Jump to content

4.006 DM Discussion


Recommended Posts

I'm having a hard time wrapping my head around how wires can *bear* a load.  It would seem like what they could do would be to *distribute* a load.  Those wires without the wings themselves aren't rigid.  Unless they're holding the wing together in some sort of "keystone" arrangement I guess.

 

Whatever.  I'm not an engineer.

Edited by J28w-Broccoli
Link to post
Share on other sites
54 minutes ago, J28w-Broccoli said:

I'm having a hard time wrapping my head around how wires can *bear* a load.

They transform a torque force on the spar into a compression force in the spar and a pull force in the supporting wire. Wood is great for compressive loads, piano wire great for pull loads.

 

The larger the angle at wich the wire intersects with the plane, the smaller the resulting push and pull force. This is why in principle, you want the wings stacked far, as the diagonal wire and the spar will receive less load for the same g you are pulling the aircraft.

  • Thanks 2
Link to post
Share on other sites

Fix the friggin glass wings of the Camel! 
It's ridiculous. Before the DM patch I could plenty of kills. Now against the D7F all they have to do is shoot a bit at your wings and wait for them to fold either right away

or after you start turn-fighting. 

Link to post
Share on other sites

I was thinking about this late at night. It seems to be that we are not going to get detailed modelled wings with every wire and strut included. Taking a single main element like spar size is also produced wildly skewed differences and killing gameplay for a section of players. Until the engine and average home GPU can cope with detailed wing models I wonder whether the focus should be to roll back to where we were pre update, with all the wings much more resilient, some more so than others. I can’t see any other way out right now.

Edited by US93_Rummell
  • Upvote 3
Link to post
Share on other sites

I agree. As long as we don't have the spars and wires to hit, any predictive model cannot have to much difference in damage tolerance. It's a game after all and whatever realism we get, it has to be permissive for good gameplay. It's good that we have differences in damage tolerance but at some points, one has to cut corners to keep it a good game. There are many places in this sim where we have "not so realistic" options and that is fine. I feel it would be out of place insisting on (theoretical) truth here, when in fact the whole IL2 series is a fantastic balance of what is realistic and what works in a game. The current odds for demise after few hits are just way too big for some aircraft in the game.

  • Like 1
  • Upvote 5
Link to post
Share on other sites

 

4 minutes ago, US93_Rummell said:

 I wonder whether the focus should be to roll back to where we were pre update, with all the wings much more resilient, some more so than others. I can’t see any other way out right now.

 

I don't think the devs are likely to do that because it means also rolling back WWII. Plus it wasn't just the DM changed in the update(s). Besides, the dummy spitting would be so severe over there a random projectile would likely take your eye out. No-one wants that. This is supposed to be a safe environment where we can kill each other without nonsense like danger or harm to ones person.

 

I really do hope they tweak the current DM though. I don't want to see tumbleweeds. I WANT TO SEE ACTION!!!  S!   Please ….

 

In the meantime any skinners around who can emblazon "Fragile! Handle with  Care!" on the upper side of the top wing and the underside of the bottom wing of my Camel. That should give enough authority to stop people shooting at my wings don't you think?

 

 

  • Haha 2
Link to post
Share on other sites
58 minutes ago, US93_Rummell said:

It seems to be that we are not going to get detailed modelled wings with every wire and strut included.

 

You don't need that.

 

WITHOUT IT you can perfectly well define a probability model (specific to each type of plane) of what kind of damage a bullet hitting the hitbox will do. First you need to guess which part was hit, then how severely it was hit.

 

WITH IT you could ensure that your probability model was finely grained, i.e. you would know which part was hit, and the probability model would then decide how severely it had been hit.

Edited by J2_Bidu
  • Upvote 1
Link to post
Share on other sites
1 hour ago, US93_Rummell said:

I was thinking about this late at night. It seems to be that we are not going to get detailed modelled wings with every wire and strut included. Taking a single main element like spar size is also produced wildly skewed differences and killing gameplay for a section of players. Until the engine and average home GPU can cope with detailed wing models I wonder whether the focus should be to roll back to where we were pre update, with all the wings much more resilient, some more so than others. I can’t see any other way out right now.

A.N.Petrovich has said on more than one post this is not an option as the whole game engine has been changed .

  • Upvote 1
Link to post
Share on other sites
1 hour ago, US93_Rummell said:

I was thinking about this late at night. It seems to be that we are not going to get detailed modelled wings with every wire and strut included. Taking a single main element like spar size is also produced wildly skewed differences and killing gameplay for a section of players. Until the engine and average home GPU can cope with detailed wing models I wonder whether the focus should be to roll back to where we were pre update, with all the wings much more resilient, some more so than others. I can’t see any other way out right now.

 

The effect of wires is already factored into the DM.  The undamaged breaking strains are based on documented limits - as far as we know - not just on the thickness of the spars.  If the undamaged limits are historically correct, they already include the effects of everything holding the wings together, whether that is spar, wires or trained hamsters.  Do you think you can do 4 G loops in a Camel after taking all the wires out?

 

You are assuming that, if you had additional explicitly modeled wires hit boxes, this would mean that the damaged G limits, for wire braced planes, should be higher than they are currently modeled. I have yet to see an even partially coherent argument to support this assertion. Wires cannot take the load from broken spars.  They can spread the load from damaged spars - but they are already doing that, implicitly, just as they are already spreading the load from undamaged spars.


If you add the wires as an explicit hit box, with a specified load bearing effect, then you have to have a lower G limit for the spars to reach the same historic maximum load for the complete wing structure.   

    

 

 

Edited by unreasonable
Link to post
Share on other sites
59 minutes ago, J5_Spyboy said:

A.N.Petrovich has said on more than one post this is not an option as the whole game engine has been changed .


Any value can be assigned to any wing. The numbers are purely notional and a best guess, not some set in stone physical certainty.  What happens with the damage model and the values assigned to different elements aren’t necessarily one and the same.

  • Upvote 2
Link to post
Share on other sites

I understand we can’t roll back the whole DM, I’m suggesting that wings are strengthened somehow back to where they were prepatch. Perhaps a realism option in the menu that buffs wing toughness for the Dv etc, just as we have had options in the past to help with gunnery etc. Id expect that would be welcomed by a lot of the MP community. 

Link to post
Share on other sites
42 minutes ago, HagarTheHorrible said:


Any value can be assigned to any wing. The numbers are purely notional and a best guess, not some set in stone physical certainty.  What happens with the damage model and the values assigned to different elements aren’t necessarily one and the same.

 

For example a "quick fix" (said with no notion of how easy this is in reality) would be to take the values of one of the current aircraft as the weakest any should be and then scale all the aircraft values up so the weakest wings are at the new minimum and all the others are remain at the same values relative to one another. That way the individual aircraft remain consistent relative to one another but the baseline moves up.

