WW1 aircraft flight modeling as mentioned in the developers “Brief Room Episode 3” video

[Holtzauge]

In this video, the developers refer to (at about 5 and 30 minutes) a book that I have written, WW1 Aircraft Performance, and which I sent them a copy off free of charge already in February 2023. And in this video they express the viewpoint that my book is insufficiently referenced and that I don’t support my claims with data. To put it mildly, I do not agree with this take, since my book is heavily referenced, and all important steps in the flight modeling are supported by clearly referenced material. This claim therefore comes as a surprise to me, especially since I offered to answer any questions about the book already almost a year ago, and which I would have gladly provided had it been asked for. So I strongly protest any claim that my book is poorly supported by data and references, which is a claim I think is unfounded, especially seeing that I even include a dedicated chapter about validation, which I have attached below.

 

That having been said, one of the major takeaways from the book is that the turn rate modeling in it differs substantially from Flying Circus, which as far as I understand has many similarities (if not being identical?) to that in Rise of Flight. And in my opinion, the book shows that the Flying Circus flight modeling of turns has room for many improvements. Taking an example, in my book I present turn data for the Nieuport 28 C1 and Albatros D.Va, showing that the Nieuport should handily out-turn the Albatros at both 1 and 5 km altitude: Taking the turn times (complete 360 degree turns) at 1 km altitude from the WW1 Aircraft Performance book as an example:

 

Nieuport 28 C1: 13.6 s

 

Albatros D.Va: 15.3 s

 

Turn times measured in Il-2 and documented in this forum thread:

 

Nieuport 28 C1: 16.3 s

 

Albatros D.Va: 11.6 s

 

So without mincing words, a totally different result showing a huge difference, including a reverse order in how the aircraft compare to one another. But then who is right and who is wrong? Here are two flight models that come to two different conclusions? Is there some way to determine which model gives the more valid results?

 

Luckily there is. Because another forum user who’s handle is JtD, has make an Excel-based tool that can be used to make turn rate calculations, and for which I have posted an output example below. The observant reader will notice that I have used something called the effective span in the calculations, which is covered in detail in the book. But in summary, the effective span is a way to model a bi-plane by calculating what is called an effective aspect ratio. This is established aerodynamic theory, and was developed by Ludwig Prandtl and Max Munk  in the early parts of the last century. And for those so inclined, you can read more about it in NACA reports 116, 256, 317, and 151. Or you can see it in condensed form in Sighard Hoerner’s book Fluid dynamic Drag, pages 7-12 and 7-13. Anyway, entering this effective span into JtD’s Excel-based calculation model, we get the following results (See attached figure below):

 

Nieuport 28 C1: 12.9 s

 

Albatros D.Va: 14.05 s

 

Granted, this is also a different result from mine. However, note that the main difference is just that I’m slightly more pessimistic, and JtD a bit more optimistic: Since by simply subtracting 1 s from my results we get 12.6 and 14.3 s respectively, showing that there is a complete agreement on how these two aircraft compare in relation to one another, thus showing that the modeling in Flying Circus has these two aircraft’s relative performance completely backwards.

 

For those of you who are wondering how much to trust JtD’s Excel-based turn performance calculation method, you can always search this forum and find out, since he has posted profusely on the subject. However, I think this thread is a good starting point and gives a good idea about his level of knowledge and why I would refer to his calculation method to validate my results.

 

So summing up, we have two very different independent methods (C++ and Excel based) from two different sources (JtD’s and mine) indicating that the Nieuport 28 C1 should turn better, not worse than the Albatros D.Va, while the situation is as we know reversed in the current Il-2 Sturmovik Flying Circus flight modeling.

 

Q.E.D

 

PS: the reason for posting this in the FM part of the forum, is that it’s a claim that the current Il-2 Flying Circus turn modeling needs tuning, with the Nieuport 28 C1 and Albatros D.Va flight models being used as examples to support this claim.

Edited January 21, 2024 by Holtzauge

[Zooropa_Fly]

I think we all need to accept that there isn't going to be a huge amount of further ww1 development.

There's stuff to come obviously, but in a year or two FC will very much be going the way of RoF I think.

They've got new stuff on the go and they're cracking on with that.

The interest in ww1 is very low, and I don't think a slightly better turning N28 will change that, even if one is to believe a couple of spreadsheets can determine how 100+ year old crates performed.

The new game is probably required at this stage, because I sense ww2 here is tailing off a bit too.

 

At the end of the day, the Devs have to make decisions about what to do to be able to make money and keep going.

[BraveSirRobin]

Starting a pissing contest in the forum instead of contacting the developer is definitely the best way to handle this. 

[ZachariasX]

12 minutes ago, BraveSirRobin said:

Starting a pissing contest in the forum instead of contacting the developer is definitely the best way to handle this. 

Has been done. Not that I would expect anything though either way. The message was very clear in the Q&A.

 

So...

[BraveSirRobin]

Just now, ZachariasX said:

Has been done. Not that I would expect anything though either way. The message was very clear in the Q&A.

 

So...

They quite clearly said in the video that they need more info from the author of the book.  But I’m sure that a pissing contest thread in the forum will work just as well.

[ZachariasX]

22 minutes ago, BraveSirRobin said:

They quite clearly said in the video that they need more info from the author of the book.

