Change of fuel tanks fire rate.

[MDzmitry]

While I've been digging through combat reports on an unrelated topic, I've noticed a significant discrepancy between cases of fuel tank fires in real life and in Il-2. For example, P-51 pilots describe in their reports quite often cases of fuel tanks being set on fire after the first or second attack on an enemy. Meanwhile from what I've seen in Il-2, one's much more likely to set an engine on fire than a fuel tank. To be frank, I don't think I've ever once seen a fuel tank fire not caused by cannon hits.

I know that HE shells have a chance of getting fuel tanks on fire or even exploding them, but from what I've seen "kinetic" rounds (AP) usually end up only causing leaks and nothing else.

 

Now, onto the suggestion:

Simple option: just give fuel tanks a % chance of fire on kinetic hits. Since incendiary effect isn't modelled it could be side-stepped by giving the fuel tanks such a characteristic. For example: 3-5% chance if hit by a rifle-caliber bullet, ~10-12,5% chance if hit by large-caliber machine guns, and ~15-20% chance if hit by 20mm. Mind that chances don't stack the longer one fires, so it won't be a guaranteed fire after 10 .50cals or 5-7 20mm rounds.

A harder option: same core concept, but on top of that to consider leaks and armour/other modules as fire chance % multipliers (2x and 0,5x respectively as an example). The latter can be quite bothersome to deal with though, since in many cases (as with Bf 109's layered fuel tank armour) it proved barely effective against anything bigger than rifle caliber. The armour either stopped the round entirely, or barely affected its capability to cause fire.

 

Скрытый текст

To support this idea: basically every round we have in the game is also supposed to have Incendiary effect: 

- API for ShVAK,

- API/APHE for MG151, 

- SAPI for Hispano

- APIs for 7,62, 7,92, .303s and all high-caliber MGs. 

- All HE shells (except for 1/2 of early ShVAK shells and not represented early French Hispano rounds) had incendiary mixture in them.

 

At this point why bother modelling incendiary damage and not just slap a fire chance on fuel tanks from aforementioned "Kinetic-type" rounds? That would also be the least bothersome way to give smaller-caliber armament a fighting chance in comparison with cannons.

UPDATE: With the new data, observations and real-life examples listed and sourced in this topic, the suggestion gets closer to the idea of "decreasing the percentile gap in fuel tank fire probability" (an example: from 2-60 12,7mm hits to 10-30 hits) with pronounced lower and upper limits to avoid infuriatingly extreme cases of RNG for both the victim (if it gets lit up with a slight burst) and the attacker (setting up an attack from a blind spot only to see a simple fuel leak). That could perhaps be accompanied by decreasing the average amount of ammo spent slightly (we're talking in the frame of the average of 25-33 changing to ~20, not to get too radical).

 

Edited July 3 by MDzmitry

[MDzmitry]

I'd also appreciate if anyone had enough time and patience to collect some data on rates of fires and their kinds (fuel tanks or engines) depending on the armament. 

That would either back up this suggestion or refute it (but put me to rest with one less thing to think about in the middle of the night).

[354thFG_Leifr]

Since incendiary is never going to happen in BoX, and it being a core component of the American ammunition belts, a good part of the effectiveness of the weapon is completely absent. I suspect this often contributes to the negative belief that .50s are unpredictable and sometimes unreliable. I am led to believe that the greater part of the BoX DM is simply reliant on effective dice rolls chances for each hit box, I therefore see no problem with either of the above implementation in absence of true incendiary.

 

12 hours ago, MDzmitry said:

I'd also appreciate if anyone had enough time and patience to collect some data on rates of fires and their kinds (fuel tanks or engines) depending on the armament. 

That would either back up this suggestion or refute it (but put me to rest with one less thing to think about in the middle of the night).

 

This might be something for @357th_KW to provide.

Edited May 31 by 354thFG_Leifr

[Aapje]

Strong agree with this change to make lower-caliber planes a little more effective. It might be relatively easy to implement, given that fuel tank fires are already modeled.

 

Between not modelling incendiary at all, and the simple option, I think that the latter is closer to reality.