Link to post
Share on other sites

 

1 hour ago, slug_yuugen said:

 

For example a "quick fix" (said with no notion of how easy this is in reality) would be to take the values of one of the current aircraft as the weakest any should be and then scale all the aircraft values up so the weakest wings are at the new minimum and all the others are remain at the same values relative to one another. That way the individual aircraft remain consistent relative to one another but the baseline moves up.

 

AnP has already described the DM: there is no single wing value that you can scale as an input.   The DM depends on hit to "spar" sections of defined length and thickness, each plane having some number. These have varying properties even within the same plane.   You can see on AnP's distribution graph that the Camel has at least two types of "spars" with different ability to take damage. (I will explain why I think that if requested).


So if you want to change the outputs by altering plane properties you have to change some variable like "spar" size, one aircraft at a time, test and repeat, unless you eliminate spar damage altogether and just have the final level of cumulative damage that is used for the damage graphic instead trigger complete wing collapse, so that there is little or no difference between planes.

 

The thing that might be possible is to reduce the damage per spar hit for each bullet.  The "thick spar" planes would then appear to be even more damage resistant than they are now, but that makes relatively little difference to them, since, when shot at, something else will usually bring them down before they can accumulate enough wing damage to fall that way.  

 

 

  

Edited by unreasonable
Link to post
Share on other sites
4 hours ago, unreasonable said:

 

The effect of wires is already factored into the DM.  The undamaged breaking strains are based on documented limits - as far as we know - not just on the thickness of the spars.  If the undamaged limits are historically correct, they already include the effects of everything holding the wings together, whether that is spar, wires or trained hamsters.  Do you think you can do 4 G loops in a Camel after taking all the wires out?

Yes, neither the struts nor the wire between the wings have their DM in level G, they only have visual DM. 

That’s why you can see this https://youtu.be/n17mf4ETUxw  Everything except the spar is damaged in the video, so the plane can fly like absolutely whole.

All the maximum G value that the structure can withstand is placed in the wing hit boxes.  Any hit in the wing subtracts a certain percentage of strength from the total value of G.  The magnitude of this percentage depends on the area of the wing, right?  And if you add to this area the one that is occupied by supports and wires?  And also take into account the fact that the probability of getting into poles and wires is extremely small?

Now we have such a situation: Any hit by a bullet in the wing causes a weakening of the spar, supports and wires at the same time.  This is a very big assumption.  I don’t understand how one can seriously talk about the realism of such a DM ??

 

Edited by emely
  • Upvote 1
Link to post
Share on other sites
17 minutes ago, emely said:

Now we have such a situation: Any hit by a bullet in the wing causes a weakening of the spar, supports and wires at the same time.  This is a very big assumption.  I don’t understand how one can seriously talk about the realism of such a DM ??

 

Exactly.

 

The bullets that could hit both spar and (several) wires at the same time are JFK material. We don't need them in FC.

 

17 minutes ago, emely said:

The model is completely accurate, are you sure?  Or do you say this as a player on d7f and dr1?  If this model corresponded to reality, Germany would have won in ww1.

 

I never said this, or implied it, or agree with it.

 

I meant a new probability model could very well better represent plane damage even without changing the hitboxes.

Edited by J2_Bidu
Link to post
Share on other sites
13 minutes ago, J2_Bidu said:

 

I never said this, or implied it, or agree with it.

 

I meant a new probability model could very well better represent plane damage even without changing the hitboxes.

I understood this a little earlier and removed my question from the message, and in the meantime you managed to quote it :-))

  • Haha 1
Link to post
Share on other sites

More musings ....

 

 AnP's graph shows the number of hits needed to break a Camel's spar by averaging the result from firing at each of the spars for 10,000 trials for each G scenario.  I do wonder about this "average" and how it was weighted.

 

What I would like to see is the distribution for each spar, as it may be the case that the the weakest one is very much weaker than the others.  That is one likely explanation for the apparent kink in the line for the high G cases in the graph.  Graph on the right is an example - not intended to exactly model AnP's results.

 

1960287577_AnPcamelbehind.thumb.JPG.e2f186baedfc1379e64a778341ce3579.JPG1374828366_WeightedDistributions.thumb.JPG.7e11ee43bbd6448a0bd78af720923fb0.JPG

 

In the case of a large sample of aircraft subject to randomly placed damage, the weakest area will fail first more often than a strong area, unless is proportionately smaller. If you shot randomly at 10,000 aircraft until a spar collapsed (with invulnerable pilots, engines etc) you would not necessarily get the same average number of hits/break as when shoot each spar individually and create a mean. The new average would be affected by how often aircraft are downed by each spar failing first. 

 

So what people are reporting - apart from the behaviour issues - could also be that the averages given so far, while being correct, do not apply to the case of random hits across the wing surfaces. 

1 hour ago, emely said:

 

Now we have such a situation: Any hit by a bullet in the wing causes a weakening of the spar, supports and wires at the same time.  This is a very big assumption.  I don’t understand how one can seriously talk about the realism of such a DM ??

 

 

The wires and struts are just visual effects to improve immersion. The question of realism is not about correspondence between exact placement of hits and visual effects, in FC any more than in GB. You know perfectly well that the damage effects to wings in GB are generic - this is no different.

 

As AnP has already explained, anyone expecting an incredibly fine grained DM with corresponding visual effects in a real time  MP environment is delusional.

 

If you will not fly like a WW1 pilot you cannot expect to get WW1 results.

 

All that said, whether the current DM creates broken wings at about the right rate, given the number of hits they take, and given the Gs put on them, is an open question.  I think that for undamaged wings the results conform very well to the posted documentation. For damaged wings I am not sure.

  • Upvote 1
Link to post
Share on other sites
16 minutes ago, unreasonable said:

The question of realism is not about correspondence between exact placement of hits and visual effects, in FC any more than in GB. You know perfectly well that the damage effects to wings in GB are generic - this is no different.

I do not know what GB is.  I'm talking about the fact that the damage model does not match the design scheme of ww1 aircraft, even in a simplified form.  And stop talking about the ability of our aircraft to withstand loads in an undamaged state, we have already discussed this 100 times and no one asks questions about this

24 minutes ago, unreasonable said:

 

As AnP has already explained, anyone expecting an incredibly fine grained DM with corresponding visual effects in a real time  MP environment is delusional.

