Mitsubishi Stuka and Thunderbolt

[Treetop64]

Originally Posted by majorfailure

The engine on the P-39 is not tough at all - if you get shot from behind and slightly off-angle.

Historically, this is one of the things about the P-39 that made it particularly vulnerable to fighters attacking from behind. Not only the engine being where it is made it vulnerable to fighters, but the Prestone overflow tank and the oil tank were positioned directly behind the engine, making them extremely susceptible in air-to-air combat.

[Pursuivant]

Originally Posted by majorfailure

Though VS. B-24 or B-17 this does not work. But using high or beam or head on attacks with good speed one nearly does not get hit. I have seen a flight of AI B-17s chopped up by AI Bf109G6s from behind with no losses once or twice, but most of the time the AI Bf109s lose one or two. And they more or less park behind the B-17s.

Yep. While I think that there's more to be done with gunner realism, things got a lot better in 4.11. My typical test is a QMB mission with 4 ace attackers going after a flight of 16 Average bombers. In those cases, it usually works out to 50% losses on both sides when late war German types go up against U.S. heavies. And that's with really dumb tactics on both sides. (i.e., Fighters never make high side or head on attacks. Bombers are always in loose echelon right formation with four planes per flight, and lose formation as they turn.)

[horseback]

Originally Posted by FC99

They can be "fixed" but the problem is that you think that AI gunners are insanely accurate while I think that they suck big time.

Actually, what I said was that they suck big time because they are so accurate, so you're half right. While I agree that they are greatly improved over 4.10, they have a long way to go before the proportional advantages of fighters vs bombers approaches the historical standard.

Test setup:
Planes : Ju88 and P47D
Distance: 200m
Test method:
Both planes are on the airfield, P47 engine running.
Player is in Ju88 rear gunner position.
P47 is behind Ju88 with front of the engine exposed to the gunner like in typical 6 o'clock attack.
Result:
Bullets Fired: 1200
Bullets Hit Air: 1047
P47 engine still running although at 90% and with some components damaged.
And as many times before FACTS>>>FEELINGS , P47 is one tough MOFO and for every FG guy's story about one ping kill there is a JG guy with the story about P47 soaking dozens of 30mm hits and flying away.

Sorry, but you'll never convince me that an ai plane has exactly the same damage model as one piloted by the Player, or that a human gunner can be as accurate as an ai one. We're talking about a routine that permits high deflection hits at over 700 meters and shot-out engines at steep angles changing at high speeds with considerably less than 1047 rounds fired.

Fly formation 200m behind an He 111 or Betty (both of whose gunners are traditionally more accurate than those noobs in the Ju-88A) in that same P-47 and I bet your engine loss ratio goes up significantly, along your PKs, loss of gunsight, ailerons, fuel leaks(and how could any rounds possibly get past the engine and firewall to reach the fuel tanks?), rudder and Prop Pitch. Of course, that's just my feeling, but it's based on several hours of experience.

AI vs AI contests may ultimately obtain 'realistic' results, but in those cases, the AI fighter knows that he's been fired at and exactly where it will hit if his vector remains constant at the moment it is fired and he makes the slight move that either results in a clean miss or a meaningless hit, but the ai gunner routine knows that he knows and quickly fires a burst at the corrected vector, but the fighter ai routine knows that he will, so they decide not to do that and move on to the next move/countermove several thousand times per second.

Think of the Dread Pirate Roberts' confrontation with the Sicilian 'with death on the line' in The Princess Bride.

cheers

horseback

[KG26_Alpha]

Test AI gunners

1. Offline QMB
2. Offline Campaign
3. Online Dogfight server
4. Online Coop

I know the difference, I wonder if you will notice it too.

[horseback]

Originally Posted by Pursuivant

I'm inclined to agree with you, but I don't have the data to prove what you're saying, other than somewhat generic stats on bomber casualties due to flak and fighters. We can trade anecdotes all day, but I'd love to see actual statistics to help back us up.

If anything, it seems like bomber gunners (at least rookie to average gunners) have been "nerfed," in 4.12, if only by unrealistic bomber formations and doctrine. But that's only by comparison with the laser-like precision with which gunners prior to 4.12 could shoot you down. It was if were were a generic Imperial TIE fighter pilot and the gunners were Han Solo and Luke Skywalker!

But, gunners must have had some usefulness, otherwise bombers would have dispensed with them earlier.

