Braced vs cantilever design - engineers needed

[=IRFC=Gascan]

15 minutes ago, unreasonable said:

This historical record, taken as a whole, clearly refutes the "win the lottery theory". Losing wings under wire was a material risk for these planes. 

 

Absolutely. Airplanes getting shot can break apart. The concern is not that it should never happen, but that it is happening more frequently than it should for airplanes with wire braced wings compared to cantilever wings.

 

I would estimate that roughly 50% of my deaths in any Entente plane (except for the Bristol) are due to wings shearing off after minor damage from a low probability shot. The rest are due to pilot snipes or control loss, depending on how nasty the stall is (I lose more Camels to control loss than the SE5a). Very Rarely will someone light me on fire, cause my engine to fail, or saw my wing off with sustained fire.

Edited September 12, 2021 by gascan
added death estimates

[unreasonable]

Just now, gascan said:

 

Absolutely. Airplanes getting shot can break apart. The concern is not that it should never happen, but that it is happening more frequently than it should for airplanes with wire braced wings compared to cantilever wings.

 

That is perfectly possible, but to establish that claim, you have to quantify:

 

How often it does happen, &

How often it should happen

 

For each of braced and cantilever wings.

 

A good start is to examine the record of what actually did happen with an open mind, rather than ignoring it or grossly misrepresenting it because you do not like what it says. 

 

[US41_Winslow]

13 hours ago, J2_Trupobaw said:

There is survivor bias in combat reports - people whose planes floped to pieces tended not to write them.

Not in combat reports.  They should give a better indication of what actually happened than memoirs since, presumably, the pilot would want to be as clear as possible to get awarded a confirmation.  So, if someone sees the wings fall off the airplane they are shooting, they most likely would report this.  The only issue is that OOC kills would be reported too often but since, from what I’ve seen, there is not a huge difference between the number of Fokkers that shed their wings and other types, we can assume there wasn’t a huge disparity.

 

1 hour ago, unreasonable said:

This historical record, taken as a whole, clearly refutes the "win the lottery theory". Losing wings under fire was a material risk for these planes. 

I have never said that braced airplanes shouldn’t lose their wings.  However, when over 95 percent of my deaths are due to either controls jamming and, more commonly, wingshedding.  If this held true for all airplanes, the argument that we are just flying too aggressively would hold up but, since there is a very clear divide between braced airplanes and cantilever airplanes, it indicates that there is likely an issue with the DM. 

[unreasonable]

32 minutes ago, Miners said:

 

 

I have never said that braced airplanes shouldn’t lose their wings.  However, when over 95 percent of my deaths are due to either controls jamming and, more commonly, wingshedding.  If this held true for all airplanes, the argument that we are just flying too aggressively would hold up but, since there is a very clear divide between braced airplanes and cantilever airplanes, it indicates that there is likely an issue with the DM. 

 

Fully agree with your first point about combat reports, which is why I went to look at MvR's systematically when the issue of wing shedding became controversial after the introduction of the  new FC DM "spars".   I was very surprised by how often it featured, perhaps because I had not thought clearly about the actual area of exposure of the wires, their vulnerability to hits, compared to that of other vulnerable areas, or the likely effects of breaking them. The other study I looked at in detail to do with the subject of "p to kill if hit in area X", related to the P47, hence wires not an issue.   

 

I do not rule out that there may be an error in the DM re wing shedding mechanics, I discussed some reservations when the new DM for FC arrived.  Jason has confirmed that the controls issue is a bug, which makes current game loss rates a poor basis for comparison, unless you just discard all of those. 

 

I am not particularly aiming my comments at you and I have no doubt  that you are reporting your "lived experience" in good faith.

 

What I find unhelpful in understanding the issue, is claiming categorically that wing shedding should be very, very rare, ignoring the area at risk by selecting a specific and very rare case designed to rule out wire hits, ignoring the well documented sample that we have showing that they were actually fairly frequent, and then justifying that by saying - and I paraphrase hyperbolically - that when MvR said "the wings came off under my fire" he actually meant "the wings did not come off under my fire",  but if he said "the pilot slumped over under my fire" he really did mean "the pilot slumped over under my fire". 

 

So unless someone has any new data, that's it from me.

[No.23_Triggers]

6 hours ago, unreasonable said:

Fully agree with your first point about combat reports, which is why I went to look at MvR's systematically when the issue of wing shedding became controversial after the introduction of the  new FC DM "spars". 

Got some data back when the DM controversy all first kicked off from the virtual 3rd PG's kill / loss reports (part of our fun in the game is putting in reports and having an in-house claiming system)...interesting compared to the MvR numbers. 

For pre-new-DM, had just 3% of kills & losses being due to wings coming off (out of 114 total aircraft downed)...from 4.006 onwards that number rocketed up to 55% (out of 92 total). 

Also went through Norman Franks' / Hal Giblin's two other "Under the Guns" books and sorted the 'fates' of downed aircraft mentioned in it - added to MvR data that brought the wing-off rate to 7% (out of 302 total, including MvR's 80). 

Currently in the process of doing the same for James McCudden's mentions of aircraft downed - not quite done with that, but so far a very quick tally (not accounting for pilot error / overspeed / etc) gives McCudden's "wing-off rate" as 23% - which I'm guessing will add up fairly closely with your figure of 15% for MvR's claims once properly divided into wings specifically shot off vs wings collapsing due to other circumstances (pilot error, death, etc) 

 

[Holtzauge]

I was not planning to post anything more but since it seems that there is still a lack of understanding about what constitutes a realistic scenario and not I think I better explain why I did the Normal distribution with the aim point in the middle of the pilot and why this is the most realistic assumption.

 

Apparently, the idea is that in many case the pilot will fire without hitting anything but thin air and miss the aircraft altogether. I totally agree on this. However, then the contention is that pilots will fire and sometimes hit air, sometimes a spar, wingtip tail or whatever and that this in the end will mean that it is more realistic to assume an even distribution. This is where I disagree: First point, let’s assume that this was true, the fire around the plane is evenly distributed: Now how far out from the center of the plane is it evenly distributed? 10, 15 or 20m? What happens at 21 m? Does it discretely go down to zero? I think most realize the fallacy of such a hypothesis.

 

Now instead let’s assume that the pilot does fire a burst of 20 rounds that all fall inside a circle of 0.5 m at 200 m range but his aim is off so he misses the plane completely. Let’s note that and let the next pilots attack each firing 20 rounds: One hits the wingtip, the next the tail, next a strut, next a wheel, next one nails the pilot etc etc. Now after a 50 such attacks we have 1000 rounds fired. Now the big question: Are those hits then evenly distributed around the airplane? I don’t think they are. I think when you take a closer look they will show a normal distribution around what each of the pilots was aiming for which is the center of the airplane. Sorry for a long winded explanation but it seem some have difficulty understanding this concept and don’t understand the post I did calculating the probability for a wire hit assuming a normal distribution with the pilot as an aim point and calls this “a special case”.

 

In addition, now that the MvR quotes are there to read, we can conclude that the wing came off in 12 cases and there is a correlation between wings coming off and planes being fired on, but, as most know, correlation is not causation: Sure, it may well be so that a certain number of those planes DID get their wing shot off, however, pilots hit and losing control, pilots overstressing their planes when being shot at are also very probable reasons and like some already have said, the same is true for cantilever planes: There are also reports of those shedding wings and it seems most agree that they should not as readily get wings shot off making overstressing a more likely cause in many cases.

 

This is the funny thing with statistics: Let’s turn it around and make a conjecture that some pilots panicked and overstressed their planes during WW1 air combat when being fired upon. Now let’s look at the statistics: Well we have MvR’s combat reports that confirms this since 15% of his victories were due to wing losses in air combat.

 

Statistics, as someone said, is the art of showing that one man with his foot in the fireplace and the other in a ice bucket is really quite comfortable……

[unreasonable]

 

2 hours ago, US93_Larner said:

Got some data back when the DM controversy all first kicked off from the virtual 3rd PG's kill / loss reports (part of our fun in the game is putting in reports and having an in-house claiming system)...interesting compared to the MvR numbers. 

For pre-new-DM, had just 3% of kills & losses being due to wings coming off (out of 114 total aircraft downed)...from 4.006 onwards that number rocketed up to 55% (out of 92 total). 

Also went through Norman Franks' / Hal Giblin's two other "Under the Guns" books and sorted the 'fates' of downed aircraft mentioned in it - added to MvR data that brought the wing-off rate to 7% (out of 302 total, including MvR's 80). 

Currently in the process of doing the same for James McCudden's mentions of aircraft downed - not quite done with that, but so far a very quick tally (not accounting for pilot error / overspeed / etc) gives McCudden's "wing-off rate" as 23% - which I'm guessing will add up fairly closely with your figure of 15% for MvR's claims once properly divided into wings specifically shot off vs wings collapsing due to other circumstances (pilot error, death, etc) 

 

   

 

Look forward to seeing the final works. I tried to go through Franks' other books but found it frustrating as, unlike in the MvR book, he does not give the combat reports verbatim. Plus you have a couple of the aces with a large proportion of dodgy claims.....  So if I were doing this whole exercise again I would sort not only by wing/non wing, but also the degree of certainty. 

