debates that this is something that people actually do not want: cannot fathom why.

 

For instance, in my developing model of JG3's attack on Eindhoven as part of Bodenplatte, I would like to place 36 guns, since that is my best guess of how many there were at the field. (If anyone has contrary sources, please let me know). 

 

Having tried that, it certainly does slow the game down: at least until most of the attackers have been shot down.

Edited January 2, 2018 by unreasonable

[6./ZG26_Custard]

 

 

Sometimes I get the impression in these debates that this is something that people actually do not want: cannot fathom why.

 

With our squad flying online in multiplayer as Luftwaffe (most of the time) we have flown many hours in 110 - E2s and Stukas against targets with high AI AAA  we have developed our own tactics that in the main work quite well but the laser guided accuracy of some of the AAA on multiplayer servers can make more than a frustrating experience at best.

 

A more realistic AAA would be very much appreciated.  The K-61 in particular is completely OP when set anything above low. For those individuals that believe tactics are the problem, i would recommend flying an online campaign in Stukas or 110s exclusively.

 

We have also had experience flying IL2s and Pe-2s with much more satisfying results. One of the main reasons is we didn't have to fly against the K-61.As your tests have pointed out, even set on low, the AAA in game is twice as effective as its real life counterpart. I don't want to sound like I'm moaning, I would just like the AAA to be more realistic. I understand that in a game you will have certain limitations but hopefully in the future we will have something that emulates AAA more effectively. 

[7.GShAP/Silas]

People don't believe it until they try it and those who surrender themselves totally to fighting gamey AA with gamey tactics don't care.

[von-Luck]

My most recent hilarity concerning AA was the fact that it shot me down while I was dogfighting an LA 5 in high G maneuvers - lost my wing while performing rolling scissors at 350+ KPH! The AA/AAA in this game is frustratingly wonky. Maybe one day they will fix it.

von Luck

Edited January 3, 2018 by von-Luck

[hames123]

Well, try attacking a Tank Column on TAW with a Stuka, 110E or 129, while Fighters are providing Anti-Fighter Support. In a Formation of 4 129s they were able to Kill 2 of us before we fired the first Shot and only 1 got out alive, limping back home on 1 Engine. And that's a regular Occurence. And we are a good at what we do.

That's on a day with Overcast at 2k.

Your lot fly the cannon-birds? I expect that your losses are pretty heavy in both planes and pilots. Do you often get many T-34s though?

[Y-29.Silky]

You have to change the AAA mechanic in this game

 

I tried too... 

https://forum.il2sturmovik.com/topic/12180-aaa-fix-more-immersive-experience/

 

 

If after 4 years it's still the same, it's never going to change. With that said, the golden rule in WWII aviation was to never make a 2nd pass on an airfield. 

Edited January 11, 2018 by Y-29.Silky

[6./ZG26_Klaus_Mann]

Your lot fly the cannon-birds? I expect that your losses are pretty heavy in both planes and pilots. Do you often get many T-34s though?

Yes, we do, at least 2 Tank per good Sortie. 2 KVs in a perfect Run. 

[Lusekofte]

 

 

People don't believe it until they try it and those who surrender themselves totally to fighting gamey AA with gamey tactics don't care.

 

I have no problems believing anyone. I fly only ground pounding, with IL 2. P 40 and PE 2. Down low no one of these excell in the meeting on AA and AAA . I admit the constant problem with a wing blown away in a 110 give frustration a extra boost.

Still I like AAA to be hard, not easy. At the same time I too expirience the accuracy as a problem. I just am afraid they will become harmless , and if so I prefer it the way it is 

[Royal_Flight]

What are people's thoughts on ambient flak?

Where, for every gun that fires, there are maybe two to four times as many visual effects, either tracers or smoke bursts, but they aren't objects that can interact with the sim world and they don't cause damage.

 

It doesn't quite fix the problem of accuracy per se, but what it would do is allow for AA to be set at a lower skill level so it doesn't totally clean up anything that goes near it, but there is still the sense of flying through a furious storm of fire and metal to press home an attack, and there's still the risk of being hit and taken down anyway so it shouldn't make people complacent.

 

If I remember correctly there are already ambient explosions going off on the ground in Stalingrad, it would be the same sort of idea as this.

