The bravery of being out of range

[6./ZG26_Custard]

I just thought I'd open a thread to discuss Triple-A. ​I know this has been mentioned before but flying as a ground pounder 99.9% of the time it is somewhat noticeable that at times, the AAA can be almost "godlike" in its accuracy.

I have raised this purely as a discussion point and not as a flame thread. I remember way back 6./ZG26_Emil did some research on the Flak and AAA during WWII.

 

If memory serves it was either the 20 or 37 mm guns that reported that roughly 900 rounds of ammunition were expended to score on average, just 1 hit.

 

It’s certainly appears that our triple-A in game is far more accurate than that.

 

 

 

Edit: Any further info on RL triple-A during WWII or your thoughts on this subject are always welcome. 

Edited November 28, 2017 by 6./ZG26_Custard

[[CPT]milopugdog]

I do feel that the AAA in this game can be a little hit and miss. Either you get taken out, or you don't get hit once.

Obviously this has to do with how you are flying, as well as the kind of AAA there is at a target. I've also found that the skill of AAA differs from server to server.

 

For example. When I fly on WoL, I can fly to a target in a lone Hs-129, and make it through the AAA. But when I do the same thing on a server like TAW or DED RE, I don't even get to make it to the target area.

So I think that the accuracy of AAA gunners is defiantly case by case. It can even be from unit to unit. It's all up to the map maker to decide how they behave. 

 

However, much like with gunners, the difference of beginner and expert AAA is a little absurd. I've been at 2000m in a Bf-110, only to be taken out by the first round of a pom-pom AAA.

[=EXPEND=Tripwire]

Problem I think is there is specifically an AI setting to allow a mission designer to set up a 'no fly zone' for vulching protection if that's what they want without bogging down the server with lots of guns.

 

The problem lies with that same setting being used where it shouldn't. Targets etc.

 

Overuse of 'ACE' AI flak and aircraft gunners by mission designers is probably the issue here.

Edited November 29, 2017 by =TBAS=Tripwire

[6./ZG26_Gielow]

I believe on TAW server some targets are designed to be attacked only by level bombers. If you get close flying low level you will be hit for sure by ace gunners placed there to discourage people flying fighter bombers on some targets.

[=SqSq=Sulaco]

AAA can be juked and destroyed on TAW as long as you or a wingman drag it, going solo against it will more often than not end very poorly for you.

 

I agree that perhaps form a historical standpoint it's accuracy may be too good but from a gameplay perspective I prefer the deadly accurate AAA on a server like TAW as without it the ground targets would disappear from the map far too quickly.

Edited November 29, 2017 by =SqSq=Sulaco

[6./ZG26_Gielow]

Just can't wait for flak trucks controlled by players.

 

 

Closed configuration.

 

Open and ready to fire.

 

I vote for 4 barrels flak

[KoN_]

I think the AAA has always been a debait . I've only just start back flying this game and I see no change in AAA still dead on target . But must draw its fire with genie pig.

[Mitthrawnuruodo]

I think Tank Crew might give us better AAA as a side effect. Combat on the ground will require that the behaviour of AI gunners be more realistic.

[CUJO_1970]

The AAA in the sim behaves exactly like what it actually is - computer controlled aim bots.

 

The challenge for programmers is to get it to behave otherwise.

[Dakpilot]

The 900 rounds to hit an aircraft comes from a wiki article about Russian 37mm AA gun.

What it actually said was the statistic of total rounds expended during WWII compared to actual aircraft shot down, and a figure of 903 was the result in respect to Russian 37mm.

Now this is an interesting average figure, 903 rounds expended for 1 aircraft destroyed over the war period , but impossible to apply to any individual 'tactical' situation.

 

There is nothing wrong with using the actual statistic as a starter for a debate about actual AA effectiveness but in itself it tells us nothing

 

Cheers Dakpilot

[unreasonable]

The AAA in the sim behaves exactly like what it actually is - computer controlled aim bots.

 

The challenge for programmers is to get it to behave otherwise.

 

It is remarkably easy to get roughly realistic outcomes we are just talking about adding a randomized error around a predicted point, there is no need to explicitly model all the physical and psychological variables. 