 

Nevertheless, there is no reason to expect the proportion of wing losses to be particularly similar in different small samples. If it even means anything to talk about the "true" rate for all Entente losses I could well believe anything from 5-20%, certainly reducing after mid 1917 when incendiary use was creating many more quick flamers. 

 

It would also be interesting to know the game squadron's rate of wing loss over time: my hypothesis would be that the combination of black out and DM changes have also changed player behaviour.

 

The other potential issue for the MM/wing kill ratio is that the pilots could be better protected by the fuselage than they should be. They can certainly be shot through the fuselage, but in testing the Camel on the ground, killing the AI pilot requires a minimum of three hits if the shots have to pass through the fuselage and fuel tank. Nothing like as bad as RoF in this respect, but I am not sure if this is right.

[unreasonable]

9 minutes ago, Holtzauge said:

I was not planning to post anything more but since it seems that there is still a lack of understanding about what constitutes a realistic scenario and not I think I better explain why I did the Normal distribution with the aim point in the middle of the pilot and why this is the most realistic assumption.

 

Apparently, the idea is that in many case the pilot will fire without hitting anything but thin air and miss the aircraft altogether. I totally agree on this. However, then the contention is that pilots will fire and sometimes hit air, sometimes a spar, wingtip tail or whatever and that this in the end will mean that it is more realistic to assume an even distribution. This is where I disagree: First point, let’s assume that this was true, the fire around the plane is evenly distributed: Now how far out from the center of the plane is it evenly distributed? 10, 15 or 20m? What happens at 21 m? Does it discretely go down to zero? I think most realize the fallacy of such a hypothesis.

…

 

I understand your post perfectly, it simply is not applicable.  Your example really is a special case.  OR scientists - who really were statisticians - treated the fall of shot as independent and relative to the exposed area for good reason. 

 

1) The aiming errors were so great that the distribution of the shots is spread over a huge area relative to the target area. We have no WW1 gun camera footage, but US analysis of German WW2 footage draws the same conclusion. I am sure you can find it. 

 

Given that the target area is small compared to the overall spread of shots, within the target's area the distribution of hits is almost random. The gradient of density of shots could be ignored because it is statistically insignificant when examining the hits on the targets, for large groups of cases.

 

2) 0.5m at 200m is about the width of a typical tripod mounted MG beaten zone, a fixed aircraft gun might do less.  But dispersion is a small part of the effective group size in air-air combat except in very unusual cases, like yours.  Far more important is the lateral relative motion of the planes. 

 

Assuming two guns firing at 500 rpm each, the elapsed time to fire 20 rounds is 1.2 seconds. Even with a relative motion of a mere 10kph, the target moves 3.3m across your line of fire, and the shots fall in a line 0.17m apart (with additional individual variance due to dispersion). 

 

The relative motion in a dogfight would often have been much higher than 10kph. As you may have found by playing the sim, trying to adjust for this by moving your point of aim while firing usually just makes things worse.  At 30 kph relative movement the group's width is 10m wide - more than a Camel's wingspan. 

 

Your example assumes zero relative motion: ie both firer and target are flying straight and level. This - or close to this - could happen when someone managed to get right behind an unaware target cruising along, but it is no more representative of air combat in general than the case of a plane flying directly across your front at 150kph, where the 20 shot group is 50m wide. 

 

3) Does it not occur to you that players losing their wings in the game is also a function of how hard the pilots pull?

 

Readers will have to decide for themselves who has the right of it on this one.... 

 

  

[Holtzauge]

30 minutes ago, unreasonable said:

 

I understand your post perfectly, it simply is not applicable.  Your example really is a special case.  OR scientists - who really were statisticians - treated the fall of shot as independent and relative to the exposed area for good reason. 

 

1) The aiming errors were so great that the distribution of the shots is spread over a huge area relative to the target area. We have no WW1 gun camera footage, but US analysis of German WW2 footage draws the same conclusion. I am sure you can find it. 

 

Given that the target area is small compared to the overall spread of shots, within the target's area the distribution of hits is almost random. The gradient of density of shots could be ignored because it is statistically insignificant when examining the hits on the targets, for large groups of cases.

 

2) 0.5m at 200m is about the width of a typical tripod mounted MG beaten zone, a fixed aircraft gun might do less.  But dispersion is a small part of the effective group size in air-air combat except in very unusual cases, like yours.  Far more important is the lateral relative motion of the planes. 

 

Assuming two guns firing at 500 rpm each, the elapsed time to fire 20 rounds is 1.2 seconds. Even with a relative motion of a mere 10kph, the target moves 3.3m across your line of fire, and the shots fall in a line 0.17m apart (with additional individual variance due to dispersion). 

 

The relative motion in a dogfight would often have been much higher than 10kph. As you may have found by playing the sim, trying to adjust for this by moving your point of aim while firing usually just makes things worse.  At 30 kph relative movement the group's width is 10m wide - more than a Camel's wingspan. 

 

Your example assumes zero relative motion: ie both firer and target are flying straight and level. This - or close to this - could happen when someone managed to get right behind an unaware target cruising along, but it is no more representative of air combat in general than the case of a plane flying directly across your front at 150kph, where the 20 shot group is 50m wide. 

 

3) Does it not occur to you that players losing their wings in the game is also a function of how hard the pilots pull?

 

Readers will have to decide for themselves who has the right of it on this one.... 

 

  

 

Apparently you think adding relative motion will make the results change and the distribution will go towards random but they won’t: It will still end up a normal (or close too) distribution with the aim point in the middle in the end:

 

Sometimes the relative motion will be left, sometimes right, sometimes up, sometimes down and sometimes the shooter will fire too early and sometimes too late. Now make a mental picture of all those “lines” of 20 holes and how they will be distributed over the target area. You really think it will be random? You don’t think a pattern will emerge with the majority of the lines crisscrossing closer to the center of aim and progressively less the further you move out from the aim point?

 

In some sense it is good that you explained your ideas in such detail now because it is easier to understand your initial posts with the ideas about even distribution.

 

But we certainly agree on the last point: Readers will have to decide for themselves so let’s leave it at that.

[unreasonable]

Relative to the fixed area of the target, a wider normal distribution will lead to a lower variation in spread over the target, as well as fewer hits.

 

Illustratively, these two distributions are normal and they have the same means (0), different standard deviations( 2 vs 5). If the area of the target is the area between where the two lines cross, (about +/- 3) the difference in density for the red distribution is small, ~20% from edge to middle. The difference for the blue distribution is nearly 200%.

 

If you accept that only a small proportion of shots fired in air to air combat hit at all, you have to accept that the distribution was very wide relative to the target, for the reasons discussed above, which means that the variation over the target was small enough that we can ignore it when looking at average results for engagements. 

 

Edited September 12, 2021 by unreasonable

[Holtzauge]

So happy to finally see something approaching a normal distribution! But when it comes to the hit area being so wide now all of a sudden how does that tab with your first calculation when you assumed 100 rounds going through an area of 11.59 sqrm MvR style?

 

"If you are firing from behind at reasonable range, MvR style, this is orders of magnitude out.  If you can get an appreciable proportion of your shots between the wings, you have a roughly evens chance of hitting a lift wire after 100 shots between the wings."

[unreasonable]

The distributions above are purely to illustrate the mathematical necessity - which you really should acknowledge - that spreading the same number of shots over a wider normal distribution reduces the change in probability from the edge to the centre for any zone in the middle representing hits. The shots are normally distributed, but the hits are not, unless very shot is a hit.

 

We have looked at different ratios and I suspect there is some confusion so:

 

1) Hits to reference area / shots fired. 

 

Only a small proportion of shots fired hit the target reference area, however defined, so the target must take up only a small part of the total distribution of the shots for a population including, say, all WW1 scout attacks from +/- 45 degrees off the tail.  Given that, the difference in hit probability for an area on the aim point and an area 4m away is sufficiently small that it is not worth worrying about when doing these population wide estimates.

   

2) Hits to target part x (eg pilot or wires)) / hits to reference area (eg wing gap or whole plane))  We know from (1) and the shape of a flat normal distribution that it makes very little difference where in the reference area the part is located. ie we can treat the shot placement inside the reference area as random. All we need to know is the vulnerable areas.

 

3) Hits to part x / hits to part y. As in (2) based on vulnerable areas.

 

4) Losses due to hits on part x /  losses due to hits on part y       Based on (3) then factoring a p that a hit to the part causes a loss.

 

Which was where we are trying to end up, to discussing the ratio of wing losses to PKs in game or reality.

 

In the quotation, I did not assume 100 shots going into the area, that is just the calculated number required to have a ~50% p of hitting the lift wires, given the my calculated vulnerable area and the assumption that the hits to the reference area are randomly placed.  ie (1-p of hit)^100 ~= 0.5

 

Changing the reference area will change that number: but it will not change the ratio of hits to part a vs part b as long as the reference area includes both parts.