[CrazyDuck]

Historically, IL-2 sturmoviks were employing the tactic to cross enemy lines, U-turn towards enemy and attack them from behind (flying in the direction of friendly territory). Main reason for this was to have the possibility to extend and glide to friendly forces whenever heavily damaged by flak.

 

This reason is often overlooked today, attacking the enemy tanks from behind being stated as the main motivation.

 

Coming from heavily armoured CAS airplane, this speaks volumes about how deadly short range flak was at that time.

Edited January 15, 2018 by CrazyDuck

[A_radek]

I agree the behaviour of current AA is very simplistic, shooting through forests are among those things that annoy me. And it's not easy for an AA gunner to hit an aircraft flying at 2-3km altitude even if that aircraft is going straight and level. So primarily it's the long range accuracy of the AA I see as a problem and not short range. Despite that short range target travelling fast or maneuvering.

However (and this is not bashing those employing thought through tactics while getting murdered by aa) coming in at treetop level and rather than dropping, placing a bomb on the roof of a tank should be utter suicide. In my opinion it's not with Ai skill at low.

I base my opinion not on historical stats but on having served as an AA gunner during my military service. While the 40mm bofors was in a "stridsfordon 90" I also got to shoot the old trailer version on towed targets. I don't know if it back in ww2 was hydraulically or electrically powered. And of course it didn't have a computer compensating for all external factors though would be very surprised if trained gunners back then didn't take in to account wind direction/speed, shell temperature and so on manually. My point is:  Imagine the last airshow you've been on with various aircraft doing low fly-by's. That's what many bomber attacks online look like from the ground. Could you hit that aircraft with a rifle? Then you absolutely could with a 40mm and an old fashioned iron sight.

Edited January 15, 2018 by a_radek

[F/JG300_Gruber]

We have also had experience flying IL2s and Pe-2s with much more satisfying results. One of the main reasons is we didn't have to fly against the K-61.As your tests have pointed out, even set on low, the AAA in game is twice as effective as its real life counterpart.

 

I won't say this enough but if the ingame 61k actually had to reload their 5 round clips like it's real life counterpart, it would go a long way to reduce it's OPness

I hate that gun so bad...

[A_radek]

I won't say this enough but if the ingame 61k actually had to reload their 5 round clips like it's real life counterpart, it would go a long way to reduce it's OPness

I hate that gun so bad...

 

Googling pictures of the ki-61 loading, it looks similar to the up-right trailer version of the 40mm bofors. If it is, you would have two loaders standing on the cannon behind the gunner(s), constantly feeding clips down the gun rail from the smaller ammunition stowage behind them, also rotating together with the cannon. You would then have two or more guys running between ammunition crates and cannon mounted stowage constantly filling it up. So the clip size is limited to 5 rounds for easier carrying rather than a rifle:ish limit before some reload. It's the heat and needed barrel change that would limit you. I don't remember how long a change took on a bofors. But do remember it was not a big operation and barrel weight was 160kg. Double check me on that number, these are 20 year old memories I'm going by.

 

[unreasonable]

I agree the behaviour of current AA is very simplistic, shooting through forests are among those things that annoy me. And it's not easy for an AA gunner to hit an aircraft flying at 2-3km altitude even if that aircraft is going straight and level. So primarily it's the long range accuracy of the AA I see as a problem and not short range. Despite that short range target travelling fast or maneuvering.

 

However (and this is not bashing those employing thought through tactics while getting murdered by aa) coming in at treetop level and rather than dropping, placing a bomb on the roof of a tank should be utter suicide. In my opinion it's not with Ai skill at low.

 

I base my opinion not on historical stats but on having served as an AA gunner during my military service. While the 40mm bofors was in a "stridsfordon 90" I also got to shoot the old trailer version on towed targets. I don't know if it back in ww2 was hydraulically or electrically powered. And of course it didn't have a computer compensating for all external factors though would be very surprised if trained gunners back then didn't take in to account wind direction/speed, shell temperature and so on manually. My point is:  Imagine the last airshow you've been on with various aircraft doing low fly-by's. That's what many bomber attacks online look like from the ground. Could you hit that aircraft with a rifle? Then you absolutely could with a 40mm and an old fashioned iron sight.

 

No one has to base their opinion on historical stats, but if you do not at least check your opinion against historical stats you have no reason to be confident about your conclusions. 