These are not Americans talking.

 

What happened is that somebody told a senior expert in his field of expertise that he doesn't know how to write a paper. Now, connect the dots from there.

[BraveSirRobin]

11 minutes ago, ZachariasX said:

These are not Americans talking.

 

What happened is that somebody told a senior expert in his field of expertise that he doesn't know how to write a paper. Now, connect the dots from there.

What does the fact that they’re not American have to do with anything?   They said that they need more info.  Maybe try giving them more info before going the pissing contest route.

Edit…

Sorry for the crazy talk.  Pissing contest thread was definitely the best option.

Edited January 21, 2024 by BraveSirRobin

[LukeFF]

Guys, I'm only going to say this once: either stay on-topic or don't post at all.

[US103_Baer]

Responding to the devs reply with examples and data seems perfectly valid. As is making sure that everyone knows the backstory re the authors approaches to Dev team. 

 

A 6.4s reversal in Sustained Turn time between two adversaries at 1k isn't a minor change!

It completely alters the scenario between them and turns the N.28 from unusable to actually competitive. Its like getting a new plane. I mean, who flies it now?  

 

Although Alb v N28 is a more extreme example, FM disparity has created mismatches across the whole planeset. The flow-down effect of which is well known to all of us here.

 

Is fine to be realistic on possibility of this work getting done as soon as we'd like. But that doesn't mean agreeing that status quo is ok, which was the response that came across from the Devs. 

Actually,  that's not quite correct is it. Without re-watching,  I think they said that IF appropriate data was provided that contradicted their models, then they'd look at the possibility of fixes.

Which rather justifies the authors response. 

[No.23_Starling]

4 hours ago, US103_Baer said:

Responding to the devs reply with examples and data seems perfectly valid. As is making sure that everyone knows the backstory re the authors approaches to Dev team. 

 

A 6.4s reversal in Sustained Turn time between two adversaries at 1k isn't a minor change!

It completely alters the scenario between them and turns the N.28 from unusable to actually competitive. Its like getting a new plane. I mean, who flies it now?  

 

Although Alb v N28 is a more extreme example, FM disparity has created mismatches across the whole planeset. The flow-down effect of which is well known to all of us here.

 

Is fine to be realistic on possibility of this work getting done as soon as we'd like. But that doesn't mean agreeing that status quo is ok, which was the response that came across from the Devs. 

Actually,  that's not quite correct is it. Without re-watching,  I think they said that IF appropriate data was provided that contradicted their models, then they'd look at the possibility of fixes.

Which rather justifies the authors response. 

Well said. The next step would seem to be a direct conversation between Anders and the 1C team. 
 

This would give more of an opportunity for Mr Jonsson to explain both the engineering principles used & how they factor into the C++, and to provide further explanation to the primary sources used to tune the datasets beyond what’s in the book and above.

Edited January 22, 2024 by US103_Rummell

[Zooropa_Fly]

Nothing needs to happen.

[No.23_Starling]

2 minutes ago, Zooropa_Fly said:

Nothing needs to happen.

Edited.

[Avimimus]

I'm confused...

 

Does this excel method take into account airfoil shape at all?

[Holtzauge]

5 hours ago, US103_Rummell said:

Well said. The next step would seem to be a direct conversation between Anders and the 1C team. 
 

This would give more of an opportunity for Mr Jonsson to explain both the engineering principles used & how they factor into the C++, and to provide further explanation to the primary sources used to tune the datasets beyond what’s in the book and above.

 

For sure, I would be happy to discuss FM modeling with them. I plan to post some ideas about this later.

 

59 minutes ago, Avimimus said:

I'm confused...

 

Does this excel method take into account airfoil shape at all?

 

Indirectly, since this is captured by entering the stall speed value in the input table. If you reverse engineer this it gives you a Clmax on aircraft level which is the governing factor for turns.

 

 

Edited January 22, 2024 by Holtzauge

[Holtzauge]

Having thought some more about this, I think the flight model issues we see in-game may stem from the way 1C does the tuning of their FM’s: Based on the way the planes in-game behave in relation to one another, the tuning seems to be based on individual polar diagrams (i.e. drag versus lift coefficient Cl/Cd) for each individual aircraft. At least that’s my impression.  And if this assumption is correct, then this means that they either need to find this type of data from historical sources, or make their own estimates.

 

But the accuracy of my C++ model is not dependent on me finding data for each individual aircraft. Because my model is rooted in physics, and when I make new aircraft models, I basically only enter the aircraft’s physical characteristics into the simulation, and out comes the results.  And where I can find solid data to compare with (which is mostly for WW2 aircraft), the results are usually very close. 

 

However, basic physics naturally also applies to WW1 aircraft, and the only leap of faith that is required to also model these aircraft accurately, is to grasp and embrace the aerodynamic concept of the effective aspect ratio which I cover in detail in my book. And this is why I mean that the developers should trust my models: Because they are rooted in physics and not arbitrarily conceived.

