Aircraft ROC Review

[el cid again]

The formula would become

(Constant * speed / power) PLUS (ROC / 200) PLUS (constant * dive speed) MINUS (wing loading / 25) MINUS (power loading / 5)

right now we use

(speed / 20) PLUS (ROC / 200) MINUS (wing loading / 25) MINUS (power loading / 5)

Thinking about this, we probably could do away with the last term in the formula (power loading)
and just use speed / power. First pass guess is the constant should be on the order of 100.

The constant for dive speed might be on the order of 0.02

[herwin]

See this air combat thread. You can't patent or copyright an algorithm (at least in Europe), so it can be used for free.

[Dili]

Thank you Herwin.

The last parte seems to be not so detailed.
Curious that Arado 240 position and Typhoon a plane that was disliked for fighting and was relegated to Jabo work. I would say this overstates the speed factor apparently because the fact that Maneuverality is only considered to be maximum turn rate.

[herwin]

ORIGINAL: Dili

Thank you Herwin.

The last parte seems to be not so detailed.
Curious that Arado 240 position and Typhoon a plane that was disliked for fighting and was relegated to Jabo work. I would say this overstates the speed factor apparently because the fact that Maneuverality is only considered to be maximum turn rate.

The database I used to calibrate the model didn't extend to a lot of those aircraft. What you see is the results of applying the model. Remember the high score aircraft were competing among themselves late in the war.

[el cid again]

ORIGINAL: herwin

See this air combat thread. You can't patent or copyright an algorithm (at least in Europe), so it can be used for free.

Well - in general - you cannot patant or copyright any principle of mathmetics. But you CAN copyright computer code. More than that, you can keep it secret, and licence it on the basis that anyone who buys it AGREES NOT to "decompile it" or "reverse engineer" it. The Wall Street Journal did publish a lengthy article about a US Supreme Court decision that licencing can TURN INTO OWNERSHIP - something notably absent in the fine print of every licence agreement I ever saw - but ONLY IF the owning company drops active support for the product. Then ownership passes collectively to the combined body of all licencees, or at least to those who want it.

For an active product that is sold, you cannot simply take a routine that exists in a different product and use it (even if it was perfect for your application). You would have to get permission to use that code. And that permission almost always would not permit you to then resell it. Exceptions would generally require that you acknowledge where you got it from, every time you sold it.

IF Matrix really wanted to have published algorithms for its routines, and IF it wanted to write and publish technical manuals on its products, you would see at least some cases where this was being done. There are sound business reasons it does NOT want to do either - starting with it becomes too easy for someone else to write comparable products and too easily take market share. From that point of view it is better to use unpublished code, and not to use anything from an outside source. If they really want something from outside, Matrix is likely to hire the person who does that to write an entirely new routine to do it, protected by a non-disclosure agreement and the fact the person was paid. This is all philosophy of doing business stuff - and because they are providing us with products that we cannot capatalize or provide enough manpower to produce - we need to be understanding of how that works.

There are exceptions: in the early days UV had AKs without any ability to carry fuel. I suggested they should be able to carry it at a 2:1 ratio penalty. They adopted it. That is, technically, an ultra simple algorithm - and they took it and ran with it - exactly as proposed. Matrix has Forums, and it does listen to the Forums from time to time. But it is very unlikely to import major game routines from alien sources.

Now if you broke the spirit of the licence agreement, and reverse engineered the code, and tried to use just an algorithm,
you would end up having to write the entire routine yourself, and debug it, and that is not any different than doing the same thing in the first place without the algorithm. It is going to be cheaper to just define what you want to do, and do it, and not bother with the reverse engineering part.

[witpqs]

ORIGINAL: Dili

He did not provide information on specifically what ROC the game engine expects to find in the database.

He said it. That you might dont take his words at face value is another matter, and i agree that an official "real manual" would be better.

I am trying to communicate with you, please help. [:)] "He said it." He said what? That he saw the code for the game engine? That Matrix confirmed what is in the code for the game engine? Or did he just assert that it works a certain way without citing any evidence that he had to confirm it works that way? He has expertise and he is honest - that is not the issue. The issue is just knowing for certain what Grigsby/Matrix did in the code.

