4.12.2 de-bugging

[Pursuivant]

Troubleshooting the DM on the Beaufighter this time.

Cockpit armor and armor glass does what it's supposed to against frontal hits, although it can be penetrated at close range.

One oddity is that compared to most other recently added planes, it's damned near impossible to set the plane on fire or start a fuel leak. Engines are also extremely tough compared to the same engine when mounted in the Wellington, which can be made to die VERY quickly.

http://forum.1cpublishing.eu/attachm...1&d=1405722665

Notice the Beaufighter with it's engines and the leading edges of its wings - directly forward of the inboard fuel tanks - turned into a sieve with no leakage and little damage to the engines.

Unless there's some armor plate there that I don't know about, those fuel tanks ought to be leaking or on fire with that much damage!

[Pursuivant]

Troubleshooting the DM on the Bf-109E series.

This is an old DM and it shows. While I've complained previously about the tendency for the engine to shut down after just a few hits from rifle-caliber bullets fired at extreme range, I've also noticed a few more bugs.

1) The rear fuselage fuel tank is extremely vulnerable - even to hits from the front. Hits from 1 o'clock or 11 o'clock can just bypass the pilot's armor plate to hit the fuel tank, and even one or two rifle-caliber bullets are sufficient to cause unstoppable leaks or even fires. While this is realistic in that the pilot's armor plate didn't fully protect the fuel tank from the front, it's unrealistic for the same reason that a couple of small caliber bullets are unlikely to start a fire in any other self-sealing fuel tank with exhaust gas blanketing.

2) There is a gap between the armor glass and the forward armor plate (for planes where it was fitted).

http://forum.1cpublishing.eu/attachm...1&d=1405745821

This screen grab shows two major problems with the Bf-109E-4 damage model - two rifle caliber bullets fired from about 300 m were sufficient to stop the engine cold. Another bullet has penetrated the gap between the armor glass and the forward firewall (which might not have been armored in the E series) to wound the pilot.

Also notice a penetrating bullet hit from the side which also passes through the pilot model, but which doesn't kill or cause injury.

3) It's very hard to get coolant leaks for the radiators which are located just outboard from the cockpit. But, there's plenty of gun camera footage showing exactly this sort of damage for the Bf-109 series (albeit mostly for the G model).

[Pursuivant]

Ace level AI bombers regularly crash into each other when maneuvering in formation, at least on the Bessarabia map and when setting up QMB missions.

On a similar note, often the third plane in a four plane formation will arbitrarily go into a very steep dive of at least 300 m, then climb to rejoin the formation, possibly to avoid collisions.

By contrast, rookie AI never seems to have this problem, perhaps because they're either slower to react to nearby planes, or because their formations aren't as tight.

I don't know if this is a QMB problem or an AI problem.

In either case, the cause of the problem seems to be that the AI isn't "thinking ahead" far enough to accommodate the very tight turns built into the AI flight paths created for QMB maps.

The quick and dirty solution would be to set QMB AI flight paths so that the turns are much gentler.

The better solution would be to alter AI formation behavior so that the formation "opens up" prior to a sharp turn, with planes on the outside of the formation speeding up prior to the turn and turning later to keep station, and planes on the inside slowing down and starting their turns earlier to keep station. Alternately, planes in formations of four could "cross over", so that the number 1 plane in the formation becomes number 4 and vice-versa. (Not always historical, but it works.)

http://forum.1cpublishing.eu/attachm...1&d=1405747271

Screen grab of two Ace AI Wellingtons about to collide.

http://forum.1cpublishing.eu/attachm...1&d=1405747840

Oops!

[Pursuivant]

Troubleshooting the Corsair I DM

Not surprisingly, there are a few oddities.

1) The fuel tank seems very vulnerable to small caliber bullets.

The armor plate in front of the oil tank, which also provided partial protection to the fuel tank from the front, doesn't seem to be modeled.

Even one hit by a .303 bullet to the fuel tank often starts a big, unstoppable fuel leak. It's as if the fuel tank isn't self-sealing at all.

Regardless of where a bullet hits the fuel tank (i.e., whether above or below the fuel level) you still get a fuel leak. For a big, basically rectangular fuel tank like on the Corsair, that doesn't make sense.

While you don't get the "one shot, instant fire" effect like on some planes, even one later .303 bullet will often start a fire. It's as if every bullet in the game is an Incendiary, API or APEX round.