 

And since the validation I’ve done shows that the model tabs well with the historical turn data we have for the Siemens Schukert D.IV, from this follows that we should be able to trust it also to accurately model other WW1 aircraft as well. So we don’t really need to find or estimate data for each individual aircraft. We can instead trust the physics model to do that for us.  Otherwise, one would need to make a really compelling argument for why Ludwig Prandtl’s and Max Munk’s ideas on the aerodynamic properties of stacked wings are inaccurate, which would be chasing after the wind IMHO.

 

Summing up, I think there is a good way out of all this after all: Because just as 1C benefits from tapping into forum community competence for campaigns and skins etc., they could also make use of its aeronautical and engineering competence.  Since at the end of the day, it’s obvious that we all share the same passion for WW1 aircraft, and if 1C’s management so chooses, I am perfectly willing to engage in FM discussions with their engineers.

 

PS: And ending on a constructive note: If the 1C method of tuning their WW1 FM’s is based on Cl/Cd polars for each individual aircraft as I believe it is, then I could reverse engineer those from my physics based models to aid in their in-game tuning.

 

 

Edited January 22, 2024 by Holtzauge

[Gavrick]

6 hours ago, Holtzauge said:

tuning seems to be based on individual polar diagrams

Correct (in general).

In general, there is an almost two-hour video on YouTube in which I, in particular, talk about how FM works and is created. But it’s in Russian, I’m afraid that the auto-translation in subtitles may be somewhat unclear.

[Holtzauge]

12 minutes ago, Gavrick said:

Correct (in general).

In general, there is an almost two-hour video on YouTube in which I, in particular, talk about how FM works and is created. But it’s in Russian, I’m afraid that the auto-translation in subtitles may be somewhat unclear.

 

OK, good to know. And this is where I think the physics model I have which is based on Prandtl's and Munk's effective aspect ratio model could come in handy for you because I think it could do a good job as a base to generate polar type input data to your WW1 biplane models.

 

BTW: Could you link the video? Even if the auto-translate is sketchy I think it would be very interesting to watch anyway. 

[354thFG_Panda_]

58 minutes ago, Holtzauge said:

 

OK, good to know. And this is where I think the physics model I have which is based on Prandtl's and Munk's effective aspect ratio model could come in handy for you because I think it could do a good job as a base to generate polar type input data to your WW1 biplane models.

 

BTW: Could you link the video? Even if the auto-translate is sketchy I think it would be very interesting to watch anyway. 

 

[Han]

Hello guys.

 

A short statement here.

 

Ive checked in video - I've told there only, that at the moment book does not contain all sources and all math basements for conclusions were made there. When it will - we will consider it. Nothing more.

 

So let's expand this statement:

 

I've not told that there is something wrong or right in the book. I've told that book do not uncover all it's math basis and historical data comparison used for it's verification.

 

WW2 flight models are ok because they have enough source data to be verified.

 

But WW1 still have at least some data and it was used that times 15 years ago when RoF FMs were developed. Some planes have measurement data of their performance, some planes have comparative data to other planes. 

 

So switch from strategy where at least some kind of historical data was used to strategy where only "theoretical evaluation" is used for FM adjustment - we don't like this idea.

 

So book/article (or some kind of additional appendix to it) should contain full math basis how results were obtained and historical research and comparison part where those results are compared to ALL known historical data. In this comparison part there should be strong argumentation why some historical data is not matching and should be dropped off (for instance - other historical data (controversal to this one) is better matching the model model, and there are much more historical data matching to the model rather than not).

 

In this case - the method will be proven as good-verifyed and conclusions about airplanes performance estimation can be taken as strong proposal to be used in FM adjustment.

 

So no "trust", only "proof and verification" - this how we worked all these years.

 

I hope you guys see that here is nothing personal in this scientific approach.

 

PS

Sorry, but still can't promise that everything will be changed in a day after this appendix will be done (and it will proof the book theory in scientific meaning) - FC idea author and initial producer haven't included any changes of FM to FC business-plan which we're still following in Great Battles (in part of FC), he stated this publically before FC release and it is fair enough I'm sure. Plus it was only one realistic way to make WW1 theme live and evolve again - so nothing can be told against this decision. So you know -  I do not do any promises which I'm not sure by myself (sometimes it hurts the business because some lie is kind of norm in gamedev, but I'm not using this hack). But I can say that we allways trying to find some time for attention and fixing old issues. Our engineers team don't like to have such issues in the trunk.

 

So once appendix will be done and it will realy sintifically proof the conclusions of the article - there is a good chance that some (most critical) mismatches will be changed. But once again: no "trust", only "proof and verification" - this how we worked all these years.

[No.23_Starling]

1 hour ago, Han said:

Hello guys.

 

A short statement here.

 

Ive checked in video - I've told there only, that at the moment book does not contain all sources and all math basements for conclusions were made there. When it will - we will consider it. Nothing more.

 

So let's expand this statement:

 

I've not told that there is something wrong or right in the book. I've told that book do not uncover all it's math basis and historical data comparison used for it's verification.

 

WW2 flight models are ok because they have enough source data to be verified.

 

But WW1 still have at least some data and it was used that times 15 years ago when RoF FMs were developed. Some planes have measurement data of their performance, some planes have comparative data to other planes. 

 

So switch from strategy where at least some kind of historical data was used to strategy where only "theoretical evaluation" is used for FM adjustment - we don't like this idea.