 

@LukeFF Perhaps you can ask the developers whether this is a feasible change?

[LukeFF]

We'll see, but I wouldn't expect any changes to the damage modeling at this point. 

[357th_KW]

On 5/30/2025 at 12:54 PM, MDzmitry said:

While I've been digging through combat reports on an unrelated topic, I've noticed a significant discrepancy between cases of fuel tank fires in real life and in Il-2. For example, P-51 pilots describe in their reports quite often cases of fuel tanks being set on fire after the first or second attack on an enemy. Meanwhile from what I've seen in Il-2, one's much more likely to set an engine on fire than a fuel tank. To be frank, I don't think I've ever once seen a fuel tank fire not caused by cannon hits.

I know that HE shells have a chance of getting fuel tanks on fire or even exploding them, but from what I've seen "kinetic" rounds (AP) usually end up only causing leaks and nothing else.

 

Now, onto the suggestion:

Simple option: just give fuel tanks a % chance of fire on kinetic hits. Since incendiary effect isn't modelled it could be side-stepped by giving the fuel tanks such a characteristic. For example: 3-5% chance if hit by a rifle-caliber bullet, ~10-12,5% chance if hit by large-caliber machine guns, and ~15-20% chance if hit by 20mm. Mind that chances don't stack the longer one fires, so it won't be a guaranteed fire after 10 .50cals or 5-7 20mm rounds.

A harder option: same core concept, but on top of that to consider leaks and armour/other modules as fire chance % multipliers (2x and 0,5x respectively as an example). The latter can be quite bothersome to deal with though, since in many cases (as with Bf 109's layered fuel tank armour) it proved barely effective against anything bigger than rifle caliber. The armour either stopped the round entirely, or barely affected its capability to cause fire.

 

  Reveal hidden contents

To support this idea: basically every round we have in the game is also supposed to have Incendiary effect: 

- API for ShVAK,

- API/APHE for MG151, 

- SAPI for Hispano

- APIs for 7,62, 7,92, .303s and all high-caliber MGs. 

- All HE shells (except for 1/2 of early ShVAK shells and not represented early French Hispano rounds) had incendiary mixture in them.

 

At this point why bother modelling incendiary damage and not just slap a fire chance on fuel tanks from aforementioned "Kinetic-type" rounds? That would also be the least bothersome way to give smaller-caliber armament a fighting chance in comparison with cannons.

 

 

 

The fuel tanks already have a chance of catching fire from kinetic hits - that chance is VERY small though.  I last tested this in version 4.006 so things have changed somewhat since then, but I setup a parked, running target aircraft on the ground 20m or so away from my parked bomber and used the gunner to test shoot various things to ensure perfect accuracy.  In the case of fuel tanks, I setup shots from both the rear and from the side, assuming penetrating though the aircraft would change the results.

 

As an example, the 109K took an average of 79 hits (ranging from a low of 25 to a high of 203) from the rear to ignite the fuel tank and 50 from the side with .50.  With a 12.7mm UB (which I believe has a small HE component), the results were 70 and 35, and with a 20mm MG/FF (firing a mix of HE and AP) they were 26 and 11.  So it is possible to ignite those fuel tanks with AP only, but the chance of landing that many strikes on a very small portion of a moving target are so low that it's essentially impossible to do in actual game play (a very similar effect to the changes just made to the 109, discussed in the wing durability thread).  

 

Here are some various historical documents on the subject:

Incendiary Effectiveness of 20mm projectiles

Effectiveness of Small Arms Incendiary Ammunition at High Altitudes

Effectiveness of Incendiary Ammunition Against Aircraft Fuel Tanks

50 caliber M1 and M8 tests

 

In testing, the probabilities of ignition given are very high for .50 and 20mm incendiaries - often in the 50-100% range.  It's seems likely that in the real world they weren't quite that effective just from reading combat reports and watching gun camera film.  And of course, things like range, angle, how full the fuel tank was, altitude, temperature, differences between different nations self sealing fuel tanks etc. all played a factor (some of which are explored in depth in those studies I posted).