It is often a mistake to attribute logic to the human decision-making process. The gunners had more of a morale effect than a practical one; of course, both of the USAAF’s primary heavy bombers were designed at a time when bombers had a big technological and performance lead over fighters, and of course, no one had any idea that radar would become as effective as it did (even those who understood that radar was even possible). When the first B-17 entered the Army Air Corps, the P-26 was America’s primary fighter defense (and the P-35 & P-36 were just undergoing trials), and the Hurricane, Spitfire and Bf 109B were just entering service; the Japanese air arms were still fielding biplanes, as was the US Navy.

The basic idea was that by entering enemy airspace at high altitudes and comparatively high speeds, it would be impossible to locate them, much less intercept them with any meaningful force before their bombs were dropped and they were on their way back to base. The fighters of the period were barely capable of the performance of the big bombers at high altitudes; at those comparative speeds, the bombers’ gunners might have been almost as effective as their predecessors in the First World War, especially if only a few at a time could make an interception.

Without the invention of radar, they would have been right. Since the USAAF was run by the ‘bomber barons’, men like Claire Chennault, who not only flew fighters, but openly tried to develop a meaningful doctrine for fighter defense, were run out of the service by hook or by crook. That conviction held on long after the early rounds of World War II made it clear that it was obsolete. There were careers to be made and profits to be taken by the big aircraft companies like Boeing and Convair (who had just coincidentally, been displaced by more innovative companies in the fighter business). You could make a lot more profit (and provide many more jobs) with a single bomber than you could with three or four fighters.

Hubert Zemke’s memoir written with Roger Freeman mentions a conference in London in the early summer of 1943, where an unnamed senior bomber officer told Zemke that they didn’t need a fighter escort. I hope that SOB lived long enough to regret those words, but by mid-1943, the US had committed itself to those bombers and the idea that they needed all those guns and men to serve them, and that was the way it went for the next two years or so.

Senior military officers are politicians as much as they were ever anything else and it would not be politically wise to announce that they had been in error, so the obvious thing to do was to continue bloody-mindedly, and act as though they had simply underestimated how many big bombers would it take to bull their way into German airspace until the necessary long range fighters finally became available and then they pretended that that was what they had planned all along.

I think that damage or kills due to heavy flak is about right - as long as you take into account the fact that each gun in the game can actually represents an entire battery. Low to medium altitude flak is downright lethal, but that might actually be realistic. Veteran ground attack pilots learned to come in fast and low, make one good pass and get the hell out.

I don't have a problem with gunners starting to shoot at 500 meters range, but that should mostly be "suppression fire" with very little chance of actually hitting. Shots at anything other than minimal deflection angles against a plane flying a relatively straight course should also have almost no chance of hitting. But, if you make an attack from 6 o'clock level against a heavy bomber, without approaching at a very high closing speed, you deserve every bullet that hits your plane.

Turning speeds for turrets seem to be about right. At least for the U.S. turrets, there's pretty good performance data, and actually possibly a few turrets that actually still work. In archival film, you can see that they turn pretty quickly - something like 120 degrees per second.

But, against that, something that isn't modeled in the game, at least for human gunners, are the effects of G forces, vibrations from the plane itself and wind buffeting of the turret and guns if the guns are angled into the plane's slipstream. All those things make bomber gunnery a bit too easy, at least for a human gunner. I don't know if the AI models those things, but it should.

Agree with you here; however, in the case of flak/AAA, I would note that even the heavy guns seem to be able to shoot at you at relatively close range and with remarkable anticipation. I would like to see some sort of hesitation or delay built in when the first low-level attacks are made, because like a seasoned prostitute, they seem to know when you’re coming before you do.

As I recall, the turrets were generally spun by means of a foot or knee switch and they weren’t all that precise (but the speed was impressive, which would count for a lot in a newsreel). In any case, they were still dependent upon a human being’s estimations of angle and range. Their greatest contribution to accuracy was that the gunner was strapped in and could continue to shoot a multiple gun battery under maneuvering conditions, while the guy sitting on a bench or standing was just hanging on for dear life.

There is no reduction in accuracy for maneuvering aircraft that I have noticed. My tests against the Betties got me hit just as often (and from as far away) while the aircraft was in a steep turn as when they were flying straight and level (which was rarely the case if my cross-hairs were pointed in their immediate vicinity). As I’ve pointed out, the Bf 110 gunners (among others) are beyond ridiculous; put yourself in the in-game position yourself and you’ll see why. They have a tiny, tiny cone of fire, they are seated on a glorified (backless) lawn chair, and when the aircraft is flung about the way the AI pilots of these things routinely do when under attack, they will hit your engine or nosecone (and therefore your prop pitch control) with great regularity from angles that would seem impossible for a human gunner to achieve even with the silly-assed no-recoil mouse gunnery model.