If TheElf can tell us how he confirmed what the game engine uses, I am willing to believe him. But I haven't seen that, only (basically) 'it works this way'. And details were missing too. I mean that if it is 'time to altitude', then:

What altitude? Is it one altitude for all aircraft, or is it the service ceiling of each aircraft (obviously different for each one), or something else?

What mathematical formula must be used to convert 'time to altitude' (which is a [time] value) to ROC (which is a [change in altitude]/[time] value)? Is it simply altitude/time, or is it a more complex formula so certain altitude ranges are favored?

I'll put in WITM40 MVR:

Speed vs power: to calculate drag
power vs weight: to calculate acceleration( the ROC issue made me think of averages in this too)
wing loading
dive speed (drag and weight?)

Still not tought of weight this factors will have for all that (power vs weight will be the most important) and i'll have to manage a soup that will expel numbers similar to those in data base.


Dunno if this would help:
http://avhistory.org/communityserver/default.aspx

This stuff is great - it just doesn't tell us what Matrix did.

[Dili]

Or did he just assert that it works a certain way without citing any evidence that he had to confirm it works that way? He has expertise and he is honest - that is not the issue. The issue is just knowing for certain what Grigsby/Matrix did in the code.

 
Maybe true, but he should have made that caveat then.

What altitude? Is it one altitude for all aircraft, or is it the service ceiling of each aircraft (obviously different for each one), or something else?
What mathematical formula must be used to convert 'time to altitude' (which is a [time] value) to ROC (which is a [change in altitude]/[time] value)? Is it simply altitude/time, or is it a more complex formula so certain altitude ranges are favored?

 
Time to 6000m at least. Yes it will be a simple altitude/time. Imperfect but better than sea level ROC.
 

This stuff is great - it just doesn't tell us what Matrix did.

 
True.
 
 

[TheElf]

ORIGINAL: el cid again


Nevertheless, it should be pointed out that the original statement is more true than false for air combat maneuvers. You don't make them at "moderate speed" - and that is the rub.

Corner Speed is where optimum instantaneous-turn-rate performance and maximum G are obtained. Corner Airspeed is typically a moderate Airspeed.

Speeds higher than corner are however less detrimental to turn performance than slow speeds since deceleration is generally faster than acceleration.

Aircraft at higher airspeeds than corner will be at there G-limit until they reach corner and as such will have a wider turn circle until they bleed excess speed.

[el cid again]

The current changes to ROC and maneuverability are included in LEVEL 7 ONLY in 7.783 microupdate. But 5 and 6 contain aircraft eratta reported (as for the torpedo bombers of FAA). x.783 mainly is device reform re gun accuracy (reform in the sense of consistency - with a single change to range - so Allied 75 mm infantry guns may engage to 3,000 yards instead of only 1,000 yards).

IF a CONSENSUS is reached on a revised maneuverability formula, I will do it all over again - for x.785. Otherwise x.784 will include the rest of the changes - if there are any more (most Allied planes are already using initial ROC). CONSENSUS means what it says.

No one has commented on the fact that initial ROC was used for Allied planes, but not for Japanese ones. This is a build in bias in the data - and it is not bothering the critics.

For the time being Levels 5 and 6 will retain the old data (except when eratta are detected). It saves me data entry time, and they are supposed not to be in development any more. But significant errors will be fixed.

[el cid again]

Post x.782 Allied aircraft thread (for types not already at initial ROC in data set):


A-20B initial ROC 1200 fpm. This reduces maneuverability to 10 (down from 12).
A-20G initial ROC 2000 fpm. This increases maneuverability to 12 (up from 11).