  

Edited September 13, 2021 by unreasonable

[US41_Winslow]

22 hours ago, unreasonable said:

What I find unhelpful in understanding the issue, is claiming categorically that wing shedding should be very, very rare, ignoring the area at risk by selecting a specific and very rare case designed to rule out wire hits, ignoring the well documented sample that we have showing that they were actually fairly frequent, and then justifying that by saying - and I paraphrase hyperbolically - that when MvR said "the wings came off under my fire" he actually meant "the wings did not come off under my fire",  but if he said "the pilot slumped over under my fire" he really did mean "the pilot slumped over under my fire". 

 

I’m sorry if what I said wasn’t clear.  Airplanes should lose their wings sometimes; however, it should, from the evidence that I’ve seen, be similar between braced and cantilever designs.  It should not be one of the leading causes of aircraft loss and it should not regularly occur after less than 20 hits to the wings, which it does now.  In Richthofen’s reports, they seem to rarely mention the exact number of rounds fired as well as many of them mentioning that the target was maneuvering.  Only in one of the reports does he mention that he fired a small number of rounds and the other two that mention the number of rounds fired show he needed 100 and 400 rounds before it broke up, which is nothing like what we are currently seeing on multiplayer for braced designs while the Fokkers seem to be able to take about this before going to pieces, though I will add that I don’t think I’ve ever seen a DVII go to pieces while being shot.

[unreasonable]

23 minutes ago, Miners said:

I’m sorry if what I said wasn’t clear.  Airplanes should lose their wings sometimes; however, it should, from the evidence that I’ve seen, be similar between braced and cantilever designs.  It should not be one of the leading causes of aircraft loss and it should not regularly occur after less than 20 hits to the wings, which it does now.  In Richthofen’s reports, they seem to rarely mention the exact number of rounds fired as well as many of them mentioning that the target was maneuvering.  Only in one of the reports does he mention that he fired a small number of rounds and the other two that mention the number of rounds fired show he needed 100 and 400 rounds before it broke up, which is nothing like what we are currently seeing on multiplayer for braced designs while the Fokkers seem to be able to take about this before going to pieces, though I will add that I don’t think I’ve ever seen a DVII go to pieces while being shot.

 

TLDR - It is possible you are correct, such is our uncertainty about many key variables, but no evidence I have seen substantiates your opinion.

 

Your first point is that wing loss should similar between braced and cantilever designs. Why? If the no damage g limits are different, which in some cases they are, why should a bullet hit erase that difference? Even if that is true, loss rates will only be similar between MP and WW1 if pilots in MP are flying like pilots in WW1, which is open to doubt. 

 

Comparing the two when undamaged: wing loss is then a function of gs pulled and the g limit. You could create a braced design with an equal g limit to a D.VII if you made the spars thick enough and put in enough flying wires, but that has other disadvantages: AFAIK it did not happen. The g limits we have seen mentioned for D.VIIs are noticeably higher than for SE5s and Camels, not sure about SPADS, (Pflaz I agree seems to be just wrong). There is a table somewhere. 

 

The second case is what happens when the designs take damage leading to a reduction in the g limit, for instance a single hit to either main spar (both types) or flying wire.  The problem for the braced design, IMHO, is that a hit to either is more likely to cause an g reduction of x amount in the case of the braced design, since the two areas that can be hit are each much thinner. 

 

Your next point is that "it should not regularly occur after less than 20 hits to the wings". This is not clear: define "regularly occur". If you mean that a low number of hits cannot break a flying wire I disagree. If a single hit can do this, which I think is undisputed - then they will occur just as regularly with the first hit as the second, and so on. In the case of spars you may need cumulative damage from hits close together, so the number of hits is important, but then we have to worry about the relative area and vulnerability of large box spars vs smaller solid spars, which AFAIK is still unsettled. The game seems to be modelling based on thickness, but again so far no actual evidence presented that the outcomes are incorrect. 

 

On the MvR ammunition usage, the specific reports I listed in this thread only have a couple of ammo reports - but he reported ammo use in 34 claims, and there is no reason to believe that the "wing loss" sample" was systematically different from the others. See table below.  This is rounds used, not hits!  We do not know the number of hits but it maybe quite low. My prior would be 10-20%, but in his case it might have been a little higher.

 

 The actual number of hits is, however, an overestimate of the number needed: since he not only shot continuously at planes already descending to crash if they were still in control, but even in a single burst of 20 rounds, the round that caused the crash was as likely to be the 1st as the 20th. 

 

 Note the high numbers were mostly against 2-seaters, many of which were pushers, using FMJ ammunition. The later ones were against scouts, using a mix of ammunition including HE/I -  perhaps his positioning and shooting improved over time as well. What we do not know is how many rounds he used against targets he either did not hit or did not down, and I am not sure how often that happened: I think not very often, but Franks' book only gives reports for his claims.

 

 

 

 

 

 

 

 

 

 

Edited September 13, 2021 by unreasonable

[1PL-Husar-1Esk]

4 hours ago, Miners said:

I’m sorry if what I said wasn’t clear.  Airplanes should lose their wings sometimes; however, it should, from the evidence that I’ve seen, be similar between braced and cantilever designs.  It should not be one of the leading causes of aircraft loss and it should not regularly occur after less than 20 hits to the wings, which it does now.  In Richthofen’s reports, they seem to rarely mention the exact number of rounds fired as well as many of them mentioning that the target was maneuvering.  Only in one of the reports does he mention that he fired a small number of rounds and the other two that mention the number of rounds fired show he needed 100 and 400 rounds before it broke up, which is nothing like what we are currently seeing on multiplayer for braced designs while the Fokkers seem to be able to take about this before going to pieces, though I will add that I don’t think I’ve ever seen a DVII go to pieces while being shot.

DVII is flying tank,  you need to kill pilot to make it go down in reasonable time, the current engine damage model make it fly much longer than in ROF. You can snap D7f wings only in high speed dive when you pull up abruptly . British planes  lost wings without need to build that speed. Just need to pull horizontal turn in moderate G load when wing is damaged. 

Edited September 13, 2021 by 1PL-Husar-1Esk

[No.23_Triggers]

5 hours ago, Miners said:

...which is nothing like what we are currently seeing on multiplayer for braced designs while the Fokkers seem to be able to take about this before going to pieces...

Just feel the need to quickly point out that the Pfalz D.III and Bristol - both braced designs - are also pretty tanky when it comes to wing shedding. Not the SuperTank that the D.VII is, but still a lot better than "Average" for other braced designs...

 

[HagarTheHorrible]

19 hours ago, Holtzauge said:

I was not planning to post anything more but since it seems that there is still a lack of understanding about what constitutes a realistic scenario and not I think I better explain why I did the Normal distribution with the aim point in the middle of the pilot and why this is the most realistic assumption.

 

Apparently, the idea is that in many case the pilot will fire without hitting anything but thin air and miss the aircraft altogether. I totally agree on this. However, then the contention is that pilots will fire and sometimes hit air, sometimes a spar, wingtip tail or whatever and that this in the end will mean that it is more realistic to assume an even distribution. This is where I disagree: First point, let’s assume that this was true, the fire around the plane is evenly distributed: Now how far out from the center of the plane is it evenly distributed? 10, 15 or 20m? What happens at 21 m? Does it discretely go down to zero? I think most realize the fallacy of such a hypothesis.

 

Now instead let’s assume that the pilot does fire a burst of 20 rounds that all fall inside a circle of 0.5 m at 200 m range but his aim is off so he misses the plane completely. Let’s note that and let the next pilots attack each firing 20 rounds: One hits the wingtip, the next the tail, next a strut, next a wheel, next one nails the pilot etc etc. Now after a 50 such attacks we have 1000 rounds fired. Now the big question: Are those hits then evenly distributed around the airplane? I don’t think they are. I think when you take a closer look they will show a normal distribution around what each of the pilots was aiming for which is the center of the airplane. Sorry for a long winded explanation but it seem some have difficulty understanding this concept and don’t understand the post I did calculating the probability for a wire hit assuming a normal distribution with the pilot as an aim point and calls this “a special case”.

 

In addition, now that the MvR quotes are there to read, we can conclude that the wing came off in 12 cases and there is a correlation between wings coming off and planes being fired on, but, as most know, correlation is not causation: Sure, it may well be so that a certain number of those planes DID get their wing shot off, however, pilots hit and losing control, pilots overstressing their planes when being shot at are also very probable reasons and like some already have said, the same is true for cantilever planes: There are also reports of those shedding wings and it seems most agree that they should not as readily get wings shot off making overstressing a more likely cause in many cases.

 

This is the funny thing with statistics: Let’s turn it around and make a conjecture that some pilots panicked and overstressed their planes during WW1 air combat when being fired upon. Now let’s look at the statistics: Well we have MvR’s combat reports that confirms this since 15% of his victories were due to wing losses in air combat.

 

Statistics, as someone said, is the art of showing that one man with his foot in the fireplace and the other in a ice bucket is really quite comfortable……

I’m glad you wrote it, saved me a lot of time.