I once suggested on the Ubi forums that anyone who thought that having an element of motion wouldn't affect their aim should try playing out a Sturmovik mission as a gunner on a laptop while riding in a moving car; no one took me up on it (or if they did, they were too embarrassed to share the results).

cheers

horseback

[horseback]

Originally Posted by KG26_Alpha

Test AI gunners

1. Offline QMB
2. Offline Campaign
3. Online Dogfight server
4. Online Coop

I know the difference, I wonder if you will notice it too.

As I have repeatedly pointed out, my interest is strictly off-line, and I have asked for a 'reality option' specifically for offline campaigns and missions. I would even put up with 'realistic gunnery' limitations for the fighters in exchange for a realistic change to the ai gunnery model, as long as the proportional advantages enjoyed by the fighter were put in place.

I understand that the QMB ai are a bit tougher than the 'campaign' ai, but not by orders of magnitude; a breast fed baby's diaper doesn't stink as badly as one that is fed formula, but it still stinks.

As for online, not my area of interest.

cheers

horseback

[KG26_Alpha]

Its of no interest to you online, but is for many of us that do fly that way the AI gunners are totally different.

All you need to do is change the AI bomber flights to rookie and see how you get on off line or change your attack method with normal settings.

If your complaint is for a selective part of the game to be changed you need to understand whats going on in the rest of it before requesting changes and making generalized statements.

As I already said the game has different ways to fly it, offline you can do what you like on your own, online you work differently, I just hosted a CooP mission on Hyperlobby attacking 32 Betty's with Corsairs
The only Ai gunner kills were pilots getting greedy and getting 100m off the six of the bombers, all Ai Betty's were shot down easily.
The loss of pilots was mainly due to debris coming off the bombers (wings rudders ailerons) hitting the pilots Corsairs because they were too close on the attack.
3 pilots were de-winged at the same range from the 20mm tail gunner again much to close sitting on the bombers six.
On TS we were all saying the same thing "this guys gonna die" because of the wrong attack method.

As this thread is now off topic, please see fit to start a new one relating to this particular discussion.

[JtD]

Originally Posted by JtD

The part with "two".

Picture available now.

[MiloMorai]

And another,

[FC99]

Originally Posted by Jumoschwanz

If the best simulation IL2 can do right now with the P47 is to have it's tail fall off with X number of 20mm hits, then that is okay, it is just funny how you can set yourself up in the qmb behind a bunch of Friendly p47s and go down the row and get the same result on one after the other, almost like you are breaking the tails off of frozen lobsters.

Of course in real life the location and number of hits from an enemy aircraft would be so random it would take a supercomputer to model the different effects of each, I am sure no two kills in WWII looked exactly the same the way they can be reproduced in IL2.

Tail coming off is relict of the past when average player's PC were much weaker than what we have now and it is best viewed as crude visual representation of catastrophic damage. Paradoxically you will experience such damage most often on very tough planes, weaker ones will usually suffer some other types of damage sooner.

Originally Posted by horseback

Sorry, but you'll never convince me that an ai plane has exactly the same damage model as one piloted by the Player, or that a human gunner can be as accurate as an ai one.

It does but I can do the same in coop mode with human controlled plane. Fact is that even 1000+ hits in the engine area is not a guarantee that you will destroy P-47 engine. In addition I can check exact damage of each bullet and if I really want to be anal about this I can generate hits at any part of the plane with whatever parameters I want.

You are very wrong about human gunners too, they are way better than AI, obviously you don't fly online much.

Fly formation 200m behind an He 111 or Betty (both of whose gunners are traditionally more accurate than those noobs in the Ju-88A) in that same P-47 and I bet your engine loss ratio goes up significantly, along your PKs, loss of gunsight, ailerons, fuel leaks(and how could any rounds possibly get past the engine and firewall to reach the fuel tanks?), rudder and Prop Pitch. Of course, that's just my feeling, but it's based on several hours of experience.

Is it surprising that cannon will do more damage than MG bullets?

AI vs AI contests may ultimately obtain 'realistic' results, but in those cases, the AI fighter knows that he's been fired at and exactly where it will hit if his vector remains constant at the moment it is fired and he makes the slight move that either results in a clean miss or a meaningless hit, but the ai gunner routine knows that he knows and quickly fires a burst at the corrected vector, but the fighter ai routine knows that he will, so they decide not to do that and move on to the next move/countermove several thousand times per second.

You are over-thinking it a bit and you are very wrong, not 100% wrong but close.