Anson I has an initial ROC of 720 fpm. This is greatly below the 2230 we were using, and it reduces maneuverability to 7 (down from 10). The aircraft is reclassified as a bomber. While all but two initial users of this machine did not use it with bombs and depth charges, those two cannot fly their historical missions (for RAAF) unless it is classified this way, and bomber units may still do recon and naval search missions. Further - players may do what real life operators sometimes did - arm the plane. [They must upgrade twice to do that - and that has a logistic cost - but for everything there is a price] This is treated as eratta and backfitted.
Blenheim I initial ROC 1500 fpm increases maneuverability to 10 (up from 9).
Blenheim I Night Fighter initial ROC 1500 fpm increases maneuverability to 10 (up from 9).
Blenheim IV initial ROC 1540 fpm does not change maneuverability of 9.
Buffalo/F2A in RAF service - properly shown with reduced armament - was improperly shown ROC wise. The reduced weapons and ammunition and fuel meant the aircraft was lighter - and could climb fully 500 fpm faster than it could in its original (Dutch) form. ROC is 3070 fpm (RAF data). This increases maneuverability to 28 (up from 26), and now you can see the difference between it and its heavier armed cousin in NEIAF service - more maneuverability. It is treated as eratta and backfitted.
Catalina I initial ROC 920 increases maneuverability to 6 (up from 5). Operated to RAF norms, it should have a maximum speed of 190 mph, a cruising speed of 179 mph, endurance of 1461 minutes. Max load 2000 pounds. Operational Ceiling 19,200 feet. 1 forward, 2 side, 1 bottom rear .303 guns. Treated as eratta and backfitted. Yet another case where the RAF operated lighter, change armament and used different throttle settings to achieve different performance of nominally "identical" aircraft.
Dakota operated to RAF norms has initial ROC of 1130 fpm. That increases maneuverability to 7 (up from 6).
FAA data. The aircraft is important, and it was operated very differently in CW services than US ones (it has a different range for example), so using different statistics helps highlight these differences. It is backfitted.
Mitchell III operated to RAF norms has initial ROC of 1880 fpm. This increases maneuverability to 10 (up from 7).
This aircraft should have max speed = 292, cruising speed = 210, endurance = 515 min (467 is actual transfer range, but we use a modified value so operational ranges are correct), max load = 6000 lbs. It has 1 forward, two top turret, two dorsal turret, and one tail .50 cal guns. Normal bomb load = 6 x 500 lb bombs. This is treated as eratta and backfitted.
Vildebeeste initial ROC 765 fpm. This reduces maneuerability to 9 (down from 10).

[akdreemer]

ORIGINAL: m10bob

ORIGINAL: el cid again

ORIGINAL: AlaskanWarrior

Why redesign the wheel? There are plenty of good air combat sims out there? I would guess that at lest some have a following of fanatical 'accuracy' pundits. Why not get some of these on board?

From Matrix point of view, the answers likely include

a) They don't own the other routines - and would have to licence them - and pay for them;

b) The existing other routines not designed for WITP won't be written in a way that is compatable with it, in several different technical senses of that term;

c) Using any such routine would not then permit Matrix to claim ownership, and licence in its own right, nor to have non-disclosure agreements in the simple sense they can with propritary code; clearly Matrix likes to own its code - and to keep it secret - and so this would not be compatable with that philosophy of doing business.

d) There are probably a number of different kinds of routines required for WITP involving different functions related to air combat, AA combat, air-ground combat, etc. It is likely many of them have NO equalivent in other software. And whatever exceptions there may be are almost certainly not exactly what WITP needs.

--------------------------------------------------------------------------------------------------------
Since "mods" are not the property of any corporation, and are offered for free, I suspect we could put ANY value in ANY box in the editor and the program would be smart enough to follow whataver we enter.
In the "maneuver" box, I doubt the computer is figuring on which 5 apples the value represents, but is literally comparing the number to the opponent plane, verbatim.
Therefore, it is up to us to determine what values go where.
We have proveable values for ROC, Service ceiling, speed, size, weight, wing area, etc,and I believe Alaskan Warrior has a good point.
Lots of time went into producing other wargames and their rules, and we might look at how THEY compared apples to oranges, and find common ground.

I believe the air module (stock) bears a re-work, if for no other reason than the fact that I found some of the planes using *identical* values as the Gunston book, but then other planes were just in somebody elses playground, not even close to the same point of comparison, (or verifiability with ANY resource I have found,yet.)
This last is a bold statement, but I have given my sources, all readily (and cheaply) available from amazon.com...

BTW, this is the first thought out (and protracted) discussion with different viewpoints I have seen on this forum in a long time, which was able to have such great talent and ideas, without personal vitriolic comments and non-productive agendas entering the discussion.