Now delusional is a DM in FC.  I have already seen many beautiful schemes and diagrams with colored lines.  But they do not change anything.  Guys can already trade them as art objects.  Perhaps the sale of these images may exceed the sales of FC 😉

  • Haha 1
Link to post
Share on other sites
4 minutes ago, emely said:

I do not know what GB is.  I'm talking about the fact that the damage model does not match the design scheme of ww1 aircraft, even in a simplified form.  And stop talking about the ability of our aircraft to withstand loads in an undamaged state, we have already discussed this 100 times and no one asks questions about this

Now delusional is a DM in FC.  I have already seen many beautiful schemes and diagrams with colored lines.  But they do not change anything.  Guys can already trade them as art objects.  Perhaps the sale of these images may exceed the sales of FC 😉

 

GB = Great Battles - you have a BoS founders tag, I assumed you would know that.  The DM never matched the design scheme of WW1 aircraft, either in RoF or in pre 4.006.

 

The FM does not match biplanes/triplanes either:  AFAIK all planes are to all intents and purposes monoplanes for FM purposes.

 

And I will talk about whatever the hell I like.

Link to post
Share on other sites
14 minutes ago, unreasonable said:

 

And I will talk about whatever the hell I like.

Of course, you can do whatever you want and write about the same thing in dozens of messages, but sooner or later, only a mirror will remain your interlocutor.

By the way, that beautiful picture in your message above, where 10,000 shots at the wing of a camel - what is it for?  This image does not carry any valuable information, because practically everyone does not give a damn how much this wing will withstand impacts at a load of 1G, if it disappears at 1.5G with three hits

Link to post
Share on other sites

If you had looked at the picture carefully, and engaged brain before mouth,  you would see that it is precisely about the higher G cases, not 1G. To help you out - note that only the 4 and 5 G lines have dog legs.  This could explain why the effective number of wing losses you are experiencing in play could be completely consistent with AnP's graphs.   

 

As for interlocutors: we all know now that you do not like the current DM and want it rolled back, or perhaps replaced with a super DM, modelling all struts and wires as hit boxes.  That seems to be about it, so in terms of repeating yourself, I say to you "pot - kettle".  

 

I am trying to understand if there are any reasons, aside from crazy flying, that explain observed results in a way consistent with AnP's data.  

Link to post
Share on other sites
9 minutes ago, unreasonable said:

If you had looked at the picture carefully, and engaged brain before mouth,  you would see that it is precisely about the higher G cases, not 1G. To help you out - note that only the 4 and 5 G lines have dog legs.  

I already noticed that you are unusually polite)) I looked at these pictures before and now, and I say again that they make no sense.  Absolutely no matter how many hits will be fatal at high load, if the wing is broken at low.  And there is still a condition of firing from the "dead six"!  As we see from practice, ANY hit in the wing in ANY place on it causes damage to the spar.  It can be 0.1% + 0.3% + 0.2% and your wing can fall off with a small load, despite the fact that the hits were not from position six.

(as for the legs in dogs, I can’t say anything, I’m not an expert on these animals)

35 minutes ago, unreasonable said:

, I say to you "pot - kettle".  

Just wondering what that expression means?

 

Link to post
Share on other sites
1 hour ago, emely said:

I already noticed that you are unusually polite)) I looked at these pictures before and now, and I say again that they make no sense.  Absolutely no matter how many hits will be fatal at high load, if the wing is broken at low.  And there is still a condition of firing from the "dead six"!  As we see from practice, ANY hit in the wing in ANY place on it causes damage to the spar.  It can be 0.1% + 0.3% + 0.2% and your wing can fall off with a small load, despite the fact that the hits were not from position six.

(as for the legs in dogs, I can’t say anything, I’m not an expert on these animals)

Just wondering what that expression means?

 

 

"The pot calling the kettle black".  It just means that you are accusing me of doing something, namely repeating yourself, that you have been doing as well, with colourful videos, in several threads.  (BTW no-one is forced to read my posts, you can always put me on your ignore list).

 

The pictures make perfect sense.   If you fire at a Camel wing in the area of one spar until the wing breaks, at a specific G load, 10,000 times, from behind, and count the number of hits required to break the spar each time, graph the number of cases against the shots needed to break, then do the same for each of the other spars, then average the results, you get AnP's graph. Perfectly straightforward, except I do not know how he weighted the average.

 

He illustrated that the perception of the players - who mostly guessed on the low side how many shots this would take on average - did not correspond to the reality as revealed by his huge tests. That particular case - thin winged Camel, from dead six, was chosen because it would produce the lowest  number of hits to break.

 

What I am arguing is that if one spar can withstand 50 shots on average, but another can only withstand 10, the observed average from random hits all over the plane will probably not be (50+10)/2 = 30 but could be considerably less (or more).  AnP's test takes into account the relative weakness of the spar types (and their relative area?), but not, I think, how often each one would break first.   

 

So even without adventurous flying and selection bias, it is possible that players could observe significantly lower average numbers of hits to break than would appear from looking at the averages in the charts. 

 

 

 

 

Edited by unreasonable
  • Upvote 1
Link to post
Share on other sites
5 hours ago, unreasonable said:

The wires and struts are just visual effects to improve immersion. The question of realism is not about correspondence between exact placement of hits and visual effects, in FC any more than in GB. You know perfectly well that the damage effects to wings in GB are generic - this is no different.

 

Are they? The fact that they are not represented as separate hitboxes does not mean that damage is not accounted separately for each part. It might very well be. Does anyone know for sure?

 

3 hours ago, unreasonable said:

[...] a super DM, modelling all struts and wires as hit boxes.  [...]

 

We should separate the concept of

    DM = Damage Model

from the concept of

    TM = Target Model

and finally the

    MM = Match Model

 

As I said before, these are fundamentally different things.

 

Problem 1) Detecting where bullets are hitting.

   => This can be done with more or with less precision, and is what I will call TM. Apparently many hitboxes are computationally expensive, and some practical limit must be accepted.

 

Problem 2) Determining the effects of damage on individual plane parts.

   => This is comparatively cheap as hell, computationally, but probably requires some good thinking. A nice model can be figured out for every plane, a model that is robust and independent from the solution of Problem 1.

 

Problem 3) A match (MM) between the two models must be arranged.

 

Three extreme examples:

 

Case 1

DM = 1 Large Indivisible Plane to be damaged

TM = 10.000 hitboxes, for every nut and bolt.

In this case, the game knows exactly where each bullet is hitting. There is however only one thing to damage. The MM considers different weights for each of the TM hitboxes to subtract different quantities from the DM until the plane collapses. Hits on the spar may subtract more, etc.

 

Case 2

Just like case 1, but the MM is very poor - every hit subtracts the same in all places.

 

Case 3

DM = 1 finely detailed set of plane parts to be damaged

TM = 1 giant hitbox

In this case, the game knows every hit the plane takes, but unfortunately does not have a clue where. The MM determines where the hell the bullet went by assigning different probabilities of it going to different parts. Upper wing fabric piece #37: 0.5%; Spar: 7%; Left wheel: 2%, etc. And then it smartly figures out the consequences to your plane of all the damage it is incurring.