 

In the RoF system there are a variety of variables, such as the speed of traverse, which can make a difference. The main variables for heavy flak, however, are the predicted point of explosion given the target height, speed and course, and a variable that sets a sphere around the predicted point in which the actual explosion takes place randomly.  Then there are the variables for lethality when a target falls inside the danger radius of the explosion.  I am not sure how to mod RoF AA to make the target prediction less accurate, but by increasing the size of that sphere one can make the shell bursts much more random. The result is that RoF AA can be made to be somewhat more realistic than vanilla.  There are various mods that do that - I use my own version.  

 

There are still problems with the lack of a minimum range feature, but that again would be trivially easy to program. Similarly the difference between barrage fire and predicted fire should not be too hard to model for such a talented team.

 

The fundamental problem is that the developers have never, to the best of my knowledge, attempted to model realistic AA. What we have is a MP driven "vulch" preventer. Some SP players also contributed to the realism deficit when they complain that flak is "ineffective" because their flight only suffered 50% casualties.

If the team actually want to model realistic AA I have no doubt they could do it with little difficulty.  Meanwhile the release of modding (???) would allow SP to experiment.

[A_radek]

I find these AA effectiveness discussions tend to be mostly one sided, as in not enough emphasis on pilot (that's us) tactics.

While blazing down on a tank convoy at a blistering 400kmh, raining hell, death and destruction on those tiny helpless targets it might feel like your invinsible. Until a wing pops and your thinking, no way, they were still so far away. Watching that same attack from the perspective of the ground troops it looks pathetic. A lone il2/ju87 at 200 meter height growing slowly on the horizon, diving down to what feels like handgun range and then slowly soars by. Much like a fly by on an airshow.

I served as an AA gunner manning a 40mm bofors (yes same old that's been used forever but with modern aiming electronics, today mounted in a light tank "stridsfordon90"), and even with an iron sight and two crank wheels for horizontal/vertical aiming an AA gunner with any kind of training would have no problem taking out an aircraft the way most of us attack groundtargets online. There's a reason an il2 irl didn't take out 12 tanks in a sortie. Also, this is all excluding the massive small arms fire we would have to endure flying that close to a target. 

Regarding the 903 rounds per aircraft destroyed. That's roughly my hit percentage online And if my life depended on it I'd hose around even more.

Edited November 29, 2017 by a_radek

[Birdman]

Complete outsider question but if we consider a 4 37mm guns set up to defend an objective, they would not have to fire that long on average to get 903 rounds in the air ? So statistically it wouldn't take that long to get that fateful 903rd round?

[6./ZG26_Custard]

Complete outsider question but if we consider a 4 37mm guns set up to defend an objective, they would not have to fire that long on average to get 903 rounds in the air ? So statistically it wouldn't take that long to get that fateful 903rd round?

That would be around 225 rounds to score a hit based on 4 firing simultaneously at the same target?

[xvii-Dietrich]

I find these AA effectiveness discussions tend to be mostly one sided, as in not enough emphasis on pilot (that's us) tactics.

 

While blazing down on a tank convoy at a blistering 400kmh, raining hell, death and destruction on those tiny helpless targets it might feel like your invinsible. Until a wing pops and your thinking, no way, they were still so far away. Watching that same attack from the perspective of the ground troops it looks pathetic. A lone il2/ju87 at 200 meter height growing slowly on the horizon, diving down to what feels like handgun range and then slowly soars by. Much like a fly by on an airshow.

 

I served as an AA gunner manning a 40mm bofors (yes same old that's been used forever but with modern aiming electronics, today mounted in a light tank "stridsfordon90"), and even with an iron sight and two crank wheels for horizontal/vertical aiming an AA gunner with any kind of training would have no problem taking out an aircraft the way most of us attack groundtargets online. There's a reason an il2 irl didn't take out 12 tanks in a sortie. Also, this is all excluding the massive small arms fire we would have to endure flying that close to a target. 

 

Regarding the 903 rounds per aircraft destroyed. That's roughly my hit percentage online And if my life depended on it I'd hose around even more.