 

 

[Aurora_Stealth]

Yep, just to add my few cents here... I do think fuel tank fires are rarer than they should be. Statistically, fuel fires were one of the common failure modes associated with aircraft loss (up there with radiator/coolant loss) so accommodating quicker ignition of fuel tanks knowing that... at this stage incendiary rounds won't be added ingame seems like a good compromise.

 

Just to add some clear demarkation here though... wing failures were comparably very rare statistically in terms of the percentage of aircraft losses (including the Bf 109). More often than not, something else (especially fuel tanks, engine seizure or pilot kill) would almost always come first because it's so much easier to achieve practically speaking... I think the recent change ingame to wing durability for the Bf 109 has just highlighted further that fuel fires aren't occurring as much as they should be.

[MDzmitry]

2 часа назад, Aurora_Stealth сказал:

wing failures were comparably very rare statistically in terms of the percentage of aircraft losses (including the Bf 109). More often than not, something else (especially fuel tanks, engine seizure or pilot kill) would almost always come first because it's so much easier to achieve practically speaking... I think the recent change ingame to wing durability for the Bf 109 has just highlighted further that fuel fires aren't occurring as much as they should be.

Agree, having an objectively broken aspect fixed simply highlighted another problem that has been present in the meantime, but was (so to say) "smoothed out" by said broken aspect.

 

5 часов назад, 357th_KW сказал:

The fuel tanks already have a chance of catching fire from kinetic hits - that chance is VERY small though.  I last tested this in version 4.006 so things have changed somewhat since then, but I setup a parked, running target aircraft on the ground 20m or so away from my parked bomber and used the gunner to test shoot various things to ensure perfect accuracy.  In the case of fuel tanks, I setup shots from both the rear and from the side, assuming penetrating though the aircraft would change the results.

Thanks for the in-game data and real-life documentation! Seeing the statistics involved with testing incendiary effect of various ammunition types I can easily understand why the Devs don't touch this aspect of DM, but as you've said the current chances of fire are a bit too unreal to achieve in a battle.

 

Still, I would personally refrain from having the percentages akin to your last posted document, since the lower number would compensate for not having to account for altitude, fuel type, armour, incendiary type and so on.

Basically speaking, considering how far we are into the development of the project, I'd rather see a simple compromise than an overcomplicated solution given up on.

[MDzmitry]

A small addition to the ways of implementation: if fuel tank fire is subject to achieving a specific amount of damage from ammunition, said volume of damage could be lowered instead of adding any percentage. 

The approximate amount of damage could be discussed further, since in this case it would become accumulative. Still, in my opinion, percentage remains a better option to consider.

Edited June 3 by MDzmitry

[[CPT]Crunch]

What ever happened to the Fw-190's front oil cooler fires, been forever since I've seen one lite off, they used to be seen fairly frequent.  Seems a lot of damage mechanics disappear after time, someones been tweaking them.

[Roland_HUNter]

Hi everyone,

 

I just wanted to point out that the Bf 109 catching fire is not uncommon.
I've encountered it several times, and it seems to happen under various conditions.

 

To illustrate this, I’ve attached a track file you can download and review here:
https://limewire.com/d/6KPUB#cpetKLmlVt

PS: At 0:31 around 10-12 hits ignited the fuel tank.
 

Please note: the mission name is originally "P-51", but I changed the aircraft to the P-39 in the mission editor.

 

Looking forward to hearing your thoughts!

Best regards,

Roland_HUNter

[MDzmitry]

55 минут назад, Roland_HUNter сказал:

Hi everyone,

 

I just wanted to point out that the Bf 109 catching fire is not uncommon.
I've encountered it several times, and it seems to happen under various conditions.

 

To illustrate this, I’ve attached a track file you can download and review here:
https://limewire.com/d/6KPUB#cpetKLmlVt

PS: At 0:31 around 10-12 hits ignited the fuel tank.
 

Please note: the mission name is originally "P-51", but I changed the aircraft to the P-39 in the mission editor.

 

Looking forward to hearing your thoughts!