The overall effect is that the Corsair's fuel system seems very vulnerable, almost as if it were a A6M2.

2) The engine is comparatively vulnerable.

While I haven't compared the Corsair's DM to that of the Hellcat, compared to the P-47D-10, which used virtually the same engine, the Corsair can stand far less damage before the engine conks out. Counting the bullets, I'd say that the same R-2800 engine mounted in a P-47 can take 50% to 100% more damage than if it was mounted in a Corsair.

Additionally, just about any hit to the engine seems to be a cylinder hit, when much of the volume of any radial engine was radiator fins. This might be realistic for .50 caliber and larger bullets, but for a .303 bullet, there's actually a good chance that you'll miss the cylinder.

This isn't a problem unique to the Corsair, but it's one of the ways in which .30/.303 bullets are overmodeled in the game.

3) While it's not exactly an engine-modeling flaw, the Corsair's forward fuselage - the area between the engine and the the fuel tank where the supercharger, etc. were seems to be very vulnerable to damage. Even a very few .303 bullets in the engine, none anywhere near this area of the plane, will trigger a "light damage" result.

This seems strange since contemporary U.S. reports hold that the Corsair was about as tough as the P-47.

4) The machine guns aren't well modeled. Hits to a MG which go right down the barrel don't disable the gun, but hits which arguably might have missed the gun receivers or ammo trays always cause a gun jam. This is a very typical DM problem, especially on the older planes.

5) The wing oil coolers aren't modeled. Bullets that go right into the oil cooler don't cause oil leaks.

6) There's no logic to pilot hits. On one mission I collected an arm wound when none of the bullets actually hit the pilot's arms. On another mission, I merely got a "pilot wounded" result from a bullet right between my pilot's eyes. Normally, that would be a straight "PK" result, or at the very least a "Serious Wound."

Maybe this is a systemic problem, but it seems especially bad on the Corsair I.

7) The armor glass seems to be undermodeled. While I can't say for sure, since I collected most of my "PK" results at relatively short ranges where a rifle caliber bullet might conceivably penetrate armor glass, out of the many QMB missions I flew, I'd estimate that about 15% resulted in PK results through the armor glass.

http://forum.1cpublishing.eu/attachm...1&d=1406108019

Bad DM for the machine guns. Notice a bullet that goes right down the barrel of one gun but doesn't result in a gun hit.

http://forum.1cpublishing.eu/attachm...1&d=1406108155

About those Ace gunners. . . Lots of hits on a fast-moving and maneuvering target at 400+ meters.

http://forum.1cpublishing.eu/attachm...1&d=1406108337

The forward fuselage is extremely vulnerable to damage, as is the engine. Strange considering that the Corsair was considered to be as rugged as the Hellcat and the P-47. Notice, there are more and bigger bullet holes in the light damage texture than there are actual bullets in the plane!

http://forum.1cpublishing.eu/attachm...1&d=1406108531

Just one .303 bullet in the fuel tank starts a gigantic, unstoppable fuel leak.

http://forum.1cpublishing.eu/attachm...1&d=1406108677

A direct hit to the oil cooler doesn't start an oil leak.

[Pursuivant]

Hellcat F6F-3 DM

The engine can survive about as many small caliber bullets as the P-47D-10 can. The armor glass does what it's supposed to. It doesn't seem to pick up fuel tank hits like the Corsair does, and the tank hits it gets stop fairly quickly.

Those are the good points.

1) The forward fuselage between the engine and the cockpit picks up damage at a ridiculous rate. The light damage texture actually shows up more and bigger bullet holes than are required to trigger that texture!

2) The cockpit is also tremendously vulnerable to damage, with just two rifle caliber bullets knocking out the gun sight and several gauges even though they actually never got near the control panel. The damage model actually knocked out more gauges than there are bullets to hit them! This isn't a unique problem to the F6F, but it really shows badly here.

The damage model doesn't take into account the fact that the armor glass is slightly angled, and extends slightly below the level of the cockpit, so bullets fired from 12 o'clock level can get in through the gap to knock out the gunsight or wound the pilot.

3) The wing damage model outright sucks. I don't say that lightly, but it appears that the damage model is offset from the physical model, so that the game thinks that bullets which should actually miss the plane's wing hit it.