 

 It is ,I think, also important to note, that, in combat reports, it would have been advantageous to report wings falling off, a definite kill, rather than a vague “out of control”.  It is equally important to stress though, that just because the wings came off, it doesn’t necessarily mean that they came off due to combat damage, an aircraft that is out of control, due to any number of factors, might quickly exceed it’s design limits, and the earlier in the war, and the older the design, the more that might have happened.  What I’m trying to say, is that just because a combat reports a kill, because the wings came off, it doesn’t necessarily imply that is the reason for the kill, a bit like saying “The aircraft crashed into the ground”, ergo, the reason the aircraft crashed was because it hit the ground.  Certainly, crashing into the ground is conclusive proof of finality, as it is in all air combats, but there are usually any number of reasons leading to that point, some of which might be less obvious than others.

 

Essentially, if old combat reports are to be considered as proof of aircraft fragility, in combat, and some form of percentage ratio derived, then it also has to consider what percentage of those aircraft lost wings in a death dive, as conformation of death, rather than disintegrating due to bullets fired in the heat of battle.

[No.23_Starling]

On 9/12/2021 at 4:12 AM, unreasonable said:

Extracted from MvR reports as per translation in Franks et al "Under the wings of the Red Baron"

 

Rep 23 "After the 400th shot, plane lost a wing while making a curve"

Rep 24 "The wings of the plane I attacked came off"

Rep 28 "After a short fight my opponent lost both wings and fell"

Rep 39 "After a very few shots the plane broke to pieces and fell near Vimy"

Rep 47 "...shot at him from the closest range until his left wing came off."

Rep 49 "The plane I had singled out broke to pieces while curving."

Rep 51 "After a short fight my adversary's plane lost it's wings".

Rep 55 "I managed to break one of the reconnaissance planes with my fire. The fuselage fell with the inmates..."

Rep 56 "We spotted an enemy artillery flyer whose wings broke off in my machine gun fire"

Rep 66 " I shot down the last opponent, a Bristol fighter. He lost his wings, and Lt Gussmann brought him down".

Rep 67 "Under my machine gun fire, both wings broke away from the aeroplane in the air".

Rep 77 "I shot at an enemy plane some 200m away. After I had fired 100 shots. the enemy plane broke apart".

I agree that all these seem likely candidates for wings being shot off, although I’m still not convinced by “fell to pieces” as always meaning main plane failure. That could equally be a tailplane coming apart etc.

 

On the cantilever vs braced ideally we need combat reports of entente planes vs Fokkers, to get a similar toll to the one you have for MvR. 
 

I can find isolated examples, eg Sous lt Louis Risacher quoted in Guttman’s book Spad XIII vs Fokker DVII p53: “…one German passed to my left with the Spad on his tail. The Spad shot him and he fell to pieces… At the very same moment a second Spad - I knew it was an American aeroplane by his cockades - coming in at full speed took on another Hun and shot him to pieces. I said ‘God Save America!’ and at that moment I put speed to my old aeroplane and took a third DVII in a loop. He looked behind at me and then fell to pieces.”

 

You could argue this is 3 cantilever planes suffering structural failure in a single fight as a direct result of combat damage, not falling apart (wings coming off) in a power dive or overstressing the airframe.

 

Quoting MvR is great vs braced planes but you need to give data on cantilever planes suffering damage. That MvR data only gives one side of the story.

 

Perhaps Rickenbacker is a good source on this as most of his combat was vs DVIIs?

 

Mr Larner also makes a good point about some of the braced planes we have right now - the Diiia is currently almost as tough as the DVII.

Edited September 13, 2021 by US93_Rummell
Grammar

[J2_Trupobaw]

A plan suffering wing failure must be witnessed and identified as such by someone else to make it to the combat report. There is much confusion in combat reports (like the guy from No.70 seeing friendlies do down in flames and claiming them as victories on enemy (I fired at the Hun, saw a plane go down, so I think I got him. And oh, Willson never came back, I wonder what happened to him...)). 

If plane that disintegrated while trying to dive away from the fight while everyones attention was higher was likely not mentioned in anyones combat report, except as mysterious MIA.

 

Edited September 13, 2021 by J2_Trupobaw

[HagarTheHorrible]

If the game,FC, is anything like reality, or even a simulation of reality, then there should be a clear correlation of outcomes.

 

The D VII was easy to fly and, in every way, the equal of it’s opponents, if not slightly better. With it’s introduction over the front in 1918 there should, in theory, be an obvious trend, as they became more numerous, of a kill to death ratio compared to it’s opponents.  If a statistical study is going to be done then surely the first and most obvious comparison is instances of D VII combat losses to victory’s,  and where noted the cause of destruction, baring in mind my caveats about destruction from combat or, in a death plunge, as a result of combat.

 

As I’ve mentioned elsewhere, I don’t think there was an obvious observed correlation in design difference and durability, and I think history of post war development bares out this hypothesis.  That’s not even to argue that cantilever wings are more damage resistant than stick and wire, they might well be, it’s just that it wasn’t seen, at the time, as an arbiter of success or failure in combat.

 

A beneficial exercise for observing combat damage, as pertained at the time, and how the aircraft were flown, might be how aircraft were brought down when staffing  ground targets (lots of guns, lots of bullets.  The chances are they were hit far more often than in typical air to air combat and it is quit possible that the wings and wires were far more vulnerable to being hit, if only by dint of the volume of fire.

[HagarTheHorrible]

Can anyone even begin to imagine, even if we discount wounds to the pilot and just think about aircraft structural integrity , the same, or similar scenario, happening in FC, as happened to George Barker VC ?

 

On 27th October 1918, he left La Targette in his Snipe returning it to Hounslow, when he crossed enemy lines at 21,000 feet above the Forêt de Mormal. He attacked an enemy Rumpler two-seater which broke up, its crew escaping by parachute; {the aircraft was of FAA 227, Observer Lt Oskar Wattenburg killed}. By his own admission, he was careless and was bounced by a formation of Fokker D.VIIs of Jagdgruppe 12, consisting of Jasta 24 and Jasta 44. In a descending battle against 15 or more enemy machines, Barker was wounded three times in the legs, then his left elbow was blown away, yet he managed to control his Snipe and shoot down or drive down three more enemy aircraft {Two German pilot casualties were Lt. Hinky of Jasta 44, wounded; and Vfw. Alfons Schymik of Jasta 24, killed}. The dogfight took place immediately above the lines of the Canadian Corps. Severely wounded and bleeding profusely, Barker force landed inside Allied lines, his life being saved by the men of an RAF Kite Balloon Section, who transported him to a field dressing station.

 

Interestingly, his aircraft is now preserved in Canada.  I think we can be fairly certain that the wings of a Snipe would simply jump off themselves if it was attacked by 15 D VII’s, not to mention the others, who also had a pop at it on the way down. If it was FC.

 

I know, one egg doesn’t make an omelette, or even a bad omelette, but I think it is extremely instructive that an aircraft of dubious structural integrity, to combat damage (at least according to FC) can do all the things FC says it can’t against the very (several) aircraft that should be so superior as to be almost indestructible.  

 

 

Edited September 13, 2021 by HagarTheHorrible

[unreasonable]

MvR's reports have been shown to be fairly accurate, others less so. (Göring!!) You would expect that firing a belt including HE/I would give different results than pure FMJ, and the MvR reports strongly support that, so any reports that you have must be located in time. The more you divide up the evidence the greater the specificity, but the smaller the sample and the greater the role of chance. There were no gun cameras and no systematic WW2 style statistical operations research. Every airforce had different claim standards, and there is no way of checking most Entente claims because the German records are so limited (I think).  I short, if we think we can solve all of this as historians, we are f&#@ed.

 

That is why I have been most specific here in rejecting the claim that dewinging braced aeroplanes should be very very rare or like winning the lottery: I can prove, at least to my own satisfaction, that this is nonsense. I think that the chances of dewinging braced planes is higher than for dewinging cantilever planes, based on the limited evidence available, but not beyond a reasonable doubt. We do not have a P47 style study of people systematically shooting Camels and DVIIs to see what breaks. 

 

On the "causality" issue that has been raised:  in the MvR reports, there is no "the pilot slumped over in the cockpit while his engine stopped as his plane began to burn with the wings falling off" but there are some cases where there was almost certainly more than one sortie ending event by the time MvR stopped firing.  But if you are trying to estimate the p of n hits (or shots) causing wing loss that does not matter.  Additionally, if you lose your wings, you are dead. As a conditional causality, that seems solid to me, so I find that whole line of argument to be specious.  You can analyse reports and give more than one cause of death if you want: then sortie ending events will add up to more than the number of kills, but that is fine. It is true that, wings coming off being sometimes more obvious that shooting the pilot, the number of PK events might be under-reported: but the p of a hit being a PK is comparatively easy to approximate. Two analysts might disagree on a p number, but not by anything close to an order of magnitude.

 

 MvR describes fuselages breaking up on two occasions IIRC: putting all the 4 generic "broke up" reports in the fuselage box when there are 8 explicit wing mentions and only 2 fuselage mentions is biased. Put 1 of the 4 generics in with the fuselages if you want to sort pro-rata: it changes the overall picture very little.     

 

I agree that the wires/cantilever decision at the time was driven by issues other than (or at least in addition to) damage resistance, but that is not evidence that the damage resistance was the same, or different, either way, which is what we are trying to determine. It is irrelevant. 