Best regards,

Roland_HUNter

Looks like not as bad a situation as in @357th_KW case with the average of 79 hits, but from what I saw I'd still call this inconsistent. You mention the F-4 that lit up quickly, but before that G-6 ate up 8-9 bursts of varying longevity.

(Upd: I tried re-counting it with consideration for missed shots, it took around 38 hits minus the misses)

Kill number 3 was a pilot kill, so out of the equation

F-2 ate up around 20 hits, which is adequate in my opinion

G-6 Late around 30 hits

G-2 is a PK

G-4 PK

G-14 took around 28-30 hits

G-6AS also around 30 hits

K-4 around 20 hits

For now the average seems to be around 25 hits required (taking minimal assumed numbers of hits out of the ones listed) at the firing distance of <100m
 

Good news: the situation doesn't look as dire

Bad news: the test was conducted at extremely close range on static targets, so multiple concerns remain:

1) Does distance affect fire chances?

2) Considering the significantly less optimal conditions in a fight, how much time/passes is necessary to score enough hits? This question is purely a skill issue though, so Q.1 is a more interesting point for me. After all it's way more likely to see distances around 200-400m, and from what I remember the "effective range" of a .50cal is said to be 400m.

Could you please conduct a similar test, just with a different distance? If the results are the same, then damn are we just bad pilots

Edited June 3 by MDzmitry

[Roland_HUNter]

41 minutes ago, MDzmitry said:

Could you please conduct a similar test, just with a different distance? If the results are the same, then damn are we just bad pilots

I'll try to test that tomorrow.

[MDzmitry]

9 минут назад, Roland_HUNter сказал:

I'll try to test that tomorrow.

Thanks in advance, I'd suggest trying 300-400m to maximize the chance of seeing the discrepancy if there is any related specifically to distance.

[Aurora_Stealth]

2 hours ago, MDzmitry said:

Looks like not as bad a situation as in @357th_KW case with the average of 79 hits, but from what I saw I'd still call this inconsistent. You mention the F-4 that lit up quickly, but before that G-6 ate up 8-9 bursts of varying longevity.

(Upd: I tried re-counting it with consideration for missed shots, it took around 38 hits minus the misses)

Kill number 3 was a pilot kill, so out of the equation

F-2 ate up around 20 hits, which is adequate in my opinion

G-6 Late around 30 hits

G-2 is a PK

G-4 PK

G-14 took around 28-30 hits

G-6AS also around 30 hits

K-4 around 20 hits

For now the average seems to be around 25 hits required (taking minimal assumed numbers of hits out of the ones listed) at the firing distance of <100m
 

Good news: the situation doesn't look as dire

Bad news: the test was conducted at extremely close range on static targets, so multiple concerns remain:

1) Does distance affect fire chances?

2) Considering the significantly less optimal conditions in a fight, how much time/passes is necessary to score enough hits? This question is purely a skill issue though, so Q.1 is a more interesting point for me. After all it's way more likely to see distances around 200-400m, and from what I remember the "effective range" of a .50cal is said to be 400m.

Could you please conduct a similar test, just with a different distance? If the results are the same, then damn are we just bad pilots

 

Interesting.

 

Range should affect penetration power, so theoretically that should mean more rounds/time required to cause a fire as distance increases but it may not be a significant difference if within effective range.

 

In real life I'd expect something like a one second burst from a P-51 D (60 to 70 API rounds against the aircraft with perhaps 5-10 rounds actually hitting through the self sealing fuel tank) to cause a fire. But that would be at an 'ideal' range of 200 - 350m and it assumes an accurate gun solution and a concentration of hits in the fuselage, which also implies that to get a consistent result like that the target is not evading/maneuvering erratically.

 

Your second point captures precisely why even in real life the above could and would still be frustrating to achieve against a competent pilot who is taking evasive maneuvers and so multiple bursts of fire could be required. In either case it should be consistent when the target was unaware or caught straight and level.

[357th_KW]

I went ahead and recreated my old test that I mentioned above and ran 10 passes from the rear and the side against the K4 in 6.001.

