4.12.2 de-bugging

[Woke Up Dead]

Somewhere in this thread (I can't find your exact quote now), you say that you suspect that engines are modeled as one big block. There are threads a year or two old here that deal with damage modeling, they show that damage models are more sophisticated than you believe them to be, there are inf fact several small components modeled in each engine. I think there are only two or three planes in game (P-39 being one of them) that have a simple, old damage model like you describe. So if your Buffalo's engine loses power after only 2-5 bullets, that means one of those bullets hit an important component.

You also say that the engine on a Buffalo seems to damage differently than the same one on a C-47. Are you taking shots to the engines from the same angle for both planes? I.e., are you chasing a Wellington or Val with the C-47 to take hits from the front?

"The bullets that are consistently killing the Wright Cyclone R-1820 engine on the Buffalo series are 0.30 caliber bullets being fired from anywhere from 50-300 meters distance, and they cause near instant engine-stoppage or serious power loss, regardless of where they hit. We're talking about bullets that make small holes and which might not have much power on them when they hit."

Don't rifle-caliber bullets have plenty of energy even after traveling 300m, never-mind just 50m?

"all bullets are treated as being incendiary"

Sorry, you're simply wrong here, check the fourth post of this thread: http://forum.1cpublishing.eu/showthr...hlight=belting


"Since damage modeling is an art, it seems to me that IL2's developers have made planes that were notably vulnerable in combat for any reason excessively vulnerable to any sort of damage."

No, just look at the Zero. Famously fragile, and in-game the early-war Zeros are particularly easy to set on fire, but have you ever had a damaged or stopped engine in a Zero? Fragile fuel tanks in wing-roots, but super-tough engines.

[Pursuivant]

Originally Posted by Woke Up Dead

Somewhere in this thread (I can't find your exact quote now), you say that you suspect that engines are modeled as one big block. There are threads a year or two old here that deal with damage modeling, they show that damage models are more sophisticated than you believe them to be,

Regardless of how various engine systems are modeled, I'm getting very consistent results. So, it's not some sort of random "critical hit," unless I'm really running afoul of the law of averages.

Originally Posted by Woke Up Dead

You also say that the engine on a Buffalo seems to damage differently than the same one on a C-47. Are you taking shots to the engines from the same angle for both planes? I.e., are you chasing a Wellington or Val with the C-47 to take hits from the front?

I was actually wrong about the C-47. It didn't use the R-1820 engine (although the DC-3 did).

Flyable planes in the game with versions of this engine are the Buffalo series, the SBD, and the CW-21. I will try to fly them all to see if it's a problem with how the engine is modeled, or how the engine is modeled in the Buffalo series.
The P-36 also used this engine but it's not flyable.

My procedure is to get up a QMB flight in arcade mode using Ace Wellington III (or TBD-1 for the slower planes), then deliberately use stupid tactics by overtaking them from 6 o'clock level without maneuvering much. It's a good way to get a nose full of lead and test engine and front armor DM. Arcade mode lets me see exactly where the bullets hit. I then compare odd results against a 3-view drawing to see if they actually make sense.

Originally Posted by Woke Up Dead

Don't rifle-caliber bullets have plenty of energy even after traveling 300m, never-mind just 50m?

Yes, they do, but smaller bullets lose energy at a proportionately faster rate than larger rounds, so a rifle caliber bullet hit at 300 meters isn't nearly as dangerous to machinery as a hit from 50 meters or less.

It's very realistic for a .30 bullet to punch into an aluminum engine block or go through 20 mm of armor glass at 50 meters or less. But, at 300 meters, armor glass should easily be able to defeat most .30 caliber rounds, and there's even a chance that a bullet might be stopped or deflected by an engine block.

Early war planes with armor plate were specifically armored to be protected against .30 caliber bullets fired at even close combat ranges, so at anything other than point-blank range, armor plate should stop them.


"all bullets are treated as being incendiary"

Originally Posted by Woke Up Dead

Sorry, you're simply wrong here, check the fourth post of this thread: http://forum.1cpublishing.eu/showthr...hlight=belting

Thanks much for this information, since it sort of proves my point.

First, you'll notice that there is no ball ammo in the mix for any of the guns listed. It isn't even modeled! That's highly unrealistic, since supply shortages or deliberate loadout choices might have meant that ball ammo was used.

Second, you'll notice a very high percentage of bullets that can start fires - HE, API, Tracer, APIT, Minengeschoss, etc. They're not all incendiary, but they might as well be! In some beltings, there's a 5/6 chance that a particular bullet is a potential fire starter!

Third, you'll notice that many beltings have a very high ratio of tracer bullets, sometimes as low as 1:3! 1:5 or even 1:10 was more typical.


"Since damage modeling is an art, it seems to me that IL2's developers have made planes that were notably vulnerable in combat for any reason excessively vulnerable to any sort of damage."

Originally Posted by Woke Up Dead

No, just look at the Zero. Famously fragile, and in-game the early-war Zeros are particularly easy to set on fire, but have you ever had a damaged or stopped engine in a Zero? Fragile fuel tanks in wing-roots, but super-tough engines.