 

 

Well said!

 

 

Regarding claims of insane Flak accuracy, let's see some videos of recorded tracks, but with the  point of view from the ground unit.

 

Also, regarding these historical claims, are their some references on this, that we can use for reading up on the figures and, more importantly, the context.

 

Another thing I often read is "I got shot by AA", rather than "My entire squad who was flying in to saturate the defences was shot". I have not found any historical examples of lone-aircraft attacking facilities. Either they were in groups, it didn't happen or, perhaps it did, but they didn't come back.

 

Indeed, it will be interesting to see if the planned GAZ-MM + 72K and Sd. Kfz. 10 + Flak38 make a different to the battle experience, when Tank Crew adds human-controlled flak to the game engine.

[unreasonable]

Try reading "Flak - German AA defenses 1914-1945" E.B.Westermann UP Kansas.  It is the most comprehensive English discussion I have come across - while specifically on the German Flak forces, the lessons are general. It is mostly about the heavy defense of fixed targets.

 

There needs to be a completely clear distinction between the heavy flak firing a time (or altitude) fused shell and the light/medium guns which needed to get direct hits.  A level bomber formation taking heavy AA and fighter bombers attacked by medium/light AA while ground attacking are completely different cases. 

 

I actually have a track of a fighter bouncing up and down near the ground flying around a gun while the AA (37mm) tracks it's altitude and speed exceptionally well - used in making my last film - and this was with the AA set on normal in the ME.  If anyone is really interested I can make a film of it, but I am not sure that there is any point unless there is some indication that this whole topic is an area the developers intend to give some thought.

[Stig]

Something no-one seems to bring up when AI AA accuracy is discussed is the conversely extreme accuracy of the human players! IRL targets were rarely taken out by 1 or 2 bombers; even a formation of 10, 20 or more bombers may not have caused much damage, especially when bombing from high alt; they may not even have hit their intended target! 

 

 

[6./ZG26_Custard]

 

 

However, much like with gunners, the difference of beginner and expert AAA is a little absurd. I've been at 2000m in a Bf-110, only to be taken out by the first round of a pom-pom AAA.

  Agreed, the difference between the settings seems to be a fairly significant factor. 

[Birdman]

That would be around 225 rounds to score a hit based on 4 firing simultaneously at the same target?

Wikipedia says the practical rate of fire was about 80 rpm. That's about 3 mins of fire to get a plane down in that set up.

 

If the game uses the spec rof (160), you possibly have 90 secs above target before the statistic catches you.

 

Disclaimer I ain't an expert it's all guesstimating.

[PatrickAWlson]

In PWCG missions all of the AAA AI is set to Novice.  I think this works really well.  Sometimes they miss.  I often get plinked with a  couple of rounds without catastrophic damage.  Sometimes I really do get hit hard, but not usually.  If you set AAA to Ace AI level then you are getting what you asked for.  As suggested by some a better range of AAA AI - generally decreasing accuracy - would probably be a good thing.

 

IRL AAA around airfields was deadly.  Multiple passes on a field were generally considered to be suicidal.

[InProgress]

The AAA in the sim behaves exactly like what it actually is - computer controlled aim bots.

 

The challenge for programmers is to get it to behave otherwise.

I don't know, it works great in il2 cliffs of dover. So cant be that hard to make.

[unreasonable]

In PWCG missions all of the AAA AI is set to Novice.  I think this works really well.  Sometimes they miss.  I often get plinked with a  couple of rounds without catastrophic damage.  Sometimes I really do get hit hard, but not usually.  If you set AAA to Ace AI level then you are getting what you asked for.  As suggested by some a better range of AAA AI - generally decreasing accuracy - would probably be a good thing.

 

IRL AAA around airfields was deadly.  Multiple passes on a field were generally considered to be suicidal.

 

Just been reading a magazine article in "Aeroplane" about Bodenplatte. JG3 was instructed to strafe an airfield: each plane was to make three attacks, circling after each run, before returning home. Presumably the planners thought that is how many attack runs were needed to use up all the ammo. 