 

The results were an average of 33.6 hits from the rear and 39.4 from the side.  Low/high from the rear were 8/63 and 13/68 from the side.  I recorded one pass from each angle to demonstrate how this was tested.  Rear angle.  Side angle.

 

My guess is that the fire effect is simply RnG based on the damage delivered to the tank.  The improvements that the .50s received some time back made a noticeable difference vs. 4.007 when they were nearly worthless.  The fact that the rear angle (which has to penetrate more material) took less hits might indicate that there is more in play here - or it might just be noise from small sample sizes.

 

These shots in my test were made at 50m - in this historical test they were shooting at 175 yards with M1 Incendiary and M8 Armor Piercing Incendiary and achieving ~2 hits per ignition, so vastly better performance for real incendiaries vs. what our AP does in the game.

 

 

 

 

[MDzmitry]

5 часов назад, Aurora_Stealth сказал:

Range should affect penetration power, so theoretically that should mean more rounds/time required to cause a fire as distance increases but it may not be a significant difference if within effective range.

 

In real life I'd expect something like a one second burst from a P-51 D (60 to 70 API rounds against the aircraft with perhaps 5-10 rounds actually hitting through the self sealing fuel tank) to cause a fire.

From what I understand, penetrating power should have little to do with the fire probability, after all two conditions were to be met for a fire:

1) penetration in general

2) incendiary taking effect

 

As 357th_KW points out, the real-life performance is still vastly better than what we have in the game, and to be frank I'd still like to see the average amount of hits necessary lowered. Not to the average of 2 of course, but I still find 25-33 a bit of a stretch to pull off in a fight.

 

Maybe this could be achieved by lowering the "high" part of the percentile in our cases (38 in Roland's recording, 60+ in KW's), since scoring a dozen hits on a tank is a pure minimum that should be asked of any and every pilot wishing for a kill, lower numbers would perhaps make the situation a bit unbearable for the "victim".

Having less of a discrepancy in chances would also make the outcome of a pass (as in on an unaware target) more predictable for the attacker.

Edited June 4 by MDzmitry

[Aurora_Stealth]

4 hours ago, MDzmitry said:

From what I understand, penetrating power should have little to do with the fire probability, after all two conditions were to be met for a fire:

1) penetration in general

2) incendiary taking effect

 

As 357th_KW points out, the real-life performance is still vastly better than what we have in the game, and to be frank I'd still like to see the average amount of hits necessary lowered. Not to the average of 2 of course, but I still find 25-33 a bit of a stretch to pull off in a fight.

 

Maybe this could be achieved by lowering the "high" part of the percentile in our cases (38 in Roland's recording, 60+ in KW's), since scoring a dozen hits on a tank is a pure minimum that should be asked of any and every pilot wishing for a kill, lower numbers would perhaps make the situation a bit unbearable for the "victim".

Having less of a discrepancy in chances would also make the outcome of a pass (as in on an unaware target) more predictable for the attacker.

 

Sorry - I was referring to the fact that, practically speaking if you were outside of effective range or on the edge of it its going to become significantly harder to actually penetrate through the airframe and obstructing internal elements to reach the fuel tank in the first place. That's not so much a problem in a test environment when calibrated at a desired solution / angle; or you can get a good deflection angle easily because your opponent is unaware or on a straight course.

 

I have no doubt whatsoever an API round can do its job of lighting up fuel tanks provided it can actually reach them (which it will do when within effective range). It's the same story as with the DeWilde rounds the RAF switched to for the same reason... they weren't getting particularly consistent results against the Luftwaffe with standard AP but that changed a lot when they could use incendiary rounds and they were closing the distance effectively.

 

The fact that rounds are still struggling ingame at 50 - 100m is concerning and that report is something isn't it; it really highlights what an API round can do when at very close range under ideal conditions (175 yards / 160m - that's pretty close). It would be interesting to see the method and some images of the test with that replica Heinkel in the report (it sounds like a static test to me, at sea level and 'at altitude' on that mountain in Chile) but it does seem to illustrate well what the round is capable of under ideal test circumstances. Once you're into effective range I'd expect most of the ballistics issues beyond it would disappear and the emphasis is just getting rounds on target.