 

If you look at the vulnerable areas from a 25 degree rear angle and a 25 degree front angle I would take those as my expectations for the ratios of cause of losses in air-air and ground-air (small arms) scenarios respectively. More engine stopped kills from the front, more PK/flamers from the rear. But as Hagar says, the sheer volume of fire being put up would probably be the most important factor in determining losses/sorties in ground attack.

 

 On the Barker VC story: he was shot down, what is the issue? I have been shot down by the AI when fighting several enemy in a Camel, and it was not because my wings came off, but because my elevator jammed: which is another story. So yes, I can imagine this happening in FC. 

Edited September 13, 2021 by unreasonable

[ZachariasX]

8 hours ago, unreasonable said:

If a single hit can do this, which I think is undisputed

So, you are still saying that a finger thick hardened steel rod that exhibits the shape of a wedge toward a pointy bullet does come apart that easy? Especially when the tension on that rod is at least an order of magnitude below what most of the wire section can take?

 

 

These are flying wires. The yellow stick to connect them is about a thick as an index finger. Now, when you look at the "wires", you really think that resistance to gunfire is evenly distributed? Those "wires" compare to what is mounted on the SE5a. See the diagram I posted before in this thread. those wires must be shaped that way and be of considerable overstrengt to keep them from fluttering. That would be draggy and dangerous.

[unreasonable]

How do you get from a "single hit can do this (break a flying wire enough for the wing to fail)" to  "think that resistance to gunfire is evenly distributed?"

I really wish people would read what is said rather than projecting. It reminds me of the famous Peterson interview..."So you are saying that women are lobsters?"   

 

Obviously there is uneven resistance: just as a bullet to the heart is more likely to cause a PK than a bullet to the foot; but for practical purposes we treat hits to pilot as pA, and what happens when pilot hit (PK, notPK) as pB.   Similarly there is a pA to hit flying wire, and a pB that the hit breaks it. 

 

On the wires: yes I do think that a single hit can break those flying wires. If you can demonstrate otherwise, I will alter my beliefs. Similarly I do think that non-breaking hits will reduce maximum g load. The pA to hit is relatively easy to calculate, the pB is the problem. 

 

Your picture is one type of flying wires, but appears unrepresentative of what was on WW1 biplanes. I do not have one handy, and many of the photos are not too clear, but I cannot find a case with anything that looks like those solid aerodynamic mini struts. Actually, I would not be surprised if those modern wires are less resistant to bullet damage than WW1 style wires. 

 

The Sopwiths appeared to have two thin wound wires as in FC:

 

 

Also Strutter:

 

This SPAD XIII replica also has paired wound wires as does the FC model.

 

 

This Vintage Aviator SE has, I think - cannot find a photo close up - two sets of paired flying wires each side, perhaps you could get away with losing the inner set.

 

 

Edited September 13, 2021 by unreasonable

[HagarTheHorrible]

All this talk of wire is a distraction, if an otherwise interesting debate.  Wires, or the damage to them isn’t modelled in FC, so it’s all a bit academic discussing it.  How can I be so sure ?  Well, if they were, then there would be an obvious advantage to similarly rigged aircraft that had more wires, ergo, the SE5a with it’s 16 flying wires, compared to a poultry 4 on aircraft like the Albatros or Pflaz D III, would be, in an order of magnitude, less likely to become a combat casualty from a collapse due to a flying wire being damaged to the point of failure.  Also, I can’t repeat often enough, the two Allied designs that were to replace the Multi wired Camel and SE5a used single flying wires, proof if proof was needed, that combat losses because of severed flying wires were “deemed” to be irrelevant compared to other causes.

 

[unreasonable]

Unless you have a document stating the reason for the decision to go from twinned wires to singles, this is idle speculation. The decision could also be interpreted as a belief that having twinned wires was no more or less effective than having only one stronger wire, but cheaper and easier to maintain. You could as well argue that the decision not to armour the fuel tank was because the proportion of  losses to flamers was deemed irrelevant. They might have been irrelevant to the decision, but it does not follow that they were a small proportion of of total.  

 

If you want to know the proportion of losses due to some cause, count them, do not pretend that design decisions will tell you.

[No.23_Starling]

2 hours ago, unreasonable said:

MvR's reports have been shown to be fairly accurate, others less so. (Göring!!) You would expect that firing a belt including HE/I would give different results than pure FMJ, and the MvR reports strongly support that, so any reports that you have must be located in time. The more you divide up the evidence the greater the specificity, but the smaller the sample and the greater the role of chance. There were no gun cameras and no systematic WW2 style statistical operations research. Every airforce had different claim standards, and there is no way of checking most Entente claims because the German records are so limited (I think).  I short, if we think we can solve all of this as historians, we are f&#@ed.

 

That is why I have been most specific here in rejecting the claim that dewinging braced aeroplanes should be very very rare or like winning the lottery: I can prove, at least to my own satisfaction, that this is nonsense. I think that the chances of dewinging braced planes is higher than for dewinging cantilever planes, based on the limited evidence available, but not beyond a reasonable doubt. We do not have a P47 style study of people systematically shooting Camels and DVIIs to see what breaks. 

 

On the "causality" issue that has been raised:  in the MvR reports, there is no "the pilot slumped over in the cockpit while his engine stopped as his plane began to burn with the wings falling off" but there are some cases where there was almost certainly more than one sortie ending event by the time MvR stopped firing.  But if you are trying to estimate the p of n hits (or shots) causing wing loss that does not matter.  Additionally, if you lose your wings, you are dead. As a conditional causality, that seems solid to me, so I find that whole line of argument to be specious.  You can analyse reports and give more than one cause of death if you want: then sortie ending events will add up to more than the number of kills, but that is fine. It is true that, wings coming off being sometimes more obvious that shooting the pilot, the number of PK events might be under-reported: but the p of a hit being a PK is comparatively easy to approximate. Two analysts might disagree on a p number, but not by anything close to an order of magnitude.

 

 MvR describes fuselages breaking up on two occasions IIRC: putting all the 4 generic "broke up" reports in the fuselage box when there are 8 explicit wing mentions and only 2 fuselage mentions is biased. Put 1 of the 4 generics in with the fuselages if you want to sort pro-rata: it changes the overall picture very little.     

 

I agree that the wires/cantilever decision at the time was driven by issues other than (or at least in addition to) damage resistance, but that is not evidence that the damage resistance was the same, or different, either way, which is what we are trying to determine. It is irrelevant. 

 

If you look at the vulnerable areas from a 25 degree rear angle and a 25 degree front angle I would take those as my expectations for the ratios of cause of losses in air-air and ground-air (small arms) scenarios respectively. More engine stopped kills from the front, more PK/flamers from the rear. But as Hagar says, the sheer volume of fire being put up would probably be the most important factor in determining losses/sorties in ground attack.

 

 On the Barker VC story: he was shot down, what is the issue? I have been shot down by the AI when fighting several enemy in a Camel, and it was not because my wings came off, but because my elevator jammed: which is another story. So yes, I can imagine this happening in FC. 

That’s my point. Where’s the evidence of braced wings failing more than cantilever? If you’re going to take MvR’s accounts as quantitive evidence you MUST find accounts of entente pilots attacking Fokkers and tally up the results to the best that you can. Assuaging your own thoughts suggests an implicit bias. All I see from you by quoting MvR or other German aces is a one-sided dataset. I’ve presented at least one example of multiple Fokkers supposedly losing their wings in the same engagement - by your “fell to pieces” definition - as a starter. I’ll read back through Rickenbacker later.

 

From what Holtze has presented as an aerospace engineer I can’t see substantially more (I’m not saying they’re equal!!) advantages to damage resistance from cantilever designs - I can definitely see the air resistance advantages and simplification from rigging; plus no critical zones of insta doom - which would justify the extreme damage resistance of the Fokkers currently in FC.

 

I do agree with you that wing removal was not super rare, but far rarer than fires (which could in turn burn through the wings) and pilot wounds.

 

To quote MvR again: meat or metal.

[Hellequin13]

So much to unpack here...

 

In order to avoid the worlds longest post I'll just start with a couple quick points. First being about the bracing wires: there currently is only one company in all the world that makes them, and they are the original company that invented them back in 1909. Bruntons of Scotland. Their bracing wires are flattened steel rods (not round wires). A flattened rod has 1/7 the drag of a round wire, and the buggers are strong, taking 6,900 pounds of force to snap. I don't know how resistant to a bullet hit they are, but that flattened shape is going to make them a bit harder to hit, and break, depending on the angle of the hit.

 

And yes Hagar, the wires are not currently modeled in game, but they really should be. The current hit boxes are confined only to the wing profile. Spars are modeled through math, and struts and wires are not really modeled, their damage being more cosmetic in nature. But again, they should be modeled, so lets not discard them from the discussion here.

 

I think one of the failings of our current DM is the fact that the wires are not being modeled properly. There is some genius in the methods used for modeling the spars, ie the hit probabilities, and a similar system should be used for bracing wires, struts, control wires, and on WWII craft systems fuel lines, hydraulic lines, etc. While hit boxes for all these systems would be ideal, that is too much to be modeled by the engine. Whereas, using the hit box where a bracing wire is attached to check for a hit on said bracing wire, we can model the damage and get cumulative loss to the structural integrity. The same could be done for modeling failure of struts.