 

The target airfield was defended by 4 RAF Regiment AA units with 40mm Bofors ( so 12-16 guns? not sure of OB).  The attack was recorded as lasting for 25 minutes.  JG3 lost 15 out of 60 aircraft, ie 25 % all except one to ground fire.  My guess would be perhaps none on the first run, then 10% then 15% as the defenses got their act together.  

 

So not suicidal - just very dangerous.  Suicidal would be to recreate this mission in the ME and then play the attacker.  ;I   The problem is that people in game have a poor estimation of the meaning of loss rates. 25% losses in one sortie were regarded as a "catastrophe" by the GAF: in MP, they would be regarded as insufficiently immersive.

Edited November 30, 2017 by unreasonable

[RedeyeStorm]

Hoi Unreasable. You are talking about sixty planes attacking one airfield. Attacking one airfield with one or two planes is a totally different thing.

[19//Moach]

Truth be told, I am continuously befuddled by the common sight of even experienced pilots attacking targets while recklessly ignoring live AAA firing at them.

 

I find myself often advising them "Take out the guns FIRST, then worry about anything not trying to kill you" but often to no avail.   They get a pair of bombs down, then get mangled by the 3 or 4 guns surrounding them in every direction, blasting tonnes of lead into the air...

 

Then complaints erupt about "sniper AAA", neglecting the half million rounds that preceded the one that hit them, ripping their planes apart like a torn rag.

 

 

Whenever I fly fighter cover, I find that I am pressed to dive in and remove the "quick-shooters" before the whole attack force gets wiped out.   Even in groups, this rather basic mistake continues to be made, with many a formation of IL2s gunning down tanks in blatant oversight of the AAA trucks which take a few minutes to start engaging, but inevitably do so unless killed right away.

 

 

 

Whenever I attack a target, I have a simple procedure which I carry out immediately upon arriving there:   Kill the damn AAA.   

 

It works, very well too.  It becomes easier and easier with every gun removed until the target is yours and all there is to worry about are enemy fighters, which hopefully are being handled by the cover one did not forget to ask (beg) for multiple times while enroute.

 

 

Other times, I see pilots making the mistake of assuming an unmanned gun is the same as dead. It is not.  A suppressed crew might be away from their posts for a few moments after being shot at close enough, but make no mistake. AAA is always a danger until a big satisfying explosion announces their "retirement".   Never assume a gun is dead just because it isn't firing.  It can start doing so at any moment, so take the opportunity when it's silent and finish it good and proper.

 

AAA positions are easily distinguished between dead and dangerous:  Dead ones are blackened charred smudges, or entirely empty ring barricades.  Anything else is a threat, and should be considered top priority when approaching a target.

 

 

You should find enemy air defenses much less "sniperey" and more cooperative after they've been swiftly dispatched to hell.

[unreasonable]

Hoi Unreasable. You are talking about sixty planes attacking one airfield. Attacking one airfield with one or two planes is a totally different thing.

 

I agree that is relevant: JG3 attacked from one direction, in sequence, circling to make multiple attacks so they were by no means all attacking at once.  Probably in waves of 20 or so aircraft.  AA guns can only fire at one target at once, I agree, and I do not know how many there were - the article just said 4 RAF Regiment units, so I am guessing 12-16 guns but that might be wrong.  

 

So if 12-16 guns shot down 14 aircraft in three passes - or say even two if the first attack was a complete surprise: then there is say a 33-50% chance of an AA gun getting hits on a target flying more or less towards it. I have no problem with that, actually: it is the ability of the guns to track targets moving fast at a tangent while changing altitude that I find most suspect.  Another way to look at it is that 12-16 guns (?) took 25 minutes to shoot down 14 aircraft in a target rich environment. 

 

But as I said earlier the true test would be to try to recreate a large attack on an airfield in the ME - instructing the AI to fire and then fly away so that it does not loiter about, and see how many the AI can shoot down at each level. My guess would be that the AI would do significantly better than the RAF Regiment.... to start with the AI would not be taken by surprise - in RL if you made just one pass at an airfield you had a good chance of getting well over it before the AA opens up at all.