 

Again, the visual consequences of all this can be separately handled. They should reflect the DM, and not necessarily the TM.

 

This scheme is not very complicated, is scalable, and separates concerns. And the several pieces can be tested separately too.

Edited by J2_Bidu
Link to post
Share on other sites
1 hour ago, unreasonable said:

 

He illustrated that the perception of the players - who mostly guessed on the low side how many shots this would take on average - did not correspond to the reality as revealed by his huge tests. That particular case - thin winged Camel, from dead six, was chosen because it would produce the lowest  number of hits to break.

All that has been illustrated is an attempt to prove to users that with DM everything is ok.  In my opinion, she failed.

1 hour ago, unreasonable said:

The pictures make perfect sense.   

Well, let's try to find some meaning there.  Please note that complaints about DM are mainly related to the situation when the number of hits is from two to five.  This is the area of the chart and consider.

IMG_8047.PNG.58b70b7b6d6b11545ee32411ccf80df8.PNGк

 

The most interesting part of the diagram for us is the least informative.  But we can say with some approximation that when three bullets hit, the probability of wing breakage at 1G and 2G is equal and it is low (maybe about 5 cases).  With 4G these cases are already 25, and with 5G 225!  But here is the problem with getting three hits, you have a chance to break the wing for any value of G.  And it works over and over again.  My almost all 100% of deaths are due to the breakage of the wings from the presence of three holes, although the probability of this is not high even with 5G 🙂

So what practical sense does this chart have?  I think it makes no sense to us and our problem.

  • Upvote 2
Link to post
Share on other sites

The in game DM should me more generous than that statistical reality assumed, due to the simple fact that the very way the DM works, me as a pilot I have no way to infer whether my wing will go at 2 g or not. In the real world I can see if I have a nice group of hits on my spar. Or at least I could see it. If it is just an inch away from the spar, my wing will have lost *none* of its strenght.

 

The statistical weakening derived from supposed numbers of hits is a nice theory, yet it is not suitable for gameplay and nor is it realistic, quiet obviously.

 

Applying statistics of big numbers to the individual case is meaningless. It is not the way to do math. You can do this if you want a simulation of 1000 AI planes fighting each other and then have the correct nomber of wing failures if you also program the AI pilots to behave in a certain way. Then it works. But it doesn‘t work for us.

 

Remember, the only reason you do statistics is to get a good answer to a question. It is obvious that we do not have that yet.

 

It is inherently wrong for gameplay to have a dice thrown for every player instead of him be at the end of a sequnece of causes, especially when odds are extremely lopsided. If your demise in a game is in no way manageable but just random luck, then this is neither realistic nor fun. It kills realism as it kills the game, I cannot put it more plainly than that. Realism does not work like that as actual individual hits count, and the game doesn‘t work like that because it‘s not fun to turn home after hearing *clac* just once. You simply cannot have such aircraft in a game that is supposed to be fun and there weren‘t any of such aircraft back then either.

 

It is ok that the Fokkers are sturdy aircraft. They should be. But in the game they are simply in a different class. And this is clearly NOT as it was back then.

 

Realism requires more than getting some statistics right. Same for fun. (For the most of us I suppose.)

  • Upvote 2
Link to post
Share on other sites
53 minutes ago, ZachariasX said:

 

Realism requires more than getting some statistics right. Same for fun. (For the most of us I suppose.)

We have statistics on both the game and the real from " Larner & Co"))) However, I don’t see the interest of officials on the side of it.

 

Edited by emely
Link to post
Share on other sites
53 minutes ago, ZachariasX said:

[...] In the real world I can see if I have a nice group of hits on my spar. Or at least I could see it. If it is just an inch away from the spar, my wing will have lost *none* of its strenght.

 

I don't mean to be picky, but can a pilot actually see this? It doesn't seem possible to me.

 

53 minutes ago, ZachariasX said:

It is inherently wrong for gameplay to have a dice thrown for every player instead of him be at the end of a sequnece of causes, especially when odds are extremely lopsided. If your demise in a game is in no way manageable but just random luck, then this is neither realistic nor fun. It kills realism as it kills the game, I cannot put it more plainly than that. Realism does not work like that as actual individual hits count, and the game doesn‘t work like that because it‘s not fun to turn home after hearing *clac* just once. You simply cannot have such aircraft in a game that is supposed to be fun and there weren‘t any of such aircraft back then either.

 

It is ok that the Fokkers are sturdy aircraft. They should be. But in the game they are simply in a different class. And this is clearly NOT as it was back then.

 

Realism requires more than getting some statistics right. Same for fun. (For the most of us I suppose.)

 

I mostly disagree. The problem is indeed that "odds are extremely lopsided". And that's all. Most or almost all games have dice thrown, one way or the other. It is a handy way of encapsulating inherent complexities that can't really be calculated. If it fits the appropriate statistical distribution, and if it is comparatively fair, it's ok. So let's not throw the baby out with the bathwater.

Edited by J2_Bidu
  • Upvote 1
Link to post
Share on other sites
11 hours ago, unreasonable said:

 

The effect of wires is already factored into the DM. 

 

 

    

 

 

 

I disagree,  If wires, bays, or indeed, even faux bays were given some significance within the damage model, then we would see evidence for that when looking at the comparative damage models compared to aircraft such as the Albatros and Pflaz DIII.  We have three late war Allied Scouts that all have individual features, that I can only assume were added for the purpose of providing strength and damage redundancy (?), at, I might add, the cost of aerodynamic efficiency, or speed.  Never the less we find that these aircraft are deemed to be on a parity or, even worse, significantly inferior to opposition aircraft that emphasised aerodynamic efficiency over structural strength in order to eck out every ounce of performance from otherwise average performing aircraft.

 

  • Upvote 1
Link to post
Share on other sites
17 hours ago, unreasonable said:

 

The effect of wires is already factored into the DM.  The undamaged breaking strains are based on documented limits - as far as we know - not just on the thickness of the spars.  If the undamaged limits are historically correct, they already include the effects of everything holding together...

 

Perhaps, is there proof of this? And even if so, to what extent? For example, are DM weightages applied generically or per plane? 

 

Practically though, no matter how the DM works, the post-4.006 in-game experience is that spar size is all that matters for bullet damage resistance of wings. That has been established thru over 100 controlled tests and much user feedback. 

 

Edited by US103_Baer
Link to post
Share on other sites
5 hours ago, HagarTheHorrible said:

 

I disagree,  If wires, bays, or indeed, even faux bays were given some significance within the damage model, then we would see evidence for that when looking at the comparative damage models compared to aircraft such as the Albatros and Pflaz DIII.  We have three late war Allied Scouts that all have individual features, that I can only assume were added for the purpose of providing strength and damage redundancy (?), at, I might add, the cost of aerodynamic efficiency, or speed.  Never the less we find that these aircraft are deemed to be on a parity or, even worse, significantly inferior to opposition aircraft that emphasised aerodynamic efficiency over structural strength in order to eck out every ounce of performance from otherwise average performing aircraft.