Hell, it's an educational experience![;)]
[/quote]
Actually i was not referring to Matrix, but us modders. There must be some air combat affeciandos on this forum that could be of great assistence in helping design an appropriate method to determine maneuverability in the context of WITP. Frankly, I do not have the inclination for it, except for gut feelings and empirical observations of what the results should mirror.

[akdreemer]

ORIGINAL: el cid again

ORIGINAL: AlaskanWarrior

Two additional factors that no one seems to be looking at. Dive speed and acceleration. By mid war maneuverability was not as important as the ability to pounce and escape, thus the energy fighters became dominant. Why else did the admittedly poorly maneuverable plane like the P-38 turn out to be such a deadly air-air plane?

Actually - some of us (there is a PhD in the crowd, and I not only have worked on this at a USAF SIL, I have collected the dive speeds for all Japanese aircraft - including many we do not use) - are well aware of these - and yet other - factors.
But they are difficult to obtain data for. And we need data on hundreds of different aircraft, some of which never flew at all, more than a few of which do not have complete data records or have contradictory data records.

However, I think this model DOES include acceleration insofar as ROC is directly related to it. But dive speed is very tricky: not only do I have no idea how to calculate it from basic data (like weight), I have no idea how to use it in a simple air combat model like this one? I can use it in a much more complex model - where we figure out when a plane is diving - and then apply measured data to where it goes next. But how to do it here? I have no idea.

Why is there a need to model every plane the same way for air-air maneuverability?

[JeffroK]

From AlaskanWarrior
 
Actually i was not referring to Matrix, but us modders. There must be some air combat affeciandos on this forum that could be of great assistence in helping design an appropriate method to determine maneuverability in the context of WITP. Frankly, I do not have the inclination for it, except for gut feelings and empirical observations of what the results should mirror
 
Until Matrix/2by3 or whoever formally advises how the various ratings interact in the combat engines, such a effort would be a waste. (I doubt they would release such info for general usage)
 
In my simpleton view, a SWAG is made of the manouverability (using the data as mentioned by many in this thread) and then tweaked to produce results we can relate to history. Herwin published a link to a list of manouverabilties for fighters, The Spit V rated the same as the P-61, maybe technically but IRL?  Probably 1000's of hours of testing.
 

[Dili]

That reads more like a speed list not a manouverality one.

[akdreemer]

ORIGINAL: JeffK

From AlaskanWarrior

Actually i was not referring to Matrix, but us modders. There must be some air combat affeciandos on this forum that could be of great assistence in helping design an appropriate method to determine maneuverability in the context of WITP. Frankly, I do not have the inclination for it, except for gut feelings and empirical observations of what the results should mirror

Until Matrix/2by3 or whoever formally advises how the various ratings interact in the combat engines, such a effort would be a waste. (I doubt they would release such info for general usage)

In my simpleton view, a SWAG is made of the manouverability (using the data as mentioned by many in this thread) and then tweaked to produce results we can relate to history. Herwin published a link to a list of manouverabilties for fighters, The Spit V rated the same as the P-61, maybe technically but IRL? Probably 1000's of hours of testing.

I would actually think that 'imformed guess' would be a better term.[:D]

[el cid again]

ORIGINAL: AlaskanWarrior

ORIGINAL: el cid again

ORIGINAL: AlaskanWarrior

Two additional factors that no one seems to be looking at. Dive speed and acceleration. By mid war maneuverability was not as important as the ability to pounce and escape, thus the energy fighters became dominant. Why else did the admittedly poorly maneuverable plane like the P-38 turn out to be such a deadly air-air plane?

Actually - some of us (there is a PhD in the crowd, and I not only have worked on this at a USAF SIL, I have collected the dive speeds for all Japanese aircraft - including many we do not use) - are well aware of these - and yet other - factors.
But they are difficult to obtain data for. And we need data on hundreds of different aircraft, some of which never flew at all, more than a few of which do not have complete data records or have contradictory data records.

However, I think this model DOES include acceleration insofar as ROC is directly related to it. But dive speed is very tricky: not only do I have no idea how to calculate it from basic data (like weight), I have no idea how to use it in a simple air combat model like this one? I can use it in a much more complex model - where we figure out when a plane is diving - and then apply measured data to where it goes next. But how to do it here? I have no idea.