 

Take the Bodenplatte Eindhoven attack for instance:  JG3 (with about 60 aircraft) strafe a base defended by 3 RAF Regiment LAA Squadrons (which I think means an OOB of 36 40mm Bofors, but might have been a few fewer), for about 25 minutes, pilots making at least two passes.  The LAA was credited with 5 aircraft shot down, which is in line with the GAF loss assessment.

 

Even if you assume that the first pass was free because of surprise, that still means that out of some 60-100 opposed strafing attacks the LAA scored about 5 hits: ie somewhere in the 5-10% range.  Try doing that mission in BoX and see what happens.

[A_radek]

Unreasonable, I have seen that example mentioned but my opinion stands.

 

There's just to few of the circumstances mentioned there and I strongly suspect something is missing from the picture.

 

While loaders could take a break I assume Raf procedure was to have gunners taking shifts manning the cannons. Still like you said the first pass could have been a complete surprise. (I remember the surprise I felt constantly scanning for targets only to find, as if from nowhere, a bright red helicopter hovering behind us signaling we're dead)

 

But for the second pass, let's say these early bofors only fired 2 shells a second. From those 36 cannons in total, that's still 72 new shells each and every second filling the sky. And you would only need 1 hit. If this was a low level attack you have my reason for doubting this example can be taken as a clean one of 36 bofors cannons and an attacker making several low level passes.

 

As for higher altitude strikes. Or when gun/target is horizontally further away, Like mentioned in my post, I agree the aa is too accurate.

[unreasonable]

Of course we do not know the precise circumstances, how many guns were manned etc: but the basic facts are clear. It was a low level strafing attack, in the Eindhoven case with very little interference from defending fighters.  Read "Bodenplatte- the Luftwaffe's Last Hope".

 

Another airfield attacked in JG strength was Ophoven. 2 LAA squadrons claimed 4 destroyed and 11 damaged, almost certainly an overstatement, since the unit that attacked - JG11 - had 25 aircraft missing, with 25 claims made by allied fighters against them.  In the Eidhoven case, the LAA claimed 20 destroyed and 19 damaged. Post battle assessment credited them with 5 aircraft destroyed.  Clearly the LAA hit some of the attackers - but they did not inflict the kind of casualties you seem to expect. The same overall story is true of every airfield attacked.

 

The data suggests that aiming at a fast moving, low flying fighter is nothing like as easy as you seem to think - or as BoX models.

Edited January 15, 2018 by unreasonable

[F/JG300_Gruber]

Googling pictures of the ki-61 loading, it looks similar to the up-right trailer version of the 40mm bofors. If it is, you would have two loaders standing on the cannon behind the gunner(s), constantly feeding clips down the gun rail from the smaller ammunition stowage behind them, also rotating together with the cannon. You would then have two or more guys running between ammunition crates and cannon mounted stowage constantly filling it up. So the clip size is limited to 5 rounds for easier carrying rather than a rifle:ish limit before some reload. It's the heat and needed barrel change that would limit you. I don't remember how long a change took on a bofors. But do remember it was not a big operation and barrel weight was 160kg. Double check me on that number, these are 20 year old memories I'm going by.

Thanks for the info  

Edited January 15, 2018 by F/JG300_Gruber

[A_radek]

No we can't know the complete circumstances. Pure guesswork of mine, but I would assume many were caught of guard, some ran for cover and one guy couldn't hit a barn anyway. As the cannons were most likely spread out over the area the remaining "functional" ones did not all simultaneously have a clear shot against the low flying targets due to terrain. And in the end, we don't really know how many cannons there were.

 

Yet these were not the circumstances set up in the BoX test to determine what ai level AA should have, to roughly match statistical outcome.

 

The data suggests that aiming at a fast moving, low flying fighter is nothing like as easy as you seem to think - or as BoX models.

In BoX, I can, in a fighter, alone, draw the Ace AA of a target area to make life easier for the bombers behind me. Sometimes Aa gets me but going fast and maneuvering the risk I'm taking is far smaller than the one bomber guys behind would be taking going in first. So in your opinion would that not be correct AA ai behavior? Me alone going right through and getting away that is.

 

I'm enjoying the discussion but I don't think we'll agree. Your basing your opinion on how accurate Aa ai should be on historical stats that I don't think tell the whole story. I'm basing mine on having lot's of practice with these cannons against moving target, albeit with aim-assist. (Even got to laser an old j-35 draken pretending to attack us a few very low passes, dying that day was awesome) And my assumption lies in how easy it would be to aim and hit with a stripped down and slow firing version.