 

 

The SPAD XIII was a late war scout - but it used a wing design from the SPAD S.A. which was first flown in 1915   All the British scouts also used very thin wing sections. Post war aircraft all eventually ended up using unbraced,  strong cantilever wings pioneered by Fokker and Junkers - they really were so much better in every way. 

 

6 hours ago, ZachariasX said:

The in game DM should me more generous than that statistical reality assumed, due to the simple fact that the very way the DM works, me as a pilot I have no way to infer whether my wing will go at 2 g or not. In the real world I can see if I have a nice group of hits on my spar. Or at least I could see it. If it is just an inch away from the spar, my wing will have lost *none* of its strenght.

 

The statistical weakening derived from supposed numbers of hits is a nice theory, yet it is not suitable for gameplay and nor is it realistic, quiet obviously.

 

Applying statistics of big numbers to the individual case is meaningless. It is not the way to do math. You can do this if you want a simulation of 1000 AI planes fighting each other and then have the correct nomber of wing failures if you also program the AI pilots to behave in a certain way. Then it works. But it doesn‘t work for us.

 

Remember, the only reason you do statistics is to get a good answer to a question. It is obvious that we do not have that yet.

 

It is inherently wrong for gameplay to have a dice thrown for every player instead of him be at the end of a sequnece of causes, especially when odds are extremely lopsided. If your demise in a game is in no way manageable but just random luck, then this is neither realistic nor fun. It kills realism as it kills the game, I cannot put it more plainly than that. Realism does not work like that as actual individual hits count, and the game doesn‘t work like that because it‘s not fun to turn home after hearing *clac* just once. You simply cannot have such aircraft in a game that is supposed to be fun and there weren‘t any of such aircraft back then either.

 

It is ok that the Fokkers are sturdy aircraft. They should be. But in the game they are simply in a different class. And this is clearly NOT as it was back then.

 

Realism requires more than getting some statistics right. Same for fun. (For the most of us I suppose.)

 

If it does not work for you, but it works for the AI, then that can only be because you are flying and fighting in a different way to the AI.  I agree that may be partly because the AI can "know" the change in limits due to damage: then again if you use technochat perhaps you can too. (Still not sure about that). I get the impression though, that even if  you had clearer visual and noise signals that you had a weakened spar, you would still dislike the DM.

 

I completely disagree that using probability vs a sequence of causes is "inherently wrong".    Every time an AA or ground based AAMG fires at you the result is determined by a dice roll. There will never be a "sequence of causes" for every result: apart from being computationally unrealistic it is wrong in principle in the real world, because Einstein was wrong. Even without going to that level, at some point in RL calculations you have to make statistically based assumptions, although it is true that people often do not know edit  that they are doing so.

 

With the old DM you were just as much at the mercy of probability based results - including wings still braking due to RNG determined damage - the DM just gave different answers.

 

What can be a problem in statistics is people drawing conclusions from averages rather than considering the effects of the whole distribution, especially when it is unclear how the average is weighted. Hence my post with the graphs.

 

I see no valid arguments concerning the realism of the DM here at all. Whether it is fun for MP or not I have no opinion.  But what I would rant and rail about is having the DM in SP determined by people who insist on confusing a SPAD with a Fw190. 

 

7 hours ago, emely said:

All that has been illustrated is an attempt to prove to users that with DM everything is ok.  In my opinion, she failed.

Well, let's try to find some meaning there.  Please note that complaints about DM are mainly related to the situation when the number of hits is from two to five.  This is the area of the chart and consider.

IMG_8047.PNG.58b70b7b6d6b11545ee32411ccf80df8.PNGк

 

The most interesting part of the diagram for us is the least informative.  But we can say with some approximation that when three bullets hit, the probability of wing breakage at 1G and 2G is equal and it is low (maybe about 5 cases).  With 4G these cases are already 25, and with 5G 225!  But here is the problem with getting three hits, you have a chance to break the wing for any value of G.  And it works over and over again.  My almost all 100% of deaths are due to the breakage of the wings from the presence of three holes, although the probability of this is not high even with 5G 🙂

So what practical sense does this chart have?  I think it makes no sense to us and our problem.

 

At least that is a post that can be engaged with!  That is why, taking reports at face value, I constructed the other diagram. 

 

In that diagram, in the case of the hypothetical red "weak spar", the number of failures after 5 hits would be ~5% compared to close to zero in other spars.  When you average them counting all the trials in the case where you fire at every spar section to destruction, the average could be very different to the case where you fire at random over the wing area and then count the number of hits on the wing section that failed first.   The first depends on how many of each spar there are on the plane, the second on how many of each spar broke in the trials (edit - and the second case is closer to what you will observe while playing). 

 

As I said, this is speculation - but it would be very interesting to see the underlying curves and/or the weighting.

 

  

37 minutes ago, US103_Baer said:

 

Perhaps, is there proof of this? And even if so, to what extent? For example, are DM weightages applied genetically or per plane? 

 

Practically though, no matter how the DM works, the post-4.006 in-game experience is that spar size is all that matters for bullet damage resistance of wings. That has been established thru over 100 controlled tests and much user feedback. 

 

 

Given that conclusion, if the pre-damage G limits are accepted as being broadly realistic based on documented RL tests, then it follows that the resistance of spars to damage starts with a base case of "spar strength" that must include the effects of both spars and everything else. 

 

Put another way: if the wires and struts were modeled as load bearing entities in the DM, and the undamaged limits remained the same, then the spars of planes with wires and struts would make a lower contribution than they do now. Plus your hit box area would be a little bigger. Be careful what you wish for.  

Edited by unreasonable
  • Upvote 2
Link to post
Share on other sites

Just to illustrate my weighting point, in case it is not yet clear.  

 

Here I have two underlying distributions, which when weighted 5 blue:1 red give an average line (brown) very close to AnP's 4G case (light blue).

If there were five times as many red spars as blue spars, the result of AnP's test of each spar to destruction, would be as per his line, near enough.

 

2137927325_FittoAnP1.thumb.JPG.2666ea8ed8b8391facf1e3165b5ad0ba.JPG

 

Now suppose that in a large test of random firings across the whole wing, the red spars fail first five times as often as the blue spars. The results of that test would look like the brown line in this graph, and nothing like AnP's line.

 

1538916434_FittoAnP2.thumb.JPG.fb7cfcfeb8dfaddf2426febb8fb26e84.JPG

 

When you play you see the results of what fails first, not the results of each section tested to destruction, so it might be something like graph 2.