As with all other machine gun damage models in the game, the game engine can't tell which hits would cause instant stoppage (e.g., barrel, receiver), which parts will cause eventual stoppage (e.g., ammo boxes), and which parts aren't going to have an effect on the gun's functionality (ejection ports, empty ammo boxes).

Perhaps because of the offset wing DM, hits that should miss the guns actually hit them.

When you're flying the F6F-3, expect to lose at least 20% of your guns within the first few passes against ace bombers.

Lots of aileron hits. A common problem with all the older planes' DM, but especially unrealistic against rifle caliber bullets which produce minimal shrapnel and basically have to intersect perfectly with cable runs or bellcranks to sever or jam even one control surface.

I collected several of these hits over several missions despite just one or two rifle caliber bullets getting anyplace near the actual cable runs.

As is normal for IL2, any control cable hit results in complete loss of control authority, even when it's not possible in real life.

http://forum.1cpublishing.eu/attachm...1&d=1406107200

Proof that the F6F-3's damage model is offset from the visual model. Note the bullet trajectory that missed the wing, but still counts as a hit!

http://forum.1cpublishing.eu/attachm...1&d=1406107466

Proof that the F6F-3's DM for its machine guns is weird - two guns knocked out by two .303 bullets, both of which realistically would have passed under the guns.

http://forum.1cpublishing.eu/attachm...1&d=1406107568

Picture of the gap where the armor glass should be, allowing a bullet to penetrate the cockpit.

http://forum.1cpublishing.eu/attachm...1&d=1406107651

Same hit from the inside. You can see that the bullet that nominally knocked out the gunsight should actually have been stopped by the armor glass.

Also notice overmodeling of cockpit damage. There are two potentially penetrating bullets, one of which should have been stopped by the armor glass, but three damaged cockpit instruments!

[Pursuivant]

Testing the Corsair II and III DM

These aren't as bad as the Corsair I. The engines are equivalent to the Hellcat and P-47 in their ability to take damage before dying. The machine guns aren't as vulnerable to being knocked out by any random .303 hit. Armor glass works like it's supposed to. It's harder to trigger a minor damage result to the forward fuselage. Light damage to the cockpit isn't overmodeled so that it knocks out multiple gauges or the gun sight.

The problems are:

1) Hits to the oil cooler in the wing still don't trigger an oil leak.

2) Aileron Control hits are still too easy to achieve.

3) Hits to machine gun barrels still don't damage the gun.

4) There is still a gap between the armor glass and the forward cockpit armor that allows bullets to get in.

5) The vulnerability of the fuel system is just as bad as in the Corsair I. That is, even one .303 bullet starts a gigantic fuel leak that doesn't stop, catches fire easily, and runs the engine out of fuel in very short order. This occurs even for hits very high in the tank that would realistically be above the fuel fill level, especially after a leak.

[Pursuivant]

Taking a break from getting shot up by Wellingtons, to getting shot up by TBD Devastators. Why? Because there are some flyable planes in the game that can't keep up with the Wellingtons.

I initially tried a formation of 4 U-2VS and discovered that not only are they armed with 12.7mm tail guns (odd, since most of the specifications I've seen have them with 7.62 mm), but was unpleasantly reminded that they're one of the most overmodeled planes in the game in terms of accuracy and effectiveness of their guns. There are modern guided missile batteries that would envy the anti-aircraft effectiveness of the U-2VS in IL2.

So, I chose one of the most hapless planes in the game, the TBD. One formation of 4 Ace TBD Devastators vs. whatever I'm testing, so 4 .30 caliber MG aimed my way.

B-534: First off was the B-534, and I'm happy to report few problems.

The engine can take a fair bit of damage, possibly even a bit too much, before it dies.

The machine guns jam when the barrels are hit.

The fuel tank can take a bit of damage, but just one bullet won't ignite it. Even so, just a few rifle caliber bullets are sufficient to start a fire. Perhaps a bit too vulnerable there, although the tanks weren't self-sealing.

No weird control hits, but then I didn't collect that many bullets on the wing.

The pilot dies frequently, but then the B-534 didn't have any sort of armor glass or armor.

B-239: Next was the B-239 Buffalo. This one's a lot more problematic.

The main problem is the remarkable vulnerability of the engine, which suffers severe loss of power after just a few rifle-caliber bullet hits and dies after just a few more.