 

So book/article (or some kind of additional appendix to it) should contain full math basis how results were obtained and historical research and comparison part where those results are compared to ALL known historical data. In this comparison part there should be strong argumentation why some historical data is not matching and should be dropped off (for instance - other historical data (controversal to this one) is better matching the model model, and there are much more historical data matching to the model rather than not).

 

In this case - the method will be proven as good-verifyed and conclusions about airplanes performance estimation can be taken as strong proposal to be used in FM adjustment.

 

So no "trust", only "proof and verification" - this how we worked all these years.

 

I hope you guys see that here is nothing personal in this scientific approach.

 

PS

Sorry, but still can't promise that everything will be changed in a day after this appendix will be done (and it will proof the book theory in scientific meaning) - FC idea author and initial producer haven't included any changes of FM to FC business-plan which we're still following in Great Battles (in part of FC), he stated this publically before FC release and it is fair enough I'm sure. Plus it was only one realistic way to make WW1 theme live and evolve again - so nothing can be told against this decision. So you know -  I do not do any promises which I'm not sure by myself (sometimes it hurts the business because some lie is kind of norm in gamedev, but I'm not using this hack). But I can say that we allways trying to find some time for attention and fixing old issues. Our engineers team don't like to have such issues in the trunk.

 

So once appendix will be done and it will realy sintifically proof the conclusions of the article - there is a good chance that some (most critical) mismatches will be changed. But once again: no "trust", only "proof and verification" - this how we worked all these years.

Thanks Han. In terms of historical test data on turn rates there’s basically none so that’s where Anders’ approach fills some gaps. He references all the original sources used to tune the model so it shouldn’t be a huge issue sharing those.

 

I suspect 777/1C will be sitting on a ton of speed/climb sources too. Excellent to see this level of engagement.

[Holtzauge]

Thanks Han, for taking the time to expand on how you see things here in the forum and in our PM thread. And based on this new info, it's easier to get a better picture of what your position is and what more you would like to see in order to contemplate changing the flight models.

 

So in order to move things forward and make it easier for you to evaluate and review the core parts of the C++ physics model, I will extract those and put these into an Excel version that will be easier and less time consuming for you and your engineers to review.

 

I will send links to those spreadsheets later on when I have them ready in the PM thread we have going.

[Gavrick]

Once again, I'll make another important point that I really want to bring to your attention.
The main problem now is time. The time it takes to collect the data (from wherever and whatever they are), and the time it takes to make corrections to the model.
That is, let's say we have all the calculations and data in hand, but we need to comprehend them, compare them, check them, prepare data for the model based on them, and so on.
There are a lot of tasks and not enough time. That's the critical issue right now.

[Han]

That is what my "PS" is about ))

[Holtzauge]

1 hour ago, Gavrick said:

Once again, I'll make another important point that I really want to bring to your attention.
The main problem now is time. The time it takes to collect the data (from wherever and whatever they are), and the time it takes to make corrections to the model.
That is, let's say we have all the calculations and data in hand, but we need to comprehend them, compare them, check them, prepare data for the model based on them, and so on.
There are a lot of tasks and not enough time. That's the critical issue right now.

 

1 hour ago, Han said:

That is what my "PS" is about ))

 

Absolutely! I get that you are busy with a lot of other stuff as well which is why I am putting together a condensed version of the calculations in an Excel spreadsheet which I will PM to you. Then when it comes to the input data needed for each aircraft, I think you will see that it's all readily available already so if you like the theoretical model, generating the Cl/Cd polars should be possible to do more or less automatically.

 

 

Edited January 23, 2024 by Holtzauge

[Avimimus]

3 hours ago, Gavrick said:

There are a lot of tasks and not enough time. That's the critical issue right now.

2 hours ago, Han said:

That is what my "PS" is about ))

 

My thoughts:

 

WWI has a dedicated fanbase which doesn't really compete with Korea (the aircraft are different enough even from WWII that the decision to buy is usually separate). While market for WWI may be smaller, it could sustain a low pace of development (as we saw with Rise of Flight).

 

Importantly, with no significant competitors, any limitations in the current engine are unlikely to be overshadowed, so updates don't need to be as extensive as other modules.

 

So partnering with a 3rd party to introduce small amounts of new content, and routinely scheduling a small amount of programming time, could probably keep this alive as a second product line and source of revenue (people are more likely to purchase a product which has at least small amounts of ongoing support).

 

P.S. I sincerely hope that eventually the capacity of the team will reach a point where the above arguments make sense.

[Zooropa_Fly]

I've not got it, but by all accounts WOFF has got the offline experience pretty much sewn up.

They have to be considered a competitor ?

 

At the end of the day, anything has to be financially viable, so what the future holds for ww1 here will likely depend on the sales of the next few releases.

[RNAS10_Mitchell]

5 hours ago, Gavrick said:

Once again, I'll make another important point that I really want to bring to your attention.
The main problem now is time. The time it takes to collect the data (from wherever and whatever they are), and the time it takes to make corrections to the model.
That is, let's say we have all the calculations and data in hand, but we need to comprehend them, compare them, check them, prepare data for the model based on them, and so on.
There are a lot of tasks and not enough time. That's the critical issue right now.

Any chance that FC could be ported to the new environment/game/engine in the somewhat near future?