 

Note that both of these graphs assume that his data is correct and the same in each case. 

 

So it could be the case, that in AnP's "trick question", he was right that most people were wrong, because they were answering a different question, correctly, namely how many shots are required to break, on average, the spar that breaks first.   

 

 

Edited by unreasonable
Link to post
Share on other sites
3 hours ago, unreasonable said:

 

Put another way: if the wires and struts were modeled as load bearing entities in the DM, and the undamaged limits remained the same, then the spars of planes with wires and struts would make a lower contribution than they do now. Plus your hit box area would be a little bigger. Be careful what you wish for.  

 

...........and that, potentially is the nub of the problem.   The entire wing damage model is summed up by the properties of the spars, even though how those properties are quantified is a potential mystery, not to mention that the spar properties are then given the dimensions of an entire wing section,  The entire load bearing ability of the wings and their resistance to damage is summed up by the properties of the spars and not the multiple features that give a thin aero foil it’s inherent strength.  Maybe you are correct, the thin spars themselves should have a much lower ability to resist increased G forces, but at the same time the elements that need to be damaged, to result in structural failure, are more numerous, harder to hit, and sometimes designed with redundancy, or battle damage, in mind.

 

Another factor that might be missing is quality over quantity.  Nobody would question, within reason, that a metal spar, of equal dimension, would be stronger than a wooden spar but potential quality is not considered when assessing the differences that might be found in wooden spars.  The Albatros and Pflaz, it would appear, had bigger spars, not significantly, but never the less bigger and this results in, despite a lack of other additional strengthening features, a good damage resistance.   Nobody has stopped to ask if the reason the spars were bigger in the first place was less to do with their ability to resist damage and not just because the timber used was inferior to start with which leads to the question does inferior timber, of bigger dimension, resist damage better than smaller dimensions of superior timber ?  The British aero industry, for example, could only source timber, of the required grade, from Canada and N America, manufactured to exacting standards, despite having the resources of the whole Empire from which to draw.  Antony Fokker, or the Albatros works, did not.  Maybe the differences were not significant, if they existed at all, but then neither were the sizes of the spars, both would have failed without wires and yet the SE5a with 8 flying wires per side, and 6 attachment points is no more resistant to damage than an Albatros, with only two flying wires per side and only three attachment points.

Link to post
Share on other sites
8 hours ago, J2_Bidu said:

Most or almost all games have dice thrown, one way or the other. It is a handy way of encapsulating inherent complexities that can't really be calculated.

It is handy if it gives you plausible bottom line results if it was a plausible depiction of what would have happened if you had things done the complex way. I think, now in practise, we have not. I have detailed why the statistical DM box system as we have it factors in easy long range shooting. I challenge the idea that if we had a spar the strength assumed in the current DM we would see both same sooting style and the same bottom line.

 

9 hours ago, J2_Bidu said:

If it fits the appropriate statistical distribution, and if it is comparatively fair, it's ok. So let's not throw the baby out with the bathwater.

Agree. My rant was more directed at the notion that, „you wrong as statistics is right“ to kill any argument when in fact those statistics can be very wrong at times.

 

If you have to simplify things, that is ok. But you better know what you are doing further than just making things sompler. As you do introduce further gameplay  mechanics when making things simpler.

 

2 hours ago, unreasonable said:

If it does not work for you, but it works for the AI, then that can only be because you are flying and fighting in a different way to the AI. 

Nicely put. By AI, you‘re mainly talking about AnP‘s robot drones that give the statistical results you posted. We do not fly like them. Nobody would play this game if we were to fly like these target drones. 

 

2 hours ago, unreasonable said:

I completely disagree that using probability vs a sequence of causes is "inherently wrong"

At least as we have it, for the purpose of a gamer it is mathematically wrong as representation, as shooting distance must be factored in to even get to desired hit probabilities. Your calculation makes only sense in specific cases. In others, it is not correct. That is a big reason why it works for robot drones and not so much for us.

 

Any game mechanics that allows a player to maximise his chances WILL be exploited in player vs. player. The AI normally doesn‘t do that. Hence the idea that „if it works for drones, it must work for you“ is just not understanding the concept of what is wrong. Nobody really contests undamaged g load tolerances and nobody really contests the hit requirements for a spar to fail. 

 

If you have a system that encourages long range shooting due to the hit effect being more lottery than a player managing his odds, then you will force different behaviour. It is called sniping. If you now increase chances of demise fow one player twice as high as the other, you start having a real problem in terms of gameplay. None of this has much to do with realism, but it is the DM mechanics enforcing certain behaviour to be successfull. Currently, it encourages non-autentic behavior. As damage probability changes with changing shooting distance, keeing it constant will give you an error there, not representing what you actually intended. Having drones shoot each other at the perfect distance giving consistent results is no proof for anything you see in the game.

 

This is why as we have it, the mathematical model we have is in many cases (yes, it is case specific!) wrong, as it doesn‘t really apply for the situation the player finds itself in. If you were to do things that way, you either need more variables or in all cases give some leniency to players that are quick to be fragged by the chance factor you create for the DM.

 

2 hours ago, unreasonable said:

There will never be a "sequence of causes" for every result: apart from being computationally unrealistic it is wrong in principle in the real world, because Einstein was wrong.

He was not, but I see your point. But if your mechanics are „bag of rice falls over in China“ gives you a death probability of 1% as you cannot compute the butterfly effect that leads to your death, you can do is accept it and ignore it. Same as your cancer risk. It is no game, just a downer. And if it is too much of a downer, you drop the game. Any probability that you cannot manage is meaningless to you and you have to disregard it. What it does, it just sets a handicap. If the handicap is such that you don‘t see reasonable chances to overcome it by ability, you just ditch the whole thing. Looking at the Flugpark numbers, I get the impression we are seeing that now.

 

4 hours ago, unreasonable said:

I get the impression though, that even if  you had clearer visual and noise signals that you had a weakened spar, you would still dislike the DM.

Good you brought that up, because this is a fundamental misunderstanding. I should have been clearer about that. Again, in no way I am disputing the potential bullet damage. I am really fine with that. I am especially fine with the wings max. g load tolerances when intact. But I want to point out that your way of applied statistics do much more to how you play this game successfully than you care seeing.

 

The whole game is about managing your chances. Same is air combat back then. MvR wasn‘t a victim of sheer probability to be very successful. What he did was managing his chances. Every time you take away the possibility to manage your chances you take a step back from realism, regardless of how plausible you perceive your statistics to be.

 

In this, I challenge you on the idea that results from a hitbox system that assumes 50 hits having a high probability for fatal damage on a specific section would be the same as if you actually had that (small) section requiring say 30 *visible* (yes, you can see if a spar is shot) hits to fail. With results I mean player behaviour in the game as well as number if wing sheddings. This is why I say in player vs. player, the DM does not live up to its own assumption.