 

Oh I agree, its currently very difficult to produce a fire and the variation possible is surprisingly high; the upper percentile in terms of numbers are off if we were to apply the context of an API round designed to specifically cause that effect (but API is not being used)... that's why we're talking and sharing. I'll take a look at the videos uploaded later when I'm back home.

Edited June 4 by Aurora_Stealth

[MDzmitry]

@Aurora_Stealth we should take a step back and re-evaluate what's been discussed so far to be clear:

- kinetic rounds in the game are never stated to be incendiary, so we can't demand that they have high fire probability;

- this is a problem by the fact that every side in WW2 switched rapidly to API and HEI ammunition;

- yet, in the game AP ammunition also somehow manages to cause fires with all (I assume) types of fuel tanks no matter their construction.

 

So far it looks like poor kinetic rounds are having an identity crisis and cannot decide whether they want to be purely AP or API. From this come two inconsistencies:

1) the historical one with incorrect ammunition types being used;

2) the gameplay one with a severe difference in fuel tank fire probabilities (RNG be blessed).

 

I'm all for changing the names to have an "I" at the end and setting the upper hit percentile to be lower, thus making fuel tank fires more consistent. 

But it will remain a suggestion rather than an outright bug report.

(I'm pointing this out because the rhetoric could devolve into bashing the Devs instead of suggesting possible improvements, and I don't think either side would want that.)

 

[Aurora_Stealth]

31 minutes ago, MDzmitry said:

@Aurora_Stealth we should take a step back and re-evaluate what's been discussed so far to be clear:

- kinetic rounds in the game are never stated to be incendiary, so we can't demand that they have high fire probability;

- this is a problem by the fact that every side in WW2 switched rapidly to API and HEI ammunition;

- yet, in the game AP ammunition also somehow manages to cause fires with all (I assume) types of fuel tanks no matter their construction.

 

So far it looks like poor kinetic rounds are having an identity crisis and cannot decide whether they want to be purely AP or API. From this come two inconsistencies:

1) the historical one with incorrect ammunition types being used;

2) the gameplay one with a severe difference in fuel tank fire probabilities (RNG be blessed).

 

I'm all for changing the names to have an "I" at the end and setting the upper hit percentile to be lower, thus making fuel tank fires more consistent. 

But it will remain a suggestion rather than an outright bug report.

(I'm pointing this out because the rhetoric could devolve into bashing the Devs instead of suggesting possible improvements, and I don't think either side would want that.)

 

Great points, yes you're right - couldn't agree more.

 

I'm going to edit my post because it does end up reading off the wrong way in that respect.

[MDzmitry]

I don't think editing is necessary, it's clear from the context that there's no I'll intent

[Roland_HUNter]

23 hours ago, MDzmitry said:

Could you please conduct a similar test, just with a different distance? If the results are the same, then damn are we just bad pilots

Here's the promised test! Around 3:40, I managed to ignite the fuel tank from ~380 meters. I counted the hits in slow motion — it took about 20–30 rounds.

 

Apologies for the inaccurate aiming — I'm using a pretty old joystick.

 

I'd also like to add that the whole thing feels very RNG. Sometimes it catches fire after just 10–12 hits, other times it takes around 30–40.

 

https://limewire.com/d/y8WWn#b4foRNU0I3

[MDzmitry]

1 час назад, Roland_HUNter сказал:

Here's the promised test! Around 3:40, I managed to ignite the fuel tank from ~380 meters. I counted the hits in slow motion — it took about 20–30 rounds.

Thanks a lot! Then I guess the distance really doesn't matter as a factor in this case.

1 час назад, Roland_HUNter сказал:

Apologies for the inaccurate aiming — I'm using a pretty old joystick.

No worries, I can promise I'd be a worse shot lol

1 час назад, Roland_HUNter сказал:

I'd also like to add that the whole thing feels very RNG. Sometimes it catches fire after just 10–12 hits, other times it takes around 30–40.