[Holtzauge]

Status update: Attached are parts of the data I have sent to the developers (via PM) to address their expressed concern that they feel they can’t act on the data presented in my book. In summary, this is because they (Daniil and Viktor) said would like to verify it even further than what is done in the book already (See OP for details).

 

My first suggestion to address this was that they could do it like you would do it for any other complex model (and as we all have been doing the in-game testing of the Il-2 Sturmovik FM’s!):  That is you test it by defining certain key test cases, test those, and see if it produces satisfactory output.

 

Their response to this was that while yes, this is certainly a possibility, they thought this was too time and resource consuming. And in some sense I can sympathize with this view, since you do need to define a number of test cases and evaluate them, but I still think this is well within the reasonable, especially given I have been willing to run and compile all the tests they wanted to throw at it myself.

 

But since this was a no-no, I have now extracted the core part of the physics part in the C++ model, which is Ludwig Prandtl’s and Max Munk’s theory of multiplanes (See attachments) into a simplified Excel spreadsheet in which it is very easy to follow each step individually and verify the validity of the concept. This will of course not be perfect, since it’s a very simplified model, and does not capture all of the complexity included in the C++ model. However, I was happy to see that it’s still very well aligned with my C++ simulation results on the whole, thus once and for all establishing that some of the in-game FM’s are in dire need of a rework. I say this because each and every step in the Excel spreadsheet is easy to follow and rests on established aerodynamic theory.

 

Interestingly enough, note that the turn time for the Excel model of the Fokker D.VIII is spot on compared to the in-game measured turn time at 1 km altitude, thus showing that it’s well aligned with how single winged aircraft are modelled in Il-2 Sturmovik. So again, the only step on top of this needed to fix the Flying Circus FM’s, is to embrace Ludwig Prandtl’s and Max Munk’s effective aspect ratio theory as used in the attached Excel model outputs.

 

I have sent all this data (and more) via a PM thread which now includes a number of the FM engineers in addition to Han and Viktor but I am still waiting for an answer. And since this has drawn out in time, I thought it would be good to also let the forum community know how things stand as of today.

 

 

 

 

 

 

 

 

 

 

 

[No.23_Starling]

5 hours ago, Holtzauge said:

Status update: Attached are parts of the data I have sent to the developers (via PM) to address their expressed concern that they feel they can’t act on the data presented in my book. In summary, this is because they (Daniil and Viktor) said would like to verify it even further than what is done in the book already (See OP for details).

 

My first suggestion to address this was that they could do it like you would do it for any other complex model (and as we all have been doing the in-game testing of the Il-2 Sturmovik FM’s!):  That is you test it by defining certain key test cases, test those, and see if it produces satisfactory output.

 

Their response to this was that while yes, this is certainly a possibility, they thought this was too time and resource consuming. And in some sense I can sympathize with this view, since you do need to define a number of test cases and evaluate them, but I still think this is well within the reasonable, especially given I have been willing to run and compile all the tests they wanted to throw at it myself.

 

But since this was a no-no, I have now extracted the core part of the physics part in the C++ model, which is Ludwig Prandtl’s and Max Munk’s theory of multiplanes (See attachments) into a simplified Excel spreadsheet in which it is very easy to follow each step individually and verify the validity of the concept. This will of course not be perfect, since it’s a very simplified model, and does not capture all of the complexity included in the C++ model. However, I was happy to see that it’s still very well aligned with my C++ simulation results on the whole, thus once and for all establishing that some of the in-game FM’s are in dire need of a rework. I say this because each and every step in the Excel spreadsheet is easy to follow and rests on established aerodynamic theory.

 

Interestingly enough, note that the turn time for the Excel model of the Fokker D.VIII is spot on compared to the in-game measured turn time at 1 km altitude, thus showing that it’s well aligned with how single winged aircraft are modelled in Il-2 Sturmovik. So again, the only step on top of this needed to fix the Flying Circus FM’s, is to embrace Ludwig Prandtl’s and Max Munk’s effective aspect ratio theory as used in the attached Excel model outputs.

 

I have sent all this data (and more) via a PM thread which now includes a number of the FM engineers in addition to Han and Viktor but I am still waiting for an answer. And since this has drawn out in time, I thought it would be good to also let the forum community know how things stand as of today.

 

 

 

 

 

 

 

 

 

 

 

Fantastic work!

 

Just to check, the DXII model is the BMW version (same as in game)?

[VonS]

6 hours ago, Holtzauge said:

[...] note that the turn time for the Excel model of the Fokker D.VIII is spot on compared to the in-game measured turn time at 1 km altitude, thus showing that it’s well aligned with how single winged aircraft are modelled in Il-2 Sturmovik. [...]
 

 

Many thanks for these stimulating technical posts Holtzauge; hopefully they will result in more historical FMs in the FC franchise. Interesting point there about monoplane FM accuracy in IL2; have noticed a similar thing in the free FlightGear, with solid FM modeling of single-wing, tricycle-geared (small) civilian aircraft, as well as of large airliners - but questionable results with multi-plane FMs (such as for the Gloster Gladiator) - with excessive "floatiness" and a much-too-pleasant glide slope with engine off. Perhaps not all sim game/physics models are well suited to (simulation of) multi-plane contraptions that many of us enjoy. Anyway, interesting stuff - will not bore anyone with Yasim vs. JSBsim FM debates à la FlightGear.