 

Might try to set that up in the ME if I can get my head around it again: better if one of the mission wizards could do it.

[6./ZG26_Custard]

 

 

I find myself often advising them "Take out the guns FIRST, then worry about anything not trying to kill you" but often to no avail.   They get a pair of bombs down, then get mangled by the 3 or 4 guns surrounding them in every direction, blasting tonnes of lead into the air...   Then complaints erupt about "sniper AAA", neglecting the half million rounds that preceded the one that hit them, ripping their planes apart like a torn

 

Our squad does indeed try to knock out AAA in the first instance (most of the time) and we sometimes split our team to target AAA and other ground units. I think as its been mentioned by others, the "ace" setting for AAA  is probably just a bit too ace, maybe? The gunners will happily track, target and kill any aircraft coming their way even with ordnance falling all around them. 

 

 

 

Might try to set that up in the ME if I can get my head around it again: better if one of the mission wizards could do it.

 

I'd be interested to see that.

 

 

Sorry for the terrible audio to this video but it just shows some German AA at work, its interesting to note that by mid war the Germans had more or less phased out  the 20 mm stuff. 

 

 

[Trooper117]

Many, many RAF pilots were lost on airfield attacks due to a storm of light to medium flak that was put up by its defenders.

Many pilots dreaded being given that type of mission purely because of flak lethality...

How accurate that translates to what we do in game however I have no idea...

[6./ZG26_Custard]

Many, many RAF pilots were lost on airfield attacks due to a storm of light to medium flak that was put up by its defenders.

Many pilots dreaded being given that type of mission purely because of flak lethality...

How accurate that translates to what we do in game however I have no idea...

I'm sure It was no doubt terrifying (and deadly) to attack a heavily defended airfield or fortified position. The thing I notice most in game is just how accurate say a single or a couple of K-61's are, firing approximately 60 rounds a minute. When you look at real world data giving us 1 hit for every 903 rounds fired, they seem remarkable accurate in game.

[Ghost666]

Ok i'm no expert, but wouldn't the flack even light guns be deployed in batteries?  So there would be many more then just 1 or 2 guns protecting a target. And I believe that deploying 15 or 20 guns around an airfield may be quite taxing on the game. Then deploying 1 or 2 guns that are as effective as a battery or 2 may be the most realistic thing we can do.  

 

Just thinking out load.   

[7.GShAP/Silas]

With the AI skill turned all the way up they're like flying amongst a nest of SAMs.  A couple days ago a comrade and I were on TAW, flying fast and about 15 meters off the ground in fighters.  We were passing 3Km(!) away from an enemy defensive position, behind trees and one of those K-61 guns hit him once and reduced him to a cockpit sailing through the air.  

 

Like Tripwire said, overuse of ace AA gunners is the issue. 

Edited December 3, 2017 by JV.44/Silas

[unreasonable]

Ok i'm no expert, but wouldn't the flack even light guns be deployed in batteries?  So there would be many more then just 1 or 2 guns protecting a target. And I believe that deploying 15 or 20 guns around an airfield may be quite taxing on the game. Then deploying 1 or 2 guns that are as effective as a battery or 2 may be the most realistic thing we can do.  

 

Just thinking out load.   

 

This has been the historic justification, but I do not think it holds up any longer, now that PCs are more powerful and it is not a sustainable policy long term.  Take the proposed Pacific expansion, for instance: many of the ships mounted many rapid fire AA guns: the game has to be able to cope with that if that expansion every goes ahead. 

 

15-20 is probably on the high side historically anyway, although I would be interested to see any historic data. IIRC the German standard for advanced airfields was a LW ground unit with perhaps one battery of 88mm and one or two of light AA, so 8 or 12 guns maximum. 

[Eicio]

Like it has been said before the AAA seems to be really powerful but for good reasons, when you are flying a slow ground pounder, at low alt, in a straight line and alone you should not be surprised to have 50%-80% chances to be killed in a single run.

 

Thus AAA is not overpowered in those cases, and let's be honest you rarely fly in packs on mp servers. You just go on heading to your target and be aware that there is a high probability to be catched by lone ennemy planes or overwhelmed by any AAA position.