Yeah, this is currently one of the primary complaints. Discussed it in the messages above if you'd like something to read instead of the morning newspaper.

[Roland_HUNter]

12 hours ago, MDzmitry said:

Thanks a lot! Then I guess the distance really doesn't matter as a factor in this case.

No worries, I can promise I'd be a worse shot lol

Yeah, this is currently one of the primary complaints. Discussed it in the messages above if you'd like something to read instead of the morning newspaper.

Yesterday I was flying around in Berloga, and to my surprise, I managed to set a 109 on fire with just one or two hits from the Yak-9’s 12.7mm. The RNG is absolutely crazy.

[Aurora_Stealth]

21 hours ago, Roland_HUNter said:

Here's the promised test! Around 3:40, I managed to ignite the fuel tank from ~380 meters. I counted the hits in slow motion — it took about 20–30 rounds.

 

Apologies for the inaccurate aiming — I'm using a pretty old joystick.

 

I'd also like to add that the whole thing feels very RNG. Sometimes it catches fire after just 10–12 hits, other times it takes around 30–40.

 

https://limewire.com/d/y8WWn#b4foRNU0I3

 

Thanks for performing that test, just had a chance to look through it in detail - it was quite a good one! I would say some nice shooting there actually and a neat test setup.

 

So of the 10 targets, half of them caught fire plus a couple of pilot kills and a couple with radiator damage / coolant loss and one AI who fell out of formation and blacked out temporarily to recover later.

 

It's good to see that fires are actually being caused when targeting the fuel tanks, its just surprising how many bursts of fire on average it takes to cause them... especially considering how accurate you're being there (compared to a combat situation) with that nose mounted armament and taking your time to get precise hits.

 

As you say it does vary, and occasionally you'll get a quick result but frankly ~10 bullets should be sufficient here... yet in most cases its three or four times that number which seems excessive. Its easy to see how any aircraft with wing mounted guns is going to be even more inconsistent and harder to achieve this in combat situations.

Edited June 5 by Aurora_Stealth

[Roland_HUNter]

2 hours ago, Aurora_Stealth said:

As you say it does vary, and occasionally you'll get a quick result but frankly ~10 bullets should be sufficient here... yet in most cases its three or four times that number which seems excessive. Its easy to see how any aircraft with wing mounted guns is going to be even more inconsistent and harder to achieve this in combat situations.

Honestly, I feel like War Thunder has too much of an influence on us.

 

I was thinking about why we don’t see sudden fires either in gun camera footage or in the game, the way many people in this thread seem to expect.

 

I considered whether we should go into details like how projectiles deform on impact, etc.

 

But then I remembered: the Japanese front.
Japanese planes didn’t have protected fuel tanks, and the tanks were located in the wings. I know Japanese aircraft were notorious for catching fire easily, yet in the video linked below, it clearly takes quite a few hits before the Japanese plane actually catches fire.

 

Also, not every round has a tracer, and based on the slow-motion footage, the number of hits could have been 20–30 or even more.

If you watch the whole video, you’ll see that RNG was a big factor in real life too. Some planes caught fire after just 5–10 hits.

 

But, back to 109s:

Many fuel tank hits.

 

And here the same.

 

Again the same.

 

You can also see that external fuel tanks sometimes ignite or even explode after just a few hits — but the main fuel tank inside the aircraft often doesn’t.

 

If you watch both videos all the way through, you'll notice that even against 109s and 190s, there were some cases where a few hits caused a fire — but based on the footage, that wasn't exactly common.
 

[357th_KW]

I think you also have to keep in mind that a fire might not propagate instantly after ignition inside the aircraft.  These gun cameras only had a couple seconds of overrun at most (some had no overrun, particularly earlier in the war) and so the outcome of the engagement might not be captured.  As an example, here is the full film of one of those clips linked above:  

 

 

And here is the AAR from the pilot: 

 

As we can see, it's actually three separate engagements.  The third target shows no evidence of fire in the film, but the aircraft was seen to spin into the overcast in flames.