 

Next point, I think it is erroneous to look at MvR's records and think it is representative of wing failure as a whole. MvR is at the extreme right of the bell curve that represents all of WWI combat pilots. His 15% wing failure rate demonstrates what can happen, but not necessarily what did. By all accounts he was an excellent pilot, as well as tactician, but also quite the marksman and that is most important here. Unlike the average WWI pilot, his shots were more likely to be well placed. And when you consider where those rounds would be placed (meat or metal) it ups the probability that he would also be hitting where a wing failure is most likely to be generated: lower wing root/connection point of the flying wires.

 

Alright, going to take a break, get some food in me, and then come back and dig into the structural design aspects...

[AndyJWest]

It strikes me that this endless debate about the relative proportion of aircraft kills that are cause by hitting bracing wires may be based on a rather questionable assumption: that the graphical 'wings coming off' modelling is intended to always be indicating bracing wire hits, rather than 'hits causing catastrophic structural failure' in general. Gameplay-wise, a kill is a kill, and it isn't really worth the extra work involved for the developers to try to model too many different modes of failure. Much the same points have been made about damage modelling in the WWII stuff, and I think it is accepted that what appears to happen visually there doesn't always match what the 'physics damage model' says has done - hence aircraft sometimes still being flyable with more of a wing missing than seems plausible. 

 

To my mind, the only thing that matters, beyond aesthetics, is that once allowance is made for shooting accuracy (most competent FC players are likely better than the average real-world pilot, since they have the luxury of not getting killed while learning to shoot), is that the game gets the proportions of PKs, structural failures, fires, and dead-engine hits about right. If it does that, does it really matter how exactly structural failure is modelled? I don't think so. And if it is getting the proportions wrong, arguments about how likely a bullet is to break a steel wire really aren't going to resolve the issue, since we don't seem to know for sure if that is being modelled at all.

[No.23_Starling]

On 9/12/2021 at 4:12 AM, unreasonable said:

Extracted from MvR reports as per translation in Franks et al "Under the wings of the Red Baron"

 

Rep 23 "After the 400th shot, plane lost a wing while making a curve"

Rep 24 "The wings of the plane I attacked came off"

Rep 28 "After a short fight my opponent lost both wings and fell"

Rep 39 "After a very few shots the plane broke to pieces and fell near Vimy"

Rep 47 "...shot at him from the closest range until his left wing came off."

Rep 49 "The plane I had singled out broke to pieces while curving."

Rep 51 "After a short fight my adversary's plane lost it's wings".

Rep 55 "I managed to break one of the reconnaissance planes with my fire. The fuselage fell with the inmates..."

Rep 56 "We spotted an enemy artillery flyer whose wings broke off in my machine gun fire"

Rep 66 " I shot down the last opponent, a Bristol fighter. He lost his wings, and Lt Gussmann brought him down".

Rep 67 "Under my machine gun fire, both wings broke away from the aeroplane in the air".

Rep 77 "I shot at an enemy plane some 200m away. After I had fired 100 shots. the enemy plane broke apart".

 

That is 12 cases. I am counting 80 as the total, as per Franks' assessment. 15%

 

This excludes a number of others in which he uses words indicating  that the wings came off while the plane was diving OOC. In most of the quotes above it is clear that the wing loss or breakage occurred while he was attacking, not after he had stopped because he knew the enemy was finished.

 

You can quibble about any specific claim - MvR had to interpret what he thought we saw, and we have to interpret what he wrote - but MvR's claims have been analysed more than anyone else and stand up very well. 8 of these 12 explicitly say "wings": that is still 10% 

 

Analysis of his claims by historians show MvR to be a remarkably accurate observer, given the circumstances, with very few claims in serious doubt. In my view it is highly improbable that when MvR said "broke to pieces" he actually meant "did not break to pieces", although I agree that may be an appropriate reaction to a completely unverified claim by an excitable no-name pilot.

 

This historical record, taken as a whole, clearly refutes the "win the lottery theory". Losing wings under fire was a material risk for these planes. 

 

 

 

Also, make sure to read all the accompanying text from the Franks book. With kill 77, MvR claimed an SE5 which came apart - Franks can’t tie a British Se5 loss at the time and place and instead thinks it was a Camel which crash landed with the pilot surviving, and going on to gain several Fokker kills later in the war. I can’t see a wingless Camel having a pilot alive afterwards.

 

MvR’s accounts have other errors - kill no15 wasn’t a kill at all and should really be stricken from the history books. The Baron claims he saw him crash, when of course the pilot McCudden survived and went on to be one of the ace of aces.

 

 

[ZachariasX]

1 hour ago, unreasonable said:

Unless you have a document stating the reason for the decision to go from twinned wires to singles, this is idle speculation.

The reason for four going to singles is that twinned wires tend to vibrate or flutter harmonically in the slipstream. This causes drag and the load cycles will wear the anchor points. Flying wires that are profiled and directionally fixed (you have that stick in the middle to connect them) they don‘t move or vibrate at all in normal flight.

 

Nobody contests that a wire cannot be cut by a single round. It‘s just not that likely. Especially since the cables (where used) are redundant. You need to snap two, not one. And that I‘d say is extremely unlikely.

 

35 minutes ago, AndyJWest said:

If it does that, does it really matter how exactly structural failure is modelled?

This. The game doesn‘t simulate any wires. All the DM at best can do is produce what we deem a plausible result from out actions.

 

Actually, that bottom line is all we think we have. If you know (some do that here) that 15% of braced wing planes must fall apart, while cantilever ones do that only by 5% (or whatever), then you make the DM do so. If there even was a hitbox for wires, you can even calculate what must be the probability required of a bracing failure per shot fired without having to know anything about the actual bracing at all. That is actually what I did in my example above in the thread.

[unreasonable]

2 hours ago, US93_Rummell said:

That’s my point. Where’s the evidence of braced wings failing more than cantilever? If you’re going to take MvR’s accounts as quantitive evidence you MUST find accounts of entente pilots attacking Fokkers and tally up the results to the best that you can. Assuaging your own thoughts suggests an implicit bias. All I see from you by quoting MvR or other German aces is a one-sided dataset. I’ve presented at least one example of multiple Fokkers supposedly losing their wings in the same engagement - by your “fell to pieces” definition - as a starter. I’ll read back through Rickenbacker later.

 

From what Holtze has presented as an aerospace engineer I can’t see substantially more (I’m not saying they’re equal!!) advantages to damage resistance from cantilever designs - I can definitely see the air resistance advantages and simplification from rigging; plus no critical zones of insta doom - which would justify the extreme damage resistance of the Fokkers currently in FC.

 

I do agree with you that wing removal was not super rare, but far rarer than fires (which could in turn burn through the wings) and pilot wounds.

 

To quote MvR again: meat or metal.

 

Some observations, not all specifically in response to your comments:

 

1) I do not HAVE to do anything. As I have said, I do not know for sure if braced designs were more prone to wing shedding than cantilever, partly because a quantitative analysis of the losses of the cantilever planes is not something I am in a position to do, and no-one else has yet presented one here. TBH I do not even care, that is not why I am posting here. It is to criticise the arguments claiming that wing loss is super rare. 

 

2) So I am glad we can agree that wing loss was not super rare.  Also that fires were certainly much more common late war, and if you get more losses from fires you will get fewer from wing loss, so no argument there. 

3) Based on MvR 15%, is my prior probability for a braced aeroplane shot down by a scout to have wing loss as (one of) cause of death. That is not a bias, it is a prior probability, a very different beast, that I would apply if I did not know what the time period or target type might be except that it was a WW1 braced aircraft. 

 

(Yes I have read the whole book and the total wing kills is uncertain - but you also have to agree, I think, that given how well most of his claims check out, compared to those of other pilots in the other "Under the Guns books", that he was an unusually reliable witness).  

 

If you want to have an estimate for a pilot from the era of incendiary use, I would only take the late war numbers, 3/22 = ~14% but small sample size is a problem.

 

4) My opinion is that there is a small but real probability of hitting a bracing wire (or connecting point) in such a way that it breaks leading to potential wing loss ie reduction in g limit. (Lee Gould thought so too, see extract earlier). What is that probability? Hard to say, but if you take the overall area of the wires/connections you can compare to the exposed area of the M&M, and play about with the probability that a hit breaks to derive p to kill by wire hit.

 

Cantilever planes simply do not have that specific vulnerability, so there is no reason in principle to assume that they must have the same damage resistance. 

 

Whatever that probability is, it has to be broadly consistent with what we know about wing losses from reports. If a one in a million chance is leading to 5% of losses, either the 5% is wrong, the one in a million is wrong: or people were firing a lot more ammunition that you thought!  edit: that is me mixing up the ratios as well: better to say "or the chances of killing by all other methods combined was in the order of 1/53,000"

 

5) If your wings come off, your sortie is over: that is about as causal as it gets. If you get shot in the head and in the heart simultaneously, it would be ridiculous to claim that the heart shot should not count as a cause of death. 