 

And once again. I don't agree with current BoX long range Aa accuracy. For the sake of this threads constructivism.. (?) I think it's important to specify what ranges we are talking about when wishing for a nerf or boost.

[unreasonable]

I agree with the last point about being specific - this is why I have generally limited my comments in the current round of AA discussion to the low flying airfield attack case, for which we have data from no less than 16 airfields that came under attack, with OOBs and loss estimates for both sides.

 

The data is as good as we are ever going to get for a wartime event 70 years ago, and is freely available in the book I mentioned previously. I would tabulate it all in summary form if I thought that might help, but if you genuinely prefer to base your thinking on your personal experience of never having shot down an aircraft with a Bofors I am not sure that it will.   You never have actually shot down an aircraft, right?

 

The test I set up - also posted so that other people can check it, along with the results of repeated test runs, in Custard's "Bravery of being out of range" thread, is based on a schematic version of JG3's attack plan, except using only two strafing runs per plane instead of at least three, to allow for the idea that the initial runs were unopposed.  It uses 8 guns - obviously a K-61 is not a Bofors, but that is why I ran comparative tests.  The best statistic to look at is actually something like losses/gun, or even better, losses/gun/pass.  Roughly speaking the number of passes made (ie 2 in this test) times the number of guns = the number of good firing opportunities.  Ammo expended per hit being more a matter of efficiency, and whether the crews expend huge amounts of ammo on low probability shots - as the AI does.

 

As you can see in the High setting of the base 109 vs K-61 tests the guns actually score more than 1 losses per gun/pass - they can change targets so quickly and accurately that they will sometimes hit two aircraft from a formation in the same attack. The low setting is still a respectable score - the guns will get a hit almost one in four times.

 

You are right that we do not know the exact number of guns that got into action or the exact number of passes made in the historic cases, but we have a pretty good idea. Excluding a "free" first attack, if the 60 aircraft of JG3 made on average one and a half more passes, and 30 of the 36 guns on establishment got into action, 5 confirmed kills is equal to  0.17 losses/gun, and 0.11 losses/gun/pass. Ie about half the test results on "Low".  You would have to assume that less than half of the guns got into action to get similar results to the "Low" test.

 

This is why I have said before that "Low" is order of magnitude OK as a historic setting for LAA, but probably about twice as good as it should be. It could perhaps represent excellent gunners in perfect conditions. Once you get to "High" you are an order of magnitude out.

 

Edited January 16, 2018 by unreasonable

[Dakpilot]

One wonders the readiness state of Air force staff on New year's Day early morning   even at a "frontline" base at that stage of the war

 

Bodenplatte was indeed an unexpected 'surprise attack' 

 

Cheers, Dakpilot

[A_radek]

No I have never shot down an aircracft or participated in any kind of real conflict. I have however shot down 2 towed targets, one at a ~4 km range (pure luck even with the modern electronics) and one at very close range, perhaps 4-600 meters right before it exited the firing window. These towed targets were cylinder shaped fabric tubes about 8 meters long ~50cm diameter at root and 80-90cm at end. Bright red and carried a proximity sensor that in realtime would tell us how close our shots came. 4, 8 and 12 meters was the results it measured. Naturally It was every gunners goal to actually hit the thing, destroy something expensive and hopefully end a cold day. We had a surplus of direct hit HE rounds and ran these target practice days quite often coupled with simulator days when weather didn't allow the tow plane to get up. Just sharing memories, one could easily point out any experience of mine is worth nothing in this context as the old bofors only shares base mechanics or rather the gun itself with the one I used. And even that is improved upon.

Though I've never shot down an aircraft, and admittedly my opinion is just based on assumptions, I think you can understand why I have formed one from the experiences I have. The numbers you present make sense and I take good note that you take in to account first pass could have been unopposed.

I remember a video posted in a thread discussing the op:ness of current Aa. Some guy captured his ground attack on the random expert server, alone, making several passes while Aa at very short range was hardly impenetrable. I think you would agree with me by common sense, historic statistics aside, a relatively big and slow target like a pe-2 that close to the ground with those amounts of Aa does not make sense. I suspect he knew exactly what he was doing and had discovered constant direction changes throws of the simplistic predictive aiming of the ace Ai. If you were to agree with me in that single case then obviously something basic is wrong. Your test at "Low" reflects the historical statistics. I'm saying AA  at short range would be too ineffective if lowered. So for a second presume my unscientific "feeling" is not wrong. Would that not imply other factors are to blame rather than decreasing Aa accuracy in general?