My point here wasn't that all notably fragile planes have been made too fragile, but that some of them have been.

I'd also suggest that the A6M2 is another exception that proves the rule. It's very flammable - perhaps too flammable - and falls apart nicely if it's hit by a few cannon shells or a solid burst of 0.50/12.7 mm MG fire. Seemingly realistic.

But, since the A6M2 was a wonderful, well-liked airplane, and early war Sakae 21 engines were very good, arguably the designers went the other way and made the engine "too tough" (or "just right" depending on how you look at it).

After all, in terms of power, mass, compression ratio and power to mass ratio what makes the Sakae 21 so much better than contemporary radial engines like the R-1820 or Bristol Hercules?

[Pursuivant]

A couple of quick missions using the H.75-4, testing the P&W R-1830 Twin Wasp engine.

There isn't the same problem with massive loss of power like in the Buffalo, but the bad engine damage model here is that even a very few .30 caliber bullets in the engine at 100-200 meters are sufficient to trigger a massive fuel leak (i.e., one that triggers the RTB result when the autopilot is on).

Comparing these hits to a 3-view drawing of the H.75, I can see nothing which could realistically cause a massive fuel leak. After all, even if the bullets hit the fuel intake to a particular cylinder, or punch a hole in the cylinder itself, there's not much more loss of fuel than that cylinder would consume normally.

3-View here:

http://www.histaviation.com/Hawk_75_...g_1424x926.jpg

2 .30 caliber bullets in the engine from about 100-200 meters away. Notice giant fuel leak and "RTB" message from AI

http://forum.1cpublishing.eu/attachm...1&d=1411449277


Same hits, but a different angle. Notice that there's no way that the bullets in the engine could have hit the fuel tank.

Different mission, same result:

http://forum.1cpublishing.eu/attachm...1&d=1411449337

Different view of the same hit:

http://forum.1cpublishing.eu/attachm...1&d=1411449337

[ElAurens]

The Hawk 75-4 used the Wright Cyclone R 1820 single row engine, not the P&W Wasp R-1830 twin row engine.

Better do your research before posting any more of your subjective "findings".

[Pursuivant]

Originally Posted by ElAurens

The Hawk 75-4 used the Wright Cyclone R 1820 single row engine, not the P&W Wasp R-1830 twin row engine.

Hmm. I though that it was the other way around, and that the P-36 used the R-1820 and the Hawk 75-4 used the PW R-1830.

Thanks for the correction.

In my defense, I'll say that it's a pain trying to find technical details about all the various Hawk 75 models. Curtiss would sell to anyone and they'd customize just about anything but the basic airframe to make the sale. P-36, French Hawk 75, Finnish Hawk 75, Thai Hawk 75, British Mohawks, etc. It gets confusing!

[Woke Up Dead]

Originally Posted by Pursuivant

"all bullets are treated as being incendiary"

Thanks much for this information, since it sort of proves my point.

Doesn't it actually directly disprove your point? You say "all bullets are treated as incendiary" I show you they're not, you say "my point is proven?"

My research shows that at the beginning of the war incendiary bullets were in short supply for the RAF's .303 guns, but that by 1942 typical belting was 50/50 armor-piercing/some type of incendiary. This source shows that in 1944 German LMGs were also 50/50, while two out of three HMG bullets were incendiary: http://users.telenet.be/Emmanuel.Gus...n/fgun-am.html

Since belting can't change from year to year in the game it's probable that the designers chose mid-war or late-war belting for the guns, when incendiary bullets were available and highly used. In that case half or even 5/6 bullets having incendiary capabilities is historically accurate.

"You'll notice that there is no ball ammo in the mix for any of the guns listed. It isn't even modeled!"
Was ball ammo used in air forces? I could only find references to armor piercing, incendiary, explosive/incendiary, tracers.


"Different mission, same result:"
There's those two or three arrows down low that could have gone through one of the two underfloor fuel tanks, #47 and #48 in the diagram you supplied, assuming the layout of the fuel tanks is the same in the 1830 and the 1820 engined versions.

[TexasJG]

Has anyone looked into the java coding to see how ballistics are actually calculated?
Other than DT I mean...

[Pursuivant]

Originally Posted by Woke Up Dead

Doesn't it actually directly disprove your point? You say "all bullets are treated as incendiary" I show you they're not, you say "my point is proven?"

True. I didn't make my point clear. What I was trying to say was that there is an unrealistically high percentage of bullets that can start fires (tracer, incendiary, explosive) in the standard beltings. AP and ball ammo make holes in things, other types of bullets start fires.

So, if you look at the beltings, you have a very high percentage of bullets that can start fires.

Originally Posted by Woke Up Dead

My research shows that at the beginning of the war incendiary bullets were in short supply for the RAF's .303 guns, but that by 1942 typical belting was 50/50 armor-piercing/some type of incendiary. This source shows that in 1944 German LMGs were also 50/50, while two out of three HMG bullets were incendiary: http://users.telenet.be/Emmanuel.Gus...n/fgun-am.html

That squares with my research, too, but I'm sure that you've done more work on the subject.