 

I cannot overstress how important the aspect of „managing your chances“ is. Especially back then. Especially when the chances are bad. These young lads, they would not have gone on missions like they did if they were told that some divine RNG was snuffing them after 8 missions. If they did, they would have been Kamikaze pilots and acted accordingly. But almost none did. Everyone thought that if he just kept a sharp eye, flew better, aimed better, they will make it. You may say, „poor lads, they don‘t know statistics“. But in fact survival rate is not random. It is heavily skewed towards certain individuals that learned to manage their chances. This includes knowing cues. You have to have those too. And it is not because he was selected by the divine RNG selected one with a higher survival chance, it is because the individual acted.

 

It is not bad having some random functios as we do not need to overcomplicate things. But it is bad not recognizing this fundamental aspect of air combat and genaral playability. Nobody wants to openly play lottery when starting the game, despite one can calculate chances ex post. If you turn it the other way around, you‘re not having a combat simulator anymore, but a lottery. Nobody would attend.

 

Right now in the game, we have an unrealistic setting: Optical damage does not really correspond to structural damage and hit effects are also by chance rather than just by ability. Both are fundamental game mechanics that are introduced artificially that work aginst the very core of what air combat is about, managing your chances. This is why you toss realism out of the window when you do not compensate that with skewing other parameters, maybe apart from their „actual“ values to give plausible bottom line.

 

Taken together, what I observe is that the current DM for FC makes shooting easier than intended by design. THIS, and only this had to be compensated for in order to get more plausible bottom line results. Ultimately, this is what everyone would want, right?

  • Upvote 2
Link to post
Share on other sites

Reading this thread I think it would be good to go over how loads are actually carried in a WW1 type airplane because if you don’t have that pegged down it’s difficult to say how the airplane should react to being hit and there seems to be some misconceptions about how this actually works.

 

I’m going to limit this to the bending loads (basically the g-load) because that is what has been discussed here. IRL you also have to handle torsional loads as well but for the purpose of explaining the fundamental difference between a braced and an un-braced WW1 design we can skip this.

 

Beginning with the un-braced designs like the Dr1 and D7: These carry the g-load  as a compression load in the solid wood elements in the upper corners and as a tension load in the lower wood elements of the box type spar. The Plywood is there to carry what is called a shear load which can be visualized as “pull” force in one 45 deg direction and as a “push” or compression load in the other (Not a totally true description of shear in beam theory but it helps in this description!). In structural design the plywood is called a shear web. However, if the shear web is really thin then it buckles in the compression direction and all shear load is carried in tension (As in Wagner beam theory). You can picture this as thousands of thin wires taking the shear load in tension in the 45 degree angle between the spars. In a Wagner beam you also need vertical elements to carry the compression load from the “wires” because otherwise the 45 degree shear forces would collapse the beams together. Note: In a Dr1 or D7 the shear web will not buckle and act like a Wagner beam but this picture is needed to understand how braced designs work.

 

So looking at the braced designs, how do they work? Well basically the braced design is a gigantic Wagner beam element with the upper wing being the upper spar and  the lower wing the lower spar and the wire bracing the shear web. So in a braced design you really can’t separate the spars from the wires when it comes to bearing the load because they work together. In addition, when seeing this picture you realize that the only wires needed (positive g-load!) are the ones going from the bottom fuselage to the upper wing and the struts which act as the vertical elements in a Wagner beam. So fundamentally, in a braced design, you are flying one gigantic Wagner beam. (For negative loads everything is of course reversed in terms of which spar carries compression and tension and which wires are needed).

 

This also highlights the importance of the wires I described above: If those go you basically lose all structural rigidity. Sure, I have seen the picture of the Albatros with one wing who made it down but I venture he flew very, very gingerly to a landing with a U-shaped upper wing praying it would hold together because that upper wing beam was designed to carry a pure compression load only and that it held up for any bending load was purely a fortuitous side effect nothing else.

 

Finally, bearing loads in tension in the spars is much easier than compression since you can basically utilize the tensile strength of the wood in the lower wing spar whereas the upper spar will as you increase the load ultimately fail by Euler beam buckling before the tensile strength of the wood is reached in the lower spar.

Edited by Holtzauge
  • Thanks 3
Link to post
Share on other sites

 

@Holtzauge, thanks for your explanation to make us notice how coarse the current appoximation  as DM model is, and what does for catering the Fokker D.VII and the Dr.I.

 

As the static construction that carried the whole load is much smaller on the Fokker wing than on braced wing designs, it conversely takes far fewer hits in the right place to make the wing collapse.

 

While in braced wing designs, the entire wing of the top wing spar is what carries (compression) load, in the Fokker it is just the upper "roofing" of the box spar, indicated by red circles here:

Spoiler

1.jpg.7dbdc1f647c94df0102bebdd7449cdd1.jpg

If anyone has a better blueprint, please share.

 

The D.VII has two box spars. The upper section is only mabe 3 or 4 cm thick, consting of longitudinally laminated wood. If you hit that section from the back ofthe aircraft, the wood will tend to buckle as it cracks apart under the pressure of the passing bullet. A slight crack there will have significant impact on the ability to receive push loads (as we have it in the Camel).

 

Conversely, the bottom part is the box is sensible as well, but I'd say wood is more resilient in push loads after damage then in push loads.

 

On the other hand, a large part of the box spar is semi immune to bullet damage, as the vertical front and back plates of the box will tolerate a lot of perforation until they cannot hold the "roofing" and the "flooring" of the box in place anymore.

 

In principle, the Fokker wing only differs from the braced wing in that the vital part that makes the wing go is a much smaller structure than the braced wing. Then again in case of our sim, I can bona fide say that this lower probability of hitting that part (for the ones that like RNG) is offset by it lower tolerance to bullet damage, as it is smaller. If I just add the size of the upper and the lower section of the box spar, I have a construction alomst like found in the Camel and same senibility to hit. Hence I could toss the assumption of high hit tolerance as we have it right out the window and replace it with what we have for the Camel. And I am as right as people who love what we have now.

 

What we as FC damage model is well meant, but it does not work for us like this. It still needs some love. We'll get there. But it's complicated. Assumptions have consequences. We are living them now.

Edited by ZachariasX
  • Thanks 1
  • Upvote 2
Link to post
Share on other sites

Exactly @ZachariasX: This whole thing about how to calculate damage is so complicated and a Dr1 or D7 design will also fail if a few bullets hit the right spot just like on a braced design. So unless you make a very detailed model with all nuts and bolts modeled, you are back to doing very general hit box counting. Anything in between will just result in that some planes will be unfairly strong and some unfairly weak.