 

Edited September 13, 2021 by unreasonable

[HagarTheHorrible]

I know I keep banging on about it, but in the absence of hard data, we really have to rely on, as much as anything, what the aircraft designers and engineers actually did to build aircraft that were up to the rigours of combat flying and how those designs evolved.  The first place to start, in my opinion, would be to ask “What was it about the Spad VII wing that so impressed the Pflaz designers that they dumped the, according to FC, brilliant D IIIa wing”, it certainly wasn’t because it was easier to rig and I’m pretty sure it wasn’t any less of a drag. So what was it ?   (Hint-  What are little blue pills good for ?).
 

Although FC doesn’t appear to model the effects of pure speed (at least not on the Albatros ?), which is a major pity (Generally.  Not just for Alby drivers) I suspect that the major issue with WOODEN cantilever spar Scouts was wing flex, as in rigidity, something that wasn’t fully resolved until the advent of metal spars. Sure, they could have fixed it, with bracing wires, but that would have rather defeated the point, wouldn’t it ?

Edited September 13, 2021 by HagarTheHorrible

[ZachariasX]

10 minutes ago, unreasonable said:

Cantilever planes simply do not have that specific vulnerability

All you need to sever is the top layer of the box spar and you affect the whole structure. As that part of the beam is AFAIK made from spruce, I think it should be somewhat robust though.

 

12 minutes ago, unreasonable said:

Whatever that probability is, it has to be broadly consistent with what we know about wing losses from reports.

Absolutely.

[HagarTheHorrible]

14 minutes ago, ZachariasX said:

 All the DM at best can do is produce what we deem a plausible result from our actions.

 

This, THIS, is the whole nub, from start to finish, succinct and devoid of all crap.  The only bit I would add is “and have fun while doing it”

[Hellequin13]

Alright, full to contentment, so back at it...

 

I am pretty sure we are all well versed on the basics of the braced design here, so I don't think I need to go into too much detail, save for one caveat: the assumption that the loss of a flying wire would be catastrophic. Yes, the flying wires are the most critical bracing wire to the wing structure (in flight) as it works against the upward flex in the wing generated by lift. But the point of any of the bracing wires is to stiffen the structure, as in brace it against flex and twisting moments. The bracing wires do not carry the full load. That still falls the the spars.

 

If you were to remove all the flying wires, the struts still tie the upper wing to the lower wing. The lower wings function is primarily for structural support, creating a box structure via the struts. The lower wing does not generate as much lift as the upper wing (due to interference in air flow from the upper wing) and therefore has less of an upward flex from lift. With out the bracing wires, you still have some degree of resistance to flex/twist albeit at much lower loads. In other words, the loss of a single flying wire would not lead to immediate wing loss, rather a reduction in wing strength. The subsequent wing loss would be due to over stressing that wing, be it from increased G-load due to maneuvers, or simply from increased lift generation due to attempting to dive away (increased speed).

 

As for the cantilevered design, this varies from Dr.1/D.VII (also the D.VI, but that isn't relevant to FC), and D.VIII. With both the DR.1 and the D.VII the cantilever design is only partial, being confined to the leading edge/main spar region. With these two planes, there is a plywood structure that goes from the top of the main spar, curving across the front of the wing and ending at the bottom of the main spar. The structural integrity of the design is dependent on the 'skin' between ribs and spar, creating a box (or series of boxes) that resists compression. Combine that with the (vertically) thick main spar, and you get a structure that is resistant to flex and twist all along the leading edge of the wing. With the D.VIII the leading edge 'skin' is extended across the entire top half of the wing, making for a very strong wing.

 

When it comes to wing removal with this design, you would literally need to saw the wing off. However, if we were to use Unreasonable's estimated 100 rounds to hit a flying wire, and apply that to the partial cantilever of the Dr.1/D.VII, a 100 well placed rounds would likely be enough to compromise the wing. On a D.VIII, in WWI, a direct hit from AA? If the plane was parked and you shot at the leading edge from the front, you could probably get a 100 rounds in a tight enough group to saw the wing off, but during combat maneuvers, I think you'd be hard pressed to cause enough damage to lead to structural failure.

 

If we exclude the D.VIII from the equation, the probability of wing loss for braced vs. cantilevered is likely close enough as to make no difference, assuming you can get those 100 rounds on target. But getting those rounds on target is tricky in both cases, meaning both types should be rather difficult to de-wing. I have seen little data to support that either design was particularly prone to wing shedding... without the pilot contributing to the failure. Pilot accounts don't really tell us what flight state their victims were in when they "fell to pieces", but personal experience should tell us exactly what state they were in: evasive.

 

In conclusion, we do have enough data to indicate that the in game wing loss rate is greater than it should be, even if we only used MvR's 15% (which is an extreme). Personally, I think adjusting the DM to actually determine bracing wire loss would help bring the numbers down to a more reasonable degree. At least that probability is calculable, and would allow for the proverbial ' golden bb' effect while simultaneously bringing the braced designs in line with the cantilevered. You could just adjust the wing failure percentage in the RNG, but that can get you the problem of every plane is like the D.IIIa, or the bullet sponge that is the D.VII. I like that there is degrees of structural integrity based on damage level. I like having to nurse my floppy winged DH.4 home. I just think the DM needs to be refined, because I don't like having my Dolphin de-winged by a stray bullet while in a gentle, banked, climbing turn.

 

[unreasonable]

Constructive post: couple of points before I start constructing the FC Utility - Camel:

 

1) "loss of a single flying wire would not lead to immediate wing loss, rather a reduction in wing strength. The subsequent wing loss would be due to over stressing that wing,"

 

I think everyone understands that the way the wing usually fails under MG fire is by having it's maximum g limit reduced, rather than it being blown off as in a  heavy AA hit. Even so, if you are happily flying at 3g, get a wing/wire hit that reduces your maximum g limit to 1g, then your wings fold up, that hit was a necessary and sufficient condition of wing loss (ie the cause), under those conditions.  Hence in AnP's analysis of Camel spar failure, he shows the hits required to break for 1g,2g up to 5g IIRC.

 

In calculating any p of wing loss per hit, the more you can specify the g condition the more accurate the p for that case. Agreed.  But you can still aggregate the probability for a sample including more than one g condition to apply to cases where you do not know the g condition. It just has a wider confidence limit.

 

Just as you could give a p to be killed by an MG bullet for a soldier on a battlefield, when wearing body armour or helmet makes a difference in individual cases, but you do not know if this case is wearing armour or not.  If you do know the armour state, you change the p and increase your confidence. 

2) "In conclusion, we do have enough data to indicate that the in game wing loss rate is greater than it should be, even if we only used MvR's 15% (which is an extreme)""

 

The most insightful way to measure wing loss rate, is losses per hit. Analysis by AnP was about that, and it is so far the only actual data we have. Based on that he might still disagree. There is still a bit of a mystery why the "lived experience" of so many players  is so radically different to the results of his 10,000 rounds per hit box per g level robot trials. 

 

What most conclusions are based on, including yours here implicitly since you mention MvR's 15%, is the wing loss ratio, ie wing losses/all losses.

 

To get that ~right and the overall loss rate ~ right you not only have to have the wing loss rate ~right, but also all the other rates, at least the ones that comprise a significant proportion of the total.

 

Possibly, the other rates are more heavily biased to being too damage resistant. For instance, based on limited testing, I suspect that:

 

- Injuring/killing the pilot is harder than it should be: it takes a minimum of 3 MG hits to kill an AI pilot firing through the fuel tank. Nothing like as bad as RoF's concrete fuselages, but still doubtful.

 

- Setting planes on fire may be harder than it should be, especially in the 1918 era of widespread incendiary and HE/I use.

 

Since those two make up most of the non-wing losses, increasing their rates would reduce the wing loss ratio, potentially by a lot. But you could also have an increased overall loss rate, which may not be what the people would like, even if it is actually more accurate.  Alternatively reduce the wing loss rate as I assume you propose: your loss/ratio is closer to what you want, but the overall loss rate may be less accurate. (edit this is not easy to explain......)

 

Finally, On MvR - he may be an outlier, but IMHO more likely due to his target set, than his skills. (We do know he shot down a lot of BEs).  But I was amused to see you arguing that his allegedly tight groups and accurate aim at the pilot would increase his wing loss ratio, compared to a random distribution of hits, while earlier in the thread someone else was arguing that these very same factors must necessarily lead to a lower wing loss ratio: so low in fact that MvR must have making false reports!

 

edited.

 

  

Edited September 14, 2021 by unreasonable

[No.23_Starling]

5 hours ago, unreasonable said:

Constructive post: couple of points before I start constructing the FC Utility - Camel:

 

1) "loss of a single flying wire would not lead to immediate wing loss, rather a reduction in wing strength. The subsequent wing loss would be due to over stressing that wing,"

 

I think everyone understands that the way the wing usually fails under MG fire is by having it's maximum g limit reduced, rather than it being blown off as in a  heavy AA hit. Even so, if you are happily flying at 3g, get a wing/wire hit that reduces your maximum g limit to 1g, then your wings fold up, that hit was a necessary and sufficient condition of wing loss (ie the cause), under those conditions.  Hence in AnP's analysis of Camel spar failure, he shows the hits required to break for 1g,2g up to 5g IIRC.