It's not just a Buffalo problem, but one of the systemic problems with IL2's damage modeling is that multiple hits to the same part of the engine count multiple times when they should count only once.

For example, in the picture below you'll see that the B-239 has taken a nose full of lead.

Normally, you'd think that it would be fair for the engine to die after that much abuse, but look more closely and you'll notice that many hits miss the engine and just shoot up the cowl, pass between the cylinders, or hit the same cylinders multiple times.

Look more closely, and you'll notice that the bullets really have only damaged 3 cylinders, and possibly some of the spark plugs/ignition harness and possibly put some holes in the crankcase. Since the bullets were fired from several hundred meters, you will have a severely damaged engine, but perhaps not a dead one.

http://forum.1cpublishing.eu/attachm...1&d=1406486898

And what's going on with that big, black stain behind the engine? That's a severe damage texture and few of the bullets actually hit the forward fuselage.

The Buffalo is also vulnerable to strange control hits, where one bullet will take out one or more control surfaces.

It is also vulnerable to fuel tank hits, but then again, it didn't have self-sealing tanks. Despite that, it still takes several bullets to start a fire, so good damage modeling there.

TBD: Finally, let's look at the TBD damage model. I think that TD overmodeled just how fragile the TBD was. At the time of its introduction, the TBD was the first all-metal, stressed skin monoplane in the U.S. navy's inventory. For its time, it was reasonably tough.
But, in the game, it's incredibly vulnerable.

Additionally, they just made some DM mistakes.

Hits in these pictures are .50 caliber hits from a B-239.

Look at the screenshots below.

http://forum.1cpublishing.eu/attachm...1&d=1406488117

1) The damage model for the wing is offset so that a bullet that actually missed the wing is modeled as if it hit it.

2) Hits to the inboard wing section trigger a damage texture on the outboard wing.

3) Two 0.50 MG hits are enough to trigger a Light Damage texture in the wing, giving the odd situation of more bullet holes than bullets which actually hit the plane!

4) While you can't really see it, one bullet in the horizontal stabilizer was enough to trigger a severe damage texture for that part, with a completely shredded control surface.

5) This damage is also somehow severe enough that the AI considers it to be worth aborting the mission.

http://forum.1cpublishing.eu/attachm...1&d=1406488117

In the second screenshot, notice that damage to the middle fuselage actually triggers a light damage texture for the rear fuselage, and that one glancing hit to the inner wing is sufficient to trigger the light damage texture for that part.

Again, you've got the odd situation of more holes in the plane than there were bullets that hit it!

Ignore the "I'm on Fire" result, that's just part of the AI bailout routine triggered by the PK hit.

[Woke Up Dead]

You guys remember "the bar" in the 190? I think other cockpits may also be suffering from the bar, particularly the Yak 9. Someone on another forum posted a good video showing why the 190 bar should be reduced in game due to refraction of light through glass, it might apply to "bars" at the top of many gunsights and to armored glass in other planes as well; details in this thread: http://forum.1cpublishing.eu/showthr...972#post705972.

[Woke Up Dead]

A minor bug: I have edited my ffb files so that I don't get my Microsoft FFB joystick shaking when I squeeze the trigger. I still get a small "bump" when I release bombs or rockets from a fighter. Funny thing is, I then still get that bump when I squeeze the trigger for the machine guns or cannons. This does not happen if I don't add bombs or rockets to the loadout in the first place.

[Pursuivant]

Originally Posted by Woke Up Dead

You guys remember "the bar" in the 190? I think other cockpits may also be suffering from the bar, particularly the Yak 9.

I think you're right. Just about all the cockpit models in IL2 fail to take the refractive properties of armor glass into account. I didn't realize exactly what was going on myself until I saw the video.

Just about any cockpit with a thick bottom portion of the armor glass frame, or where armor glass is faired into the aircraft's fuselage, is going to have some degree of refraction which reduces the apparent size of the frame or fairing.

Realistically, though, this also means that any damage to the armor glass that destroys the refractive properties is also going to make the "bar" (i.e., fairing or frame) suddenly appear, as well as making any image seen through the armor glass appear is if it were reflected in a cracked mirror.

Also, the refractive properties of the glass will be obvious as a "step down" between the view in the armor glass and the view through the unarmored adjacent canopy areas.