 

Cheers & good flying,

?

[ZachariasX]

14 hours ago, Holtzauge said:

Interestingly enough, note that the turn time for the Excel model of the Fokker D.VIII is spot on compared to the in-game measured turn time at 1 km altitude, thus showing that it’s well aligned with how single winged aircraft are modelled in Il-2 Sturmovik. So again, the only step on top of this needed to fix the Flying Circus FM’s, is to embrace Ludwig Prandtl’s and Max Munk’s effective aspect ratio theory as used in the attached Excel model outputs.

 

I don't even remember how long I keep harping about the induced drag of the biplanes in RoF/FC. While it was clear that in the GB series, the sim made great progress, yet the biplanes still remained at times questionable. Your chart is really instructive for where to look for an underlying issue. ?

[Holtzauge]

10 hours ago, US103_Rummell said:

Fantastic work!

 

Just to check, the DXII model is the BMW version (same as in game)?

 

Yes, I have calculated the in-game BMW powered variant. The reason I included it, was because I have so far presented data that would result in two Entente scouts getting a significant up-lift in turn performance if all this was implemented, while the Albatros would lose the only performance advantage it has in-game, basically turning it into a lame duck.  But all the evidence points to that being the case IRL as well, so not much we can do about it. IIRC then this is also what Kermit Weeks is saying about his replica.

 

So this was why I thought it would be good idea to point out that also a Central scout, the Pfalz D.XII, would see a big improvement in performance if the FM’s are reworked along the lines suggested in my book.

 

9 hours ago, VonS said:

 

Many thanks for these stimulating technical posts Holtzauge; hopefully they will result in more historical FMs in the FC franchise. Interesting point there about monoplane FM accuracy in IL2; have noticed a similar thing in the free FlightGear, with solid FM modeling of single-wing, tricycle-geared (small) civilian aircraft, as well as of large airliners - but questionable results with multi-plane FMs (such as for the Gloster Gladiator) - with excessive "floatiness" and a much-too-pleasant glide slope with engine off. Perhaps not all sim game/physics models are well suited to (simulation of) multi-plane contraptions that many of us enjoy. Anyway, interesting stuff - will not bore anyone with Yasim vs. JSBsim FM debates à la FlightGear.

 

Cheers & good flying,

?

 

1 hour ago, ZachariasX said:

 

I don't even remember how long I keep harping about the induced drag of the biplanes in RoF/FC. While it was clear that in the GB series, the sim made great progress, yet the biplanes still remained at times questionable. Your chart is really instructive for where to look for an underlying issue. ?

 

Well, the way stacked wings work (or really more appropriately said don’t work! ) together is complex and I learnt a lot about this working on my book since bi-plane theory was for obvious reasons not what they taught us when I studied aerodynamics at university. It was of course more about compressibility, transonic and supersonic lift and drag and the like. So “rediscovering” how bi-planes work is probably an on-going process for most these days I would think!

 

[giftgruen]

This is absolutely great work, dear Holtzauge, interesting to read and hopefully will help the DEV team to make the sim better in terms of FM accuracy.

 

Regarding the "Albatros DVa will get a lame duck topic"

 

If we have a look e.g. to the kill list of MvR we can see that most victims are 2-seaters and not entente fighters.

This changes somewhat after MvR started to fly the Fokker DrI, but at least when flying the Albatros he seemed not to focus on hunting down enemy fighters.

https://en.wikipedia.org/wiki/List_of_aerial_victories_of_Manfred_von_Richthofen

 

 

What I want to express: 

 

The Albatros was ( afaik ) one of the first fighter wearing two Spandaus and therefore having double firepower and the weight maybe also helped to serve as a stable gun platform. 

 

We could maybe compare the Alb more to the Fw190 In IL-2 WW2. The 190 does not really shine in dogfights ( imho a 109 is superiour in 1vs1 duells ) but as a gun platform it's a great plane and hunting IL2s or Bombers in a Fw190 is quite a good thing.

 

If a fighter plane is a good plane or not - depends on more than just turn or climb rates.

 

So - if historically accurate - having a inferiour Albatros in terms of pure dogfighting performance would be completely okay - and does not mean that there is no fun and or purpose to fly it.

 

The real problem then is, that we have not enough suitable 2-seater victims ( BE2c escpecially, 1 1/2 Strutter  .... ) and the fact that Il2 WW1 is more or less implemented as a pure dogfighting game ( in terms of mission building and planeset ) is maybe the even bigger "historical bug".

 

 

Edited February 2, 2024 by giftgruen

[Holtzauge]

On 2/2/2024 at 4:46 PM, giftgruen said:

This is absolutely great work, dear Holtzauge, interesting to read and hopefully will help the DEV team to make the sim better in terms of FM accuracy.

 

Regarding the "Albatros DVa will get a lame duck topic"

 

If we have a look e.g. to the kill list of MvR we can see that most victims are 2-seaters and not entente fighters.