Why is there a need to model every plane the same way for air-air maneuverability?

How could we get valid outcomes in air combat if we deliberately biased the values in the fields to favor some aircraft over other aircraft? Granted that air combat is a special case when there are air encounters: the 90% normal case (which is almost never properly modeled in games) is that the side which first saw the other side wins, both offensively and defensively. The way that works IRL is that if you see a plane or formation you don't want to tangle with, and they have not yet seen you, you evade. On the other hand, if you see a plane or formation you do want to engage, and they have not yet seen you, you seek a favorable attack situation (altitude, bearing, attitude vs sun, etc). This works 9 times in 10 REGARDLESS of the plane type or performance of the planes. We can only hope something like this is used in the game.

Air air combat is the special case where

a) Both sides see each other at the same time, and both do not wish to evade (which, obviously, is possible if BOTH want to do so);

b) Both sides see each other at about the same time, but the side that wants to engage has sufficient speed, altitude, relative position to force an engagement;

OR

c) Regardless of who saw the other first, one formation is committed to press on regardless (to bomb the target, drop paratroopers or supplies, to take the photographs, or whatever).

Now here we have a case where the relative performance of both sides does matter (it does not in the more common case that one side has surprise). [In air combat, the vast majority of the time, the loser never knew there was an attack, never sighted the enemy, and never made any effort to evade or engage in any sense.]

How can we determine who gets into firing position for each round - if anyone - UNLESS we use a comparison based on identical standards of performance? It appears obvious that is the way to go.

[el cid again]

ORIGINAL: AlaskanWarrior

ORIGINAL: el cid again

ORIGINAL: AlaskanWarrior

Why is there a need to model every plane the same way for air-air maneuverability?

In the case that an air combat occurs where one side has not achieved surprise,

it seems impossible to calculate who gets into firing position fairly unless we have equal standards for all planes.

What else do you have in mind?

[el cid again]

ORIGINAL: JeffK

From AlaskanWarrior

Until Matrix/2by3 or whoever formally advises how the various ratings interact in the combat engines, such a effort would be a waste. (I doubt they would release such info for general usage)

We have been formally informed - many months ago - when working out the maneuverability rating. It is probably in a thread to read. The information was by Mike Wood.

[m10bob]

ORIGINAL: AlaskanWarrior

ORIGINAL: el cid again

ORIGINAL: AlaskanWarrior

Two additional factors that no one seems to be looking at. Dive speed and acceleration. By mid war maneuverability was not as important as the ability to pounce and escape, thus the energy fighters became dominant. Why else did the admittedly poorly maneuverable plane like the P-38 turn out to be such a deadly air-air plane?

Actually - some of us (there is a PhD in the crowd, and I not only have worked on this at a USAF SIL, I have collected the dive speeds for all Japanese aircraft - including many we do not use) - are well aware of these - and yet other - factors.
But they are difficult to obtain data for. And we need data on hundreds of different aircraft, some of which never flew at all, more than a few of which do not have complete data records or have contradictory data records.

However, I think this model DOES include acceleration insofar as ROC is directly related to it. But dive speed is very tricky: not only do I have no idea how to calculate it from basic data (like weight), I have no idea how to use it in a simple air combat model like this one? I can use it in a much more complex model - where we figure out when a plane is diving - and then apply measured data to where it goes next. But how to do it here? I have no idea.

Why is there a need to model every plane the same way for air-air maneuverability?

I don't believe this is the question you meant to ask ?

Why compare a 75l40 to a 75l70 , or a Piper J3 to a M D F4e ?

A question like yours might have unintended consequences, if taken verbatim.
If we did not compare the planes on the same uniform criterium, we would be in the realm of fiction or other Sci Fi........
This is the very reason we are researching ALL the stats at present.
We *HAVE* found inconsistencies.

[herwin]

ORIGINAL: Dili

That reads more like a speed list not a manouverality one.

It's a statistical model that incorporates a number of factors including speed. Speed and pilot skill control who gets the drop on who. After that, it's energy (i.e., altitude and speed), manoeuvrability (power loading, wing loading, etc.), and firepower that produce the results. Fighters almost always get the drop on bombers if they can make the intercept, so then it's firepower and ruggedness that dominate.