 

Again it's the close vs long range accuracy I see as a problem, not general accuracy. And imagine if ace ai were to remain what it is up close while exponentially getting weaker as range increases. I don't know what ranges the Aa in your test scored the hits, I presume it was not as the fighters passed by the cannon? This kind of change would make me happy and suspect it would not have to change the overall outcome of your test, thus still matching historical stats. Only difference is that kind of change would take my assumptions in to consideration.

[unreasonable]

One wonders the readiness state of Air force staff on New year's Day early morning   even at a "frontline" base at that stage of the war

 

Bodenplatte was indeed an unexpected 'surprise attack' 

 

Cheers, Dakpilot

 

You are quite right - the accounts say that the first passes were pretty much unopposed  at some airfields, but as the attacks took place over 15-25 minutes with multiple runs we can adjust for that,as I did in my test mission.

 

@ a_radek - I respect your experience, certainly from the pov of what it is like to operate one of these guns. Having people on the forum who know what they are talking about is invaluable. 

 

In terms of the quantitative success of various weapons systems, however, we need a real world, results based check. I am sure you are familiar with the over-claiming of fighter pilots, bomber gunners etc: even in action, weapon operators have difficulty assessing the actual combat effectiveness of their systems, for perfectly understandable reasons. Extrapolating from training is even more difficult. Hence, the need for OR - Operations Research - where nerdy young men who have never seen action, in rooms far from the front line, look at the data and come up with revised targeting schedules, etc.

 

I did not specifically record the circumstances under which the shoot downs tended to take place, but having watched from the guns pov on many tests - and I have now run over fifty trial runs of various permutations of the test mission   I am fairly sure that the majority of hits - especially on "Low" - do in fact occur at very short range, especially when an aircraft has just flown over, or nearly over, a gun and is heading away.  At that range the built in variation in the group and prediction matters least: the target is in effect larger and has less time to move off line from the predicted path. Long range hits are very rare except at the end of the mission when aircraft are flying away fairly straight and several guns are firing at the last few aircraft to leave the scene. The guns are very quick to get onto the new target heading when an aircraft flies over them - something that I suspect is at the root of the unrealistically high hit rate.  You can run the test mission and watch from the pov of the guns and you will see what I mean, but to actually count the ranges you will have to make tracks and go through each hit.

 

TBH, I cannot take cherry picked examples of game play from MP seriously. I do not agree with you that your example shows anything is wrong. A Pe-2 is not a slow aircraft. The historical facts are that the vast majority of passes made over heavily defended airfields were not hit. We get comments like "attacking an airfield is suicidal" which are, strictly speaking, complete and utter rubbish, no matter how often they are repeated, sometimes by people who should know better.  Attacking an airfield was dangerous, no doubt about it: but what is "dangerous" for a real pilot is quite different from what is "dangerous" for a gamer who can reload (and has already done so hundreds of times in his virtual career).  

 

The AI predictive shooting is not simplistic: is is aided by a computer more powerful than anything imagined in the 1940s, which has perfect knowledge of the targets position, direction and speed.  Watch the AI fire at a target changing it's heading in three dimensions - the computer AI is changing it's point of aim for every single shot, only "nerfed" by the AI level dictated RNG.  It gets on target amazingly quickly, especially when changing targets.  Actually, it might get more long range hits if it predicted a flight path and then fired a short burst at a single point to make the target fly through the stream of shells: which is in effect what is happening with the short range low deflection shots. 

 

There may be some problems that make the AI unrealistically poor in some specific situations that can be "gamed": target selection might be one of them. The AI tends to select the nearest target and then get target fixated, firing at a target for some time while other, closer, targets are closing in on it from behind. A well trained gun crew might not do that but stay orientated towards the oncoming attackers.  Statistically this does not matter in my tests because the AI is so good at spotting targets behind them and shooting them down just after they have passed overhead, but this could be exploited in MP. But then again, real gun crews in war do not make optimal decisions very often.

Edited January 16, 2018 by unreasonable

[sniperton]

The guns are very quick to get onto the new target heading when an aircraft flies over them - something that I suspect is at the root of the unrealistically high hit rate.