 

And with all that said, you do see some cases of sudden fuel tank fires:  

 

[MDzmitry]

6 часов назад, Roland_HUNter сказал:

But, back to 109s:

Many fuel tank hits.

 

And here the same.

In both of these cases there's just a handful of flashes on the fuselage before a flame appears.

Mind you, that flashes seen on the fuselage are not only tracer rounds, each of those is an API, the visual recognition of which was mentioned by the pilots (hits were easier to identify). 

In your first linked video most of the first burst's hits were on the starboard wing, with approx. 4 on the fuselage around fuel tank. After the 2nd burst the tank was in flames.

On the 2nd video similar situation, there's a passing burst with only a couple of hits on the wing and the fuselage. Second burst and right before a cut off we see a something erupt, but that's more likely glycol leaking, at least seems like wing root to me.

6 часов назад, Roland_HUNter сказал:

You can also see that external fuel tanks sometimes ignite or even explode after just a few hits — but the main fuel tank inside the aircraft often doesn’t.

If you rewatch the video, most of the hits observed were around the aircraft: some on both wings, a bunch on the rear section of the tail.

And considering the deflection angle I've counted (on 0,25x) that there were around 4-6 flashes that were directly on the fuel tank.

 

Again, we're not discussing setting the limit of hits to 2 per fire, it's more around the value of 10-20 hits with definite lower and upper caps on damage so as to minimize cases of extremely hostile RNG for either side. And 10-20 seems reasonable enough based on what you've linked in your videos.

[MDzmitry]

Also I'm not going to derail the discussion too much, but War Thunder is a whole different can of worms with more serious problems than fuel tank fire probabilities. 

If I were to compare, I'd say Il-2 has WT beat in terms of how the damage feels (its impact, visuals and "time to kill") 8 times out of 10, the only nitpicks being said fire probability and large-caliber guns' damage.

 

If you need a couple of examples of the ugly part of WT DM, send me a message, but in terms of this topic I won't elaborate further.

[Aurora_Stealth]

Lets be cautious reading too much into video footage or individual accounts as its always circumstantial / situational, but yeah its not so straight forward when trying to extrapolate information from gun camera footage. The tests that have recently been performed in game with controlled conditions/parameters represent something of a baseline on which to compare effects while not introducing all the other variables that can affect this.

 

I've had some interesting conversations with technical staff at Duxford air museum here about wreckage / fuselage of an A6M3 (it ended up being a 35 minute conversation discussing material specs, thicknesses and construction of Japanese aircraft... my partner about to string me up from the hanger doors by the end) there's a lot of mythology from western perceptions around how they were built and to what comparable standard. There are some limited areas where the Japanese were actually superior in mechanical construction elements at least early on; but yes many of their aircraft did not have proper fully self-sealing fuel tanks... although they made some efforts to partially implement protective measures later in the war but those were modest. And finally, the fuel tanks of many of the Japanese aircraft are also in the wings and not just in the fuselage so its a somewhat different proposition here with the Bf 109 (bullets have to hit fuselage, if rear hit through fuselage: could hit tail wheel mech, could hit radio & its mount, could hit an 8mm thick armoured plate behind the fuel tank etc depending on model). But... lets not digress further into all that or extrapolate too much.

 

The other thing regarding specifically tracers is they tend to be a different weight to the other rounds so they can somewhat mislead the pilot. You may see tracer on target (I'm not talking API that can cause flashes) but are the rest of the bullets without tracer really going where you think? it can be helpful, but if you're targeting something very specific especially with a wing mounted armament with your guns dispersed and your wings flexing (as you're manoeuvring) it might mislead. We shouldn't go too far down these rabbit holes because that's why we're testing under certain parameters to avoid these variables with an apples to apples comparison.

 

Having played WT many years in the past, among a host of sims in the last twenty or more years - I would say IL2: GB still has far superior if not some of the best overall game mechanics and modelling... with the exception of Cliffs of Dover which details even individual part & system failure but that game has its own issues its trying to overhaul.

Edited June 6 by Aurora_Stealth

[MDzmitry]

Updated the original message slightly to include the narrowed down suggestion according to the points discussed here