 

Sure, there are advantages with the large spar box design on the Dr1 and D7: It has alternative load paths but OTOH it is so much bigger so will get hit more. The braced design spars just carry tension or compression loads and are so much smaller so will get hit less. A braced wing will fail if you hit that bracing wire but the chances of that happening? Minute. So there are pros and cons with both designs and it all comes down to WHERE did the bullet hit and we are back to a detailed model.

 

That was why I suggested a really simple DM: All aircraft start of with their design g-load capability. Then every hit in a box whittles away the same proportional strength in that box. This could of course be fine tuned somewhat but as a first order approximation I think it would be a much better solution. In addition, it would allow you to make a rough estimate of the g-load capability you have left. This would not be perfect since it would not capture the outliers but looking at the law of large numbers I think it would give the best approximation because IRL some unlucky dude would loose his wing to a single shot severing a bracing wire while the other guy comes home with hundreds of holes in the ribs and fabric. However, taking the averages, most would probably fall in the middle meaning their structural integrity was gradually ground down hit by hit.

Edited by Holtzauge
  • Like 1
  • Upvote 2
Link to post
Share on other sites
4 hours ago, ZachariasX said:

I cannot overstress how important the aspect of „managing your chances“ is. Especially back then. Especially when the chances are bad. These young lads, they would not have gone on missions like they did if they were told that some divine RNG was snuffing them after 8 missions. If they did, they would have been Kamikaze pilots and acted accordingly. But almost none did. Everyone thought that if he just kept a sharp eye, flew better, aimed better, they will make it. You may say, „poor lads, they don‘t know statistics“. But in fact survival rate is not random. It is heavily skewed towards certain individuals that learned to manage their chances. This includes knowing cues. You have to have those too. And it is not because he was selected by the divine RNG selected one with a higher survival chance, it is because the individual acted.

 

It is not bad having some random functios as we do not need to overcomplicate things. But it is bad not recognizing this fundamental aspect of air combat and genaral playability. Nobody wants to openly play lottery when starting the game, despite one can calculate chances ex post. If you turn it the other way around, you‘re not having a combat simulator anymore, but a lottery. Nobody would attend.

 

Right now in the game, we have an unrealistic setting: Optical damage does not really correspond to structural damage and hit effects are also by chance rather than just by ability. Both are fundamental game mechanics that are introduced artificially that work aginst the very core of what air combat is about, managing your chances. This is why you toss realism out of the window when you do not compensate that with skewing other parameters, maybe apart from their „actual“ values to give plausible bottom line.

 

As for survival being heavily skewed towards those who learned to manage their chances - well yes of course! Why on earth do you think that I am suggesting otherwise?   I have been a full time military officer - I know the value of training as a continuous process. 

 

Boelcke's maxims and Mannock's rules work just as well under the current DM as they do under the old one - perhaps better, since  combat is less forgiving, which makes tactical advantage - which mainly consists of not being shot at - even more important. If you fly like an idiot you will die sooner than later. Getting into a fight voluntarily without the advantages of numbers, height, etc is idiotic, albeit sometimes necessary or unavoidable.   We do it all the time because we can respawn or start a new career.  Managing your chances with this DM means being far more careful about picking your fights. Nothing intrinsically unrealistic about that.

 

War fighting is not chess.  Those WW1 (and WW2) pilots who survived many fights certainly learned to improve their chances: but they were also lucky. Their chances were still chances.

 

What I find fascinating, is that when it was proposed to reduce the number of wing-offs in the DM for BoS, we had complaints from number of MPlayers arguing that this would make MP more random and reduce the rewards for skill. If they got in the correct firing position and squeezed the trigger at the right time, they considered that they had "won" the engagement and should be rewarded by an appropriately disintegrating antagonist.   Now here you are arguing the opposite.  There is another thread for posts about MP balance. 

 

 As I have repeatedly said, I am not sure if the current effects of damage on G loads is right, either from an engineering standpoint, which I am not qualified to judge, or statistically. I do find many of the arguments that it is not realistic irrelevant, incoherent or just plain wrong, so I can understand why some people might have the impression that I am "defending the DM". ;)   What I have tried to do is to see if the DM may be giving some unanticipated results, leading to the  mismatch between player perception and AnP's idea of the statistical facts revealed by his tests. Complex models often do that.  

 

@Holtzauge As for alternate DMs etc: it is possible that all that needs to be tweaked - if such is required - is the properties of the weakest "spars".  There is no point completely redesigning the DM if the problem, if there is one, is mostly due to the weakest elements being broken too quickly. For instance, in the Camel case, assuming for the moment that my musings about the distributions are somewhat right, if the red spars were as strong as the blue spars, and consequently failed at the same rate, the observed game failure rate for the low hit numbers would be dramatically reduced, to far below AnP's "average" line, which as I have said, may be misleading.  

 

We can only make more progress on this analytically if we have more information from AnP - or perhaps if players were able to identify where their wings actually break most of the time.  Otherwise we are just trading opinions.

 

 

Edited by unreasonable
Link to post
Share on other sites
39 minutes ago, Holtzauge said:

Exactly @ZachariasX: This whole thing about how to calculate damage is so complicated and a Dr1 or D7 design will also fail if a few bullets hit the right spot just like on a braced design. So unless you make a very detailed model with all nuts and bolts modeled, you are back to doing very general hit box counting. Anything in between will just result in that some planes will be unfairly strong and some unfairly weak.

 

Sure, there are advantages with the large spar box design on the Dr1 and D7: It has alternative load paths but OTOH it is so much bigger so will get hit more. The braced design spars just carry tension or compression loads and are so much smaller so will get hit less. A braced wing will fail if you hit that bracing wire but the chances of that happening? Minute. So there are pros and cons with both designs and it all comes down to WHERE did the bullet hit and we are back to a detailed model.

 

That was why I suggested a really simple DM: All aircraft start of with their design g-load capability. Then every hit in a box whittles away the same proportional strength in that box. This could of course be fine tuned somewhat but as a first order approximation I think it would be a much better solution. In addition, it would allow you to make a rough estimate of the g-load capability you have left. This would not be perfect since it would not capture the outliers but looking at the law of large numbers I think it would give the best approximation because IRL some unlucky dude would loose his wing to a single shot severing a bracing wire while the other guy comes home with hundreds of holes in the ribs and fabric. However, taking the averages, most would probably fall in the middle meaning their structural integrity was gradually ground down hit by hit.

 

I have diminute knowkedge of all this, but I understand the Fokker DVII used internal bracing, which I see referred as a progress, rather than merely a variant. How does that differ from normal bracing, and did you consider this? Thanks!

Edited by J2_Bidu
Link to post
Share on other sites
  • LukeFF locked this topic
Guest
This topic is now closed to further replies.
×
×
  • Create New...