^This. When over the target your position relative to the gun is changing so rapidly that it should be physically impossible to adjust the firing vector accordingly. (IIRC several USN pilots at Midway remembered that the only moments they felt safe were when they flew over the deck of the target.) Another question is whether AAA should fire at outgoing aircrafts or should concentrate on incoming attackers instead. I'm sure that target fixation played a role in RL as well, but I'm also sure that doctrine dictated to give preference to incoming attackers posing a direct threat to the gun and its personnel. Just speculation   

[unreasonable]

You can get my original mission here, about half way down:  https://forum.il2sturmovik.com/topic/32578-bravery-being-out-range/page-2

 

Save in "My Missions" folder. Uses the Kuban map which I see you have. It does not have a player, so when it starts, press the keys for next air unit (Ctrl F2) or next ground unit (Ctrl F5). Repeat presses to cycle through.  There is an Fw190 that is just there so that you can count the survivors easily.

 

Note that it says 48 aircraft but for some reason best known to the ME two of them do not attack - "bad waypoint" - same two in every variant of the test. The summary table reflects that adjustment: the number of planes hit is unaltered, I just had to alter the denominator for the percentages.

 

The other versions, the results of which are summarized in the table in this thread, just change one variable at a time. I have not posted them, since only a few people have downloaded the test and no-one has made any specific comments, so I assume that people are happy to take my results on trust.  But would post some other versions on request.  I am currently half way through testing versions with a player plane (set onto autopilot once the mission is started), since it has been claimed that the presence of a player would make everything completely different. So far I can see no evidence of this at all: it does take a lot of tests to be sure, however, since 10 tests of 46 aircraft = 460 sorties, and 920 passes (less those shot down first pass) so I can get a fairly robust average even though the number of aircraft shot down on each test may vary from zero to 5+ on the "Low" test. (Actually up to 46 if everyone on the planet ran the test full time until the red giant sun consumes the earth )    10 tests looking at what happens to a single aircraft has a lot more potential for random variation.

 

BTW good discussion, I appreciate your comments - my own opinion is that of the settings we have "Low" is the only one that can claim to be close to RL on a one game gun = one real gun basis, in terms of the general likelihood of scoring hits. Whether this generates the sort of game-play that people enjoy most is of course another matter. edit: I also recommend Retnek's post #58 in the same thread - lots of interesting data. 

Edited January 16, 2018 by unreasonable

[A_radek]

@unreasonable

 

While trying to edit my last reply (I had screwed myself over with my own wording in the last paragraph :D ) I accidentally deleted the post.

 

Short version without the fluff: Can I please have a look at the mission file? And thank you!

[unreasonable]

Click on blue highlighted link in my post #65 above yours in this thread. Go to post #45 in the linked thread.  Click on the blue "Airfield defense" zip on the bottom of the thread. Everything is in there. Unzip in your "My Missions" folder to watch in game. 

[CUJO_1970]

Flying online in DED server this morning  on a bad weather map with rain and low visibility, it seems flak is continuing to track you like it's a clear day.

[Livai]

personal attacks towards the AAA mechanic -

 

You need to undersand the AAA mechanics. Not sure if somebody ever notice that the AI always set the first enemy plane that enter their range under endless shower of shots as long it stays in their range - Means for the other planes clear approach to disable the AAA mechanics. 

Edited January 16, 2018 by Livai

[A_radek]

@unreasonable. I trust you on the results and considering how many rounds you had to do to get a reasonably stabilized average result for every test, understand you have put much time and effort in these scenarios.

It's not to question your testing procedure I asked for the mission but figured it would help me watch in what situations 'low' gunners make a successful hit and how they behave against a human players actions. If there is any behavior change at all that is. And skill level is not just a multplier to Rng result. Only need to figure out how to ad a player.
 

 

10 tests looking at what happens to a single aircraft has a lot more potential for random variation.

Yesyes, in a nice way your telling me if I alone make 10 inverted passes, gear down, on low aa without getting whacked. It still means didley squat    as 50 out of a hundred could fail.

 

 

I also appreciate the discussion and find the real world statistics an eye-opener. Did read Retneks post and while these collected statisctics are both surprising and impressive in themselves, my knowledge here is far to limited to know what to make of them. Are these actually fired number of rounds while aiming or trying to at a target?

Edited January 16, 2018 by a_radek

[unreasonable]

@unreasonable. I trust you on the results and considering how many rounds you had to do to get a reasonably stabilized average result for every test, understand you have put much time and effort in these scenarios.