 

It's also true that IRL pilots were obviously much more careful than we are, if we're shot we just spawn in an other plane so we oftenly face AAA in some ways the real pilots never ever would have.

 

The AAA power is not really historicaly inacurate, our behaviour is and always will be, therefore make comparisons with the reality isn't quite useful in all the situations.

[unreasonable]

Well we shall see: BoBodenplatte after all will need to be able simulate that particular series of attacks, and we know quite a bit about what happened.   

 

 

My hypothesis is that if we replicate those attacks with the game AA on normal, the German losses will be much higher than the historic results. (Waiting for my just ordered book before I finish my test mission "Flakenplatte"    ) 

[7.GShAP/Silas]

Like it has been said before the AAA seems to be really powerful but for good reasons, when you are flying a slow ground pounder, at low alt, in a straight line and alone you should not be surprised to have 50%-80% chances to be killed in a single run.

 

Thus AAA is not overpowered in those cases, and let's be honest you rarely fly in packs on mp servers. You just go on heading to your target and be aware that there is a high probability to be catched by lone ennemy planes or overwhelmed by any AAA position.

 

It's also true that IRL pilots were obviously much more careful than we are, if we're shot we just spawn in an other plane so we oftenly face AAA in some ways the real pilots never ever would have.

 

The AAA power is not really historicaly inacurate, our behaviour is and always will be, therefore make comparisons with the reality isn't quite useful in all the situations.

 

 

The losses suffered and efficacy of anti-aircraft fire can often be, depending on the server and AA gunner AI settings, extreme even when flying in force(4-5 aircraft I'm speaking of here) and with well-drilled disciplined tactics.  If these ace AI settings are used, I believe the idea is not attempting to simulate real AA but using it as a tool for other ends(preserving limited ground targets, forcing high-altitude bombing, etc) .  Maybe that's the best choice one can make given the options available, but I really think that is the nature of it.

Edited December 4, 2017 by 7.GShAP/Silas

[Eicio]

The losses suffered and efficacy of anti-aircraft fire can often be, depending on the server and AA gunner AI settings, extreme even when flying in force(4-5 aircraft I'm speaking of here) and with well-drilled disciplined tactics.  If these ace AI settings are used, I believe the idea is not attempting to simulate real AA but using it as a tool for other ends(preserving limited ground targets, forcing high-altitude bombing, etc) .  Maybe that's the best choice one can make given the options available, but I really think that is the nature of it.

I dont't know if this is really a way to "force" high level bombing, back on the History: the battle of France. The french losses on close air support were really really severe, a massacre...

 

The AAA should not be underestimated, it may be a tool for missions makers but that'd still be immersive.

 

In 44' the allies had squadrons specially trained to take out AAA in zones where an operation was to take place, typhoons with rockets or hurricanes with rockets for the unlucky ones.

 

That is to say : there were missions before missions only to get rid of the AAA, that tells you how feared the AAA was.

 

And it was rightfully, I heard that in the Pacific something like 1\3 of japanese planes losses was due to oerlikons canons.

[Gambit21]

As with Patrick I set the AI to "low" in my missions - seems about right.

Even on "normal" they are beyond radar-guided accurate - let alone Ace.

[Blooddawn1942]

Just read Pierre Clostermans Book. This might give You an idea, how the AAA was feared.

Since I read 'The great Show' several years ago, I tread the AAA with more respect then ever before.

[Hauksbee]

I notice that quite a few people in this thread are listed as "Founder". Founder of what, exactly? IL-2, 777, this Forum?

[unreasonable]

I notice that quite a few people in this thread are listed as "Founder". Founder of what, exactly? IL-2, 777, this Forum?

 

IIRC it is those people who purchased the first game in the series in early access right at the start.  Given my flying, it should probably say "Flounder".

[6./ZG26_Custard]

 

 

As with Patrick I set the AI to "low" in my missions - seems about right. Even on "normal" they are beyond radar-guided accurate - let alone Ace.

Agreed, but as Gielow said above, it will be interesting to see what player controlled flak is going to be like?