This changes somewhat after MvR started to fly the Fokker DrI, but at least when flying the Albatros he seemed not to focus on hunting down enemy fighters.

https://en.wikipedia.org/wiki/List_of_aerial_victories_of_Manfred_von_Richthofen

 

 

What I want to express: 

 

The Albatros was ( afaik ) one of the first fighter wearing two Spandaus and therefore having double firepower and the weight maybe also helped to serve as a stable gun platform. 

 

We could maybe compare the Alb more to the Fw190 In IL-2 WW2. The 190 does not really shine in dogfights ( imho a 109 is superiour in 1vs1 duells ) but as a gun platform it's a great plane and hunting IL2s or Bombers in a Fw190 is quite a good thing.

 

If a fighter plane is a good plane or not - depends on more than just turn or climb rates.

 

So - if historically accurate - having a inferiour Albatros in terms of pure dogfighting performance would be completely okay - and does not mean that there is no fun and or purpose to fly it.

 

The real problem then is, that we have not enough suitable 2-seater victims ( BE2c escpecially, 1 1/2 Strutter  .... ) and the fact that Il2 WW1 is more or less implemented as a pure dogfighting game ( in terms of mission building and planeset ) is maybe the even bigger "historical bug".

 

 

 

Well AFAIK early on the first Albatros versions could hold their own but it was basically always a design that was too heavy for the engines they put in it. Not even the over-compressed Mercedes DIIIäu could change that so even if we get that engine option in-game eventually, it will still not provide the quantum leap the Alby needs. However, if a FM review is done, maybe from a Central perspective the Albatros’ loss will to some extent be offset by the gain for the Pfalz D.XII since this scout will see quite a boost in turn performance.

 

Then when it comes to actual usage, teamwork, numbers and disciplined flying can off course compensate to some extent for lackluster performance. In fact I’ve seen some very interesting MP videos where a few Alby’s take on a smaller number of enemy scouts, and by popping in and out of the fray can maintain an altitude advantage, and then by keeping a constant pressure on the opposition force them downwards and eventually finish them off. How sporting the guys flying the Camels and what have you find this is of course questionable, but it’s certainly effective and shows that even a “lame duck” like the Alby can kill you provided you have the numbers stacked against you. And that kind of teamwork was probably also used back in the days as well. However, what they did not have back then was of course the instant communication via discord enabling the seamless hand off between different scouts to keep up the pressure.

 

Another big difference between now and then is that in FC all planes are flown by gorillas able to roll effortlessly at high IAS which is not very realistic at all is it? Since IRL many WW1 scouts get their controls set in concrete when you speed them up. So the type of BnZ attacks you can get away with in FC would have been a big no-no IRL due to the excessive stick forces. But modelling stick forces correctly is probably an order of magnitude more complicated than getting the speed, climb and turn rate sorted out, so that can of worms is probably best left unopened. 

[BMA_Hellbender]

I can't believe I completely missed this thread. Great work so far, @Holtzauge, and hopeful news from the deveteam.

 

On 2/3/2024 at 5:00 PM, Holtzauge said:

 

Well AFAIK early on the first Albatros versions could hold their own but it was basically always a design that was too heavy for the engines they put in it. Not even the over-compressed Mercedes DIIIäu could change that so even if we get that engine option in-game eventually, it will still not provide the quantum leap the Alby needs. However, if a FM review is done, maybe from a Central perspective the Albatros’ loss will to some extent be offset by the gain for the Pfalz D.XII since this scout will see quite a boost in turn performance.

 

The Albatros needs better speed and climb to be able to fulfill its historical role as from 1917 onwards: diving on "customers entering the store". It can technically speaking already do that, since it won't necessarily be going any faster in a dive than it does now after an FM update, but to get in a high energy position is an exercise in sheer frustration today.

 

As for its offensive and dogfighting capabilities? None. Zero. The D.Va, especially, was a pig. MvR said so himself.

 

Dogfighting is a task left for elite pilots flying elite machines such as the Fokker Dr.I and the Halberstadt D.II. Yes, you read that right: the Halberstadt D.II. Yes, that one, from 1916.

 

And before bitter tears start flowing on Central, by the time the Fokker D.VII was available with its standard D.IIIaü 200hp engine, the balance had completely shifted again. Sadly MvR just didn't live to see it.

[J2_Trupobaw]

On 2/3/2024 at 8:00 AM, Holtzauge said:

Well AFAIK early on the first Albatros versions could hold their own but it was basically always a design that was too heavy for the engines they put in it. Not even the over-compressed Mercedes DIIIäu could change that so even if we get that engine option in-game eventually, it will still not provide the quantum leap the Alby needs. However, if a FM review is done, maybe from a Central perspective the Albatros’ loss will to some extent be offset by the gain for the Pfalz D.XII since this scout will see quite a boost in turn performance.

What about Oeffag. D.IIIs which, with 225 Austro-Daimpler, annegdotically held their own till the end of the war? They don't apply to the western front situation, but they do demonstrate what the airframe was potentially capable of if properly retrofitted.

I am still somewhat suspitious about the impact of wires in the modelling - The Albatros / Fokker planes, with much fewer wires between wings (and interferences between these wires) may have a hidden drag advantage - or rather, the heavily wired Entente planes may be producing much more induced drag than airfoil analysis would suggest.  With different wire profiles and drag changing with speed any comparison of wiring impacting two planes sounds like a headache, but I take it density of wires has drastic impact on drag, and always believed that minimal wiring was one of reasons why German planes performed comparatively better than what bare power / weight / airfoil values would suggest.