 

In calculating any p of wing loss per hit, the more you can specify the g condition the more accurate the p for that case. Agreed.  But you can still aggregate the probability for a sample including more than one g condition to apply to cases where you do not know the g condition. It just has a wider confidence limit.

 

Just as you could give a p to be killed by an MG bullet for a soldier on a battlefield, when wearing body armour or helmet makes a difference in individual cases, but you do not know if this case is wearing armour or not.  If you do know the armour state, you change the p and increase your confidence. 

2) "In conclusion, we do have enough data to indicate that the in game wing loss rate is greater than it should be, even if we only used MvR's 15% (which is an extreme)""

 

The most insightful way to measure wing loss rate, is losses per hit. Analysis by AnP was about that, and it is so far the only actual data we have. Based on that he might still disagree. There is still a bit of a mystery why the "lived experience" of so many players  is so radically different to the results of his 10,000 rounds per hit box per g level robot trials. 

 

What most conclusions are based on, including yours here implicitly since you mention MvR's 15%, is the wing loss ratio, ie wing losses/all losses.

 

To get that ~right and the overall loss rate ~ right you not only have to have the wing loss rate ~right, but also all the other rates, at least the ones that comprise a significant proportion of the total.

 

Possibly, the other rates are more heavily biased to being too damage resistant. For instance, based on limited testing, I suspect that:

 

- Injuring/killing the pilot is harder than it should be: it takes a minimum of 3 MG hits to kill an AI pilot firing through the fuel tank. Nothing like as bad as RoF's concrete fuselages, but still doubtful.

 

- Setting planes on fire may be harder than it should be, especially in the 1918 era of widespread incendiary and HE/I use.

 

Since those two make up most of the non-wing losses, increasing their rates would reduce the wing loss ratio, potentially by a lot. But you could also have an increased overall loss rate, which may not be what the people would like, even if it is actually more accurate.  Alternatively reduce the wing loss rate as I assume you propose: your loss/ratio is closer to what you want, but the overall loss rate may be less accurate. (edit this is not easy to explain......)

 

Finally, On MvR - he may be an outlier, but IMHO more likely due to his target set, than his skills. (We do know he shot down a lot of BEs).  But I was amused to see you arguing that his allegedly tight groups and accurate aim at the pilot would increase his wing loss ratio, compared to a random distribution of hits, while earlier in the thread someone else was arguing that these very same factors must necessarily lead to a lower wing loss ratio: so low in fact that MvR must have making false reports!

 

edited.

 

  

Your 15% is wrong. Kill 77 was not an SE5a that fell apart; if you read Franks, the book you quote, you’ll see he identifies the plane as a Camel which made an emergency landing with the pilot surviving. It should be 13-14%.

 

That being said, McCudden wasn’t actually shot down by MvR so that kill should never have counted giving you 11/79 or 14%. Either way your 15% is too high using your own source.

 

I didn’t say you personally had to supply data on Entente planes shooting up DVIIs, I’m saying that your dataset is heavily biased and flawed in answering my initial question of cantilever vs braced, without citing qualitative evidence for both sides of the argument. I’m also critiquing your personal conclusion which eschews contrary evidence as personal bias.

 

I still need to reread Rickenbacker…

 

 

13 hours ago, ZachariasX said:

All you need to sever is the top layer of the box spar and you affect the whole structure. As that part of the beam is AFAIK made from spruce, I think it should be somewhat robust though.

 

Absolutely.

Lots of fantastic contributions from you all. I think I’m starting to form some conclusions:

 

1) wings did fail, particularly after a long burst into an single area e.g. McCudden, but rarely failed from a golden BB. The 303 round is a precision weapon not an explosive shotgun like a cannon or AA, far more likely to tear through surfaces. The vast majority of hits to either the braced or cantilever wing would be fabric holes

 

2) a long burst in the vicinity of the pilot would be far more lethal - bigger target, more squishy, and MvR’s tally accounts for this

 

3) Pilot wounds or fires accounted for far more losses (86% in the case of MvR’s kills, most likely)

 

4) Cantilever designs offered advantages but given that most losses were to wounds, engine damage, or fires, wing loss prevention wasn’t the main motivation for the design

 

5) we need someone (Larner??!) to provide a similar tally to Unreasonable’s looking at Fokker losses to a single pilot and tally reasons for loss so we have some kind of compatible dataset to MvR’s

[unreasonable]

1 hour ago, US93_Rummell said:

Your 15% is wrong. Kill 77 was not an SE5a that fell apart; if you read Franks, the book you quote, you’ll see he identifies the plane as a Camel which made an emergency landing with the pilot surviving. It should be 13-14%.

 

I am perfectly happy to allow you to sum up what is your thread, but not if you do it with gratuitous snark.

 

***

According to Franks there were two possible candidates for Claim 77, and he says "The identity of the victim is not clear". "Perhaps the luckiest of Richthofen's victims"  in the box on Gallie, is ambiguous, probably deliberately so. It could mean either "He was a victim of R, and perhaps the luckiest" or "Perhaps he was R's victim, in which case he was the luckiest".      

 

The pilot who was told by the Germans that he was MvR's claim (Adams) did not make an emergency landing, he crashed, with the aeroplane "smashed to pieces" and the pilot "left for dead under a tarpaulin", all consistent with the sortie ending because, in MvR's words, "the enemy plane broke apart".

 

When faced with uncertainty of this kind, you can either throw the case out completely, (ie the total is now 79), put it in an undecided box, or go with the claim. IMHO the evidence for the Adams case is strong enough to go with the claim. You can disagree, but what you cannot reasonably do is say, if it was not certainly a wing kill then it was certainly a PK. (Especially when Gallie was "completely unscathed!")

 

*** 

 

More importantly, why do you think I care whether the number is 14% or 15%? Or even 5% or 25%?  I raised the MvR case to illustrate the problem with the "very very rare" or "one in a million chances" pov, not to defend a specific number. It could be 5% and the lottery pov would still be nonsense.  I do, however, care about you implying that I have not read my source or accusing me of bias. Continue to do so and I will continue to post in rebuttal. 

 

*** 

 

MvR, like McCudden, was by all accounts a good shot and very aggressive in getting close, which is partly why they are both outliers in terms of victories.  Staying alive for a long time has a lot to do with it as well.  But none of that demonstrates that MGs were "precision weapons" in the context of air warfare. The maths does not support that. See the post earlier about lateral relative motion.

 

We do not know his hit/shot ratio for his victories, though we can try to guess. Worse, we do not know how many shots he fired in other engagements when he did not score and possibly did not hit. You cannot deduce his overall shot pattern from his best results. 

 

Hence the assumption, used by the OR statisticians in WW2, that for a given angle of an attack, the hits are randomly distributed over the projected area of the target for the population of attacks as a whole. 

 

It is illogical to say that MvR is an outlier, because he got unusually close and was unusually accurate, therefore we should take the very best of his shot patterns as typical when we calculate hit distribution!

 

PKs plus flamers were ~65% of MvRs total claims.  There were quite a few engine kills where the crew were able to walk away, plus a couple of structural collapses other than the wing kills discussed.  As always there is some doubt over which box to put some of these in and even what the denominator should be, and I really do not think anyone worry too much about the odd percent or three. 

 

I have no doubt that his changing target mix over time would influence the results, as would changing his ammo. When you look at the late period only (incendiaries) the proportion of flamers and PKs was was ~78%, more flamers than PKs. Still 5-10% structural. But a  very small sample: we have to accept that we are dealing with wide confidence limits, but I would take that as a prior for 1918 fighting for the GAF and a good starting point for sensitivity analysis of hits to a Camel, in my currently under development spreadsheet, which will be out soon in a separate thread.  

 

 

Edited September 14, 2021 by unreasonable

[No.23_Triggers]

31 minutes ago, unreasonable said:

I do, however, care about you implying that I have not read my source or accusing me of bias. Continue to do so and I will continue to post in rebuttal. 

If I'm following correctly, I think Rummell is stating that - directly related to the question of "Braced VS Cantilever" - the MvR dataset is 'biased' towards providing data on the rates at which Braced designs broke up because, well, none of MvR's victims flew planes with Cantilever spars. I don't think he's implying any personal bias on your part...

...whereas you're saying that the purpose of posting the MvR stats in this thread is a rebuttal to the claims that braced designs being 'shot to pieces' is an exceptionally rare thing to happen, and that the MvR stats aren't attempting to address the "braced vs cantilever" discussion. 

I'm inclined to agree about Braced designs vs structural failure from battle damage, from what I've read in pilot accounts and from what the community has already researched and posted here. I think it was an uncommon occurrence (with more likely outcomes of aircraft destruction being more frequent), but not overtly rare. 

Comparing to the types of aircraft 'fates' I tallied up from pre and post 4.005, I'd say that pre-4.005 was far too lenient, but post-4.005 was far too high.

For reference (taken from our virtual unit's reports): 



EDIT: "other" here referencing pilot killed / forced to land / crashed intact / etc, etc

Edited September 14, 2021 by US93_Larner