Originally Posted by Woke Up Dead

Since belting can't change from year to year in the game it's probable that the designers chose mid-war or late-war belting for the guns, when incendiary bullets were available and highly used. In that case half or even 5/6 bullets having incendiary capabilities is historically accurate.

That makes sense. After all, IL2 models planes at the peak of their performance.

Originally Posted by Woke Up Dead

Was ball ammo used in air forces? I could only find references to armor piercing, incendiary, explosive/incendiary, tracers.

It was used for early war beltings by the RAF due to shortage of better bullets. I could also see it being used by other air forces which had supply problems, or for ground attack missions against personnel and light vehicles. But, getting good info is very hard and I'm not an expert.

Originally Posted by Woke Up Dead

"Different mission, same result:"
There's those two or three arrows down low that could have gone through one of the two underfloor fuel tanks, #47 and #48 in the diagram you supplied, assuming the layout of the fuel tanks is the same in the 1830 and the 1820 engined versions.

Possibly, but the two suspect bullets exited the side of the plane just behind the engine. They might have penetrated the turbocharger/fuel injection system, though, which could also have triggered a big fuel leak. In either case, I could possibly see the hits starting a big fuel leak.

There's also the issue of the bullet completely blowing through an engine cylinder to penetrate whatever is behind the cylinder. That's a lot of energy for a small caliber bullet at 200-300 m, even if it is AP.

But, the big problem is that I was consistently getting that particular result. When I post a screenshot, it might represent a sample of 10 or more trials. I don't post weird "on-off" results, since in combat anything can happen.

That said, I think that the H.75 series isn't bad in terms of damage modeling, although it might be a bit too vulnerable to control run hits, and like all the other planes in the game it's a bit too flammable in that bullets instantly start fires. I'm also still trying to determine which models had armor glass and self-sealing fuel tanks. There are just so many variants!

[Pursuivant]

Woke Up Dead inspired me to run a lot of QMB missions flying planes equipped with the R-1820 engine to see if it's problem with the engine or the DM for particular planes.

As I suspected, it's a DM problem with the Buffalo series, particularly the B-239 (old model, initially had problems with oil leaks) and Buffalo MkI (serious design flaw which meant that the engine was highly vulnerable to overheat, especially in the tropics).

The SBD had a reputation for ruggedness, which carries over in engine DM. The CW-21 was such a rare bird that it didn't have much of a reputation either way, and as a much later model in the game, the DM is better. DT does very good FM and DM work.

Again, screenshots are representative of trends - 10-20 missions flown in QMB.

Dauntless - takes a licking and keeps on ticking!

http://forum.1cpublishing.eu/attachm...1&d=1411517213

CW-21 Demon - a bit more vulnerable to engine damage, but not bad. I had far more problems with fires (unsealed tanks) and pilot kills (no forward armor).

http://forum.1cpublishing.eu/attachm...1&d=1411517213

F2A - Could the fact that it was a lousy airplane, forced into a role it was never designed to fill, which had to fight at a disadvantage against vastly superior foes have influenced DM decisions?

http://forum.1cpublishing.eu/attachm...1&d=1411517213

[Woke Up Dead]

Originally Posted by Pursuivant

True. I didn't make my point clear. What I was trying to say was that there is an unrealistically high percentage of bullets that can start fires (tracer, incendiary, explosive) in the standard beltings. AP and ball ammo make holes in things, other types of bullets start fires.

So, if you look at the beltings, you have a very high percentage of bullets that can start fires.

(...)

That makes sense. After all, IL2 models planes at the peak of their performance.

So in the end I think we agree while still thinking we're disagreeing. There might be too many incendiary bullets for 1939-1941 planes and scenarios, but by 1942 the beltings in-game start to look like like they did in the war. It would be nice if players or mission builders could pick their beltings, but that may not be so easily done.


"There's also the issue of the bullet completely blowing through an engine cylinder to penetrate whatever is behind the cylinder."
But without knowing what the modeled components look like, we can't actually tell what the bullet went through, right? I can't find the thread that showed an image of the components of a radial engine in-game, but I remember that it wasn't just a disk, it was detailed enough that there were V-shaped spaces between the cylinders. So a bullet could pass freely through the V to hit something behind the engine.


"But, the big problem is that I was consistently getting that particular result. When I post a screenshot, it might represent a sample of 10 or more trials. I don't post weird "on-off" results, since in combat anything can happen."
It could be that there are enough critical components modeled in and behind the engine that when you test the way you do (flying directly behind a bomber's six and eating bullets), you will consistently take critical damage.

To simplify things, think of the engine is a big, two-dimensional disk when you look at it head-on. What percentage of the disk's surface area is taken up by components that would cause decreased power, leaks, or other damage that you experience: 20%, 30%, 50%? Let's say that 25% of that disk is vulnerable. If you hit the disk with 3 bullets, your chances of hitting a critical component are already 58% (http://stattrek.com/online-calculator/binomial.aspx). If you assume that 30% of that disk is vulnerable, then your chances of sustaining damage with at least one of the 3 bullets is 66%. If half of that disk is vulnerable then at least one of those 3 bullets will hit a vulnerable spot 88% of the time.