[Holtzauge]

On 3/25/2024 at 4:34 PM, J2_Trupobaw said:

What about Oeffag. D.IIIs which, with 225 Austro-Daimpler, annegdotically held their own till the end of the war? They don't apply to the western front situation, but they do demonstrate what the airframe was potentially capable of if properly retrofitted.

I am still somewhat suspitious about the impact of wires in the modelling - The Albatros / Fokker planes, with much fewer wires between wings (and interferences between these wires) may have a hidden drag advantage - or rather, the heavily wired Entente planes may be producing much more induced drag than airfoil analysis would suggest.  With different wire profiles and drag changing with speed any comparison of wiring impacting two planes sounds like a headache, but I take it density of wires has drastic impact on drag, and always believed that minimal wiring was one of reasons why German planes performed comparatively better than what bare power / weight / airfoil values would suggest.

 

I have not looked at the 225 Daimler (Can you post a power versus altitude chart for the engine you have in mind?) but in order to give the Albatros the quantum leap it needs (due to its poor P/W ratio) the engine would have to be really good and provide much better altitude performance than even the Mercedes D.IIaü because even that engine is not enough to bring it back to being competitive against for example the S.E.5a and SPAD XIII.

 

When it comes to the drag of wires, the profiled solid rods or "RAF-wires" as they are known, were really good in reducing drag while the round wires the Germans used were bad and had caused a lot of drag. We had a big discussion about the Pfalz D.XII's performance over at The Aerodrome forum where this was brought up that you can find here:

 

https://www.theaerodrome.com/forum/showthread.php?t=75796&highlight=wire

 

[NO.20_Krispy_Duck]

Your charts and excel data are excellent. Making FC the best it can be is important for the multiplayer aspect. WOFF single player is outstanding (the single player campaign is an achievement there), but they have no multiplayer. FC's niche is the WW1 MP environment, which has a long tradition going back to INN and Red Baron days, then WON RB2/3d, RS, OFF beta, ROF, and so forth. Improvements are always welcomed.

Edited March 27, 2024 by NO.20_Krispy_Duck

[Todt_Von_Oben]

Holtzauge, 

 

In the latest Dev Blog #359, Albert Zhiltzov and Daniel Tuseyev said the following:

 

“…we need an artificial intelligence programmer who is familiar with navigation in three-dimensional physics-based space. If you exist, can write in C++, and want to solve problems that probably only the aerospace industry solves except for us, please write and come — we are very much waiting for you, we have a terrible shortage of personnel for this task.”

 

If you are the kind of C++ programmer they describe, maybe you could offer them your services?   

[Holtzauge]

6 hours ago, NO.20_Krispy_Duck said:

Your charts and excel data are excellent. Making FC the best it can be is important for the multiplayer aspect. WOFF single player is outstanding (the single player campaign is an achievement there), but they have no multiplayer. FC's niche is the WW1 MP environment, which has a long tradition going back to INN and Red Baron days, then WON RB2/3d, RS, OFF beta, ROF, and so forth. Improvements are always welcomed.

 

Even though the WOFF FM is a bit gamey, I think it’s still a very enjoyable game due to the completeness of it in terms of environment, career and the way AI works. And in WOFF the FM’s are not as important for a good SP experience since one can enjoy it as the game it is. OTOH, if they were to fix the FM’s in it, I would shelve FC (in its current shape) in a heartbeat since other than the FM’s I think it’s superb. FC OTOH targets the simulator niche so I think we are right to hold it to higher standards. In addition, I believe their stated aim with it is to be within 5% of the known data as well?

 

Pristine skins with gleaming aluminium and all the rivets in the right place on WW2 kites or immaculate wood graining and leather stitching details on the WW1 crates all very much take a back seat compared to the FM’s IMHO. So to me it was bad news indeed that Albert and Daniel talked so much about attention to detail and graphics and even MORE visual improvements but said absolutely nothing about FC FM’s in the last #359 dev blog…….

 

2 hours ago, Todt_Von_Oben said:

Holtzauge, 

 

In the latest Dev Blog #359, Albert Zhiltzov and Daniel Tuseyev said the following:

 

“…we need an artificial intelligence programmer who is familiar with navigation in three-dimensional physics-based space. If you exist, can write in C++, and want to solve problems that probably only the aerospace industry solves except for us, please write and come — we are very much waiting for you, we have a terrible shortage of personnel for this task.”

 

If you are the kind of C++ programmer they describe, maybe you could offer them your services?   

 

In the Briefing room episode 3 video I linked in the OP, Albert’s and Daniel’s excuse for not addressing the Flying Circus flight models was that they could not verify my book. I therefore provided an alternative “proof” via the Excel model and referenced the supporting data listed in this thread. So in my opinion their initial concerns are all addressed now and they have what they need to do an FC FM revision. In addition, I have already offered to help with that (for free) both via the forum here and via PM’s so they know I’m willing to do that. So I don’t know what’s keeping them? Unless of course, there are no serious plans to do anything about the FC FM’s? In that case a continued silence makes perfect sense......