 

It's not to question your testing procedure I asked for the mission but figured it would help me watch in what situations 'low' gunners make a successful hit and how they behave against a human players actions. If there is any behavior change at all that is. And skill level is not just a multplier to Rng result. Only need to figure out how to ad a player.

 

 

Yesyes, in a nice way your telling me if I alone make 10 inverted passes, gear down, on low aa without getting whacked. It still means didley squat    as 50 out of a hundred could fail.

 

 

I also appreciate the discussion and find the real world statistics an eye-opener. Did read Retneks post and while these collected statisctics are both surprising and impressive in themselves, my knowledge here is far to limited to know what to make of them. Are these actually fired number of rounds while aiming or trying to at a target?

 

Did you manage to get the mission to run? Watching from the pov of the first gun to come up is pretty good as it is in the middle of the side of the field and a number of planes fly almost directly over it.

 

 If you make 10 inverted passes with gear down and not get hit - - it  would certainly mean something, since it would massively change the estimated average result.  I am fairly sure that the number of losses per round approximates to a binomial distribution, so once you have a large enough sample to get a decent average, you can calculate the probability of getting one loss, five losses, or forty-six losses in any given test run.  As you add more tests of a specific scenario you can progressively adjust your average, but each incremental test makes less and less difference.

 

On the last point, I assume you are asking about the rounds fired per kill data? I am confident that these are high level aggregate: some researcher knows the total number of rounds expended - or a proxy, like rounds delivered through the logistics chain - and the number of claims or confirmed kills by the LAA.  

There are some ammo expenditure figures in the Bodenplatte book relating to the specific attack, and IIRC put into a ammo/kill format they would be much lower than the aggregates in Rednek's post, which I am sure would include, for instance, barrage fire put up against medium/low level bombers. 

 

I will try to summarize the book's data when I get some time: maybe start a new thread for that.

Edited January 17, 2018 by unreasonable

[A_radek]

Did you manage to get the mission to run? Watching from the pov of the first gun to come up is pretty good as it is in the middle of the side of the field and a number of planes fly almost directly over it.

 

I have the archive but also have a very limited evening play time these days. Hoping to get a few hours later this week and will be sure to view from gun 1 first.

 

 

If there are ammmo expenditure numbers for a specific attack. Would that not give us an idea on number of cannons or rather how many on average were firing simultaneously during this one attack?

 

Regarding Retneks numbers, I see. I suspected those shell numbers would be something like that and probably same proxy or source used for all the different sums mentioned. As the relationships between them would answers questions like what type was more effective.

Edited January 17, 2018 by a_radek

[unreasonable]

There are a few figures in the text and tables but they are not systematic. Certainly, even including an allowance for the usual overclaiming or multiple claims for the same hit, the numbers are much lower than for the aggregates previously posted. They are also lower than my test run - although that does not surprise me, as Livai says, the AI gets target fixated and keeps constantly firing and traversing at a retreating target until it is out of range, which I find implausible for this type of low strafing attack. Certainly also not all guns got manned in time, or at all, some were suppressed (although I think none destroyed), so although I am fairly sure that most airfields were defended by 24 guns, the actual number that fired would certainly have been considerably lower. 

 

In brief, it is a mess, making comparison difficult. Nevertheless, this is about the best data we are likely to get about this sort of attack, so I will persevere  in my attempt to make sense of it.

[A_radek]

In brief, it is a mess, making comparison difficult. Nevertheless, this is about the best data we are likely to get about this sort of attack, so I will persevere  in my attempt to make sense of it.

Yes lot's of factors not in consideration and I also see them pointing to both ways. More examples:

 

Small arms fire certainly played a role in this and perhaps even accounted for some of the Aa claims. A machinegun or several at close range is hardly harmless.

 

Terrain and gun placement is a large factor, we don't really know the exact flight path through it. Guessing it would exclude some of the guns. Then again like I think I previously mentioned gunners should be taking shifts day and night to have every gun constantly manned. So even with loaders drunk on new years day gunners could still get a few rounds in the air, or even towards any possible non flying attacker. 40 mm rounds against unarmored targets is certainly effective.

 

In the end your low Aa performs better than the probably over-claimed hits and no we can't ever in detail simulate a specific attack with this info. I agree your method for determining appropriate Ai level in this case is as valid as we can make it.

Edited January 17, 2018 by a_radek