RHS Maneuverability Review: Data [ALL Data Done]

[el cid again]

Based on a specific private request for review of the P-38 in particular

and general discussion of maneuverability ratings

I am attempting to review the Maneuverability rating system used by RHS
for a new RHS series (which will be numbered 5.xx for clarity that there is a difference).

This is a low intensity (not time intensive) review - and it may take a long time to execute if we find a better system - because I fear vast amounts of data research (and no little calculation) may be required.

The best I can say about the RHS system is that it is openly defined, and it is subject to immediate change if a simple improvement is found, and subject to long term change if a difficult improvement is found. It at least has the virtue of consistency - given the data available it is possible to say "that is the right value".

And the formula is

[ (speed / 20) + (rate of climb / 500)] / Number of Engines.

I wish to have proposals for changes - not in abstract form - and not because "P-38 is not high enough" - but in specific form -
that is in the form of formulas that can be used with data we can actually
obtain to CHS/RHS standards in less than a lifetime of research.

Such proposals may take two forms:

Immediate: using data PRESENTLY in the data set (or clearly universally available, such as aircraft weight)

Distand: using data NOT in the data set (but which can be obtained)
and this ONLY is germane if you have specific sources for the data (including wierd types like J7W1) - if you cannot point to sources I cannot use the formula.

[Nemo121]

Could you list what the current formula is so we could see where the weightings have been changed and so what is now accorded greater importance than previously?

[el cid again]

ORIGINAL: Nemo121

Could you list what the current formula is so we could see where the weightings have been changed and so what is now accorded greater importance than previously?

The formula is in the middle of the above post.

There is one qualifier (never used yet): a plane with multiple engines on the same axis does not count these as multiple - but as one.

I am thinking of using "square root of the number of engines" vice "number of engines" - which works well for the two engine case -
but not for the four engine case.

I am thinking of changing the algorithm from 80% speed in the upper body of the formula to 60% - that is to double the non-speed factor again (it was originally 10% and now is 20% in my view of things). Further, we could make the 40 % be more than just ROC. Instead, divide it something like this:

10% for ROC
10% for power loading
10% for wing loading
10% for something else (ideally roll rate - but I have no way to get the data for that on a comprehensive, consistent and authoratative basis)
If we cannot get roll rate, maybe we use another factor - or maybe we make wing loading matter 20%?

All these proposals would then be subject to "calibration" - comparing the result with what we have and with other data - over a large number - to see if it is closer - and closer by type?

Note I am not proposing a special rule for P-38 - I want a general rule for everything that happens to rate P-38 better in relative terms. Also note that P-38 must be better in durability vs any single engine plane with otherwise equal statistics - so it will be a better plane overall due to its second engine (sometimes it can go home with one engine turning). No one factor tells the whole story for an aircraft.

[el cid again]

I think that the present algorithm says:

Speed is decisive 80% of the time in determining the tactical advantage a plane will have on a statistical average basis. Other things are decisive 20% of the time - and they are crudely represented by ROC.

I think we might prefer doubling that ratio - and making the non- speed component include things different than at present. The math is not important in that we can do it quickly. But the data is a problem - saying you want roll rate for example is lovely in theory and possibly impossible in practice. But roll rate is related to things we can know - so maybe using loadings etc will get us close enough?

[Terminus]

The formula has been proved to be correct, I presume?

[el cid again]

Regretfully, no. It is better to say we are engaged in a process of calibration. It appears the formula we now use is better than the data set used in other mods - I am not sure it is even right to say any of them have a consistent formula at all? But that is a very iffy statement
because:

a) It takes a LOT of data to validate - and we need a lot more to be confident this is true over time - particularly later in the war with higher performing aircraft;

b) There are still problems between plane types - in my view the 2 E fighter/night fighter case is a problem in ALL mods - RHS has reversed the problem but it is STILL a problem (stock and CHS and Nik Mod all give 2E fighters too high a performance, RHS too little). The ONLY point ALL mods seem to have right (a reference point?) is 4E types - and that is common in a wierd sense - they almost always are the same value!

c) Even if RHS is better than others in relative terms, it is NOT right.
That is, I am sure maneuverability can be better - if only we can figure out how to get there? This is a "quick and dirty" mod - anything better is used - but it is not theoretically going to fit all things to use such a simple algorithm.

It is in recognition of the latter point I am asking for input - I LIKE the criticism that the algorithm used is not ideal - I AGREE with it and ALWAYS agreed with that - and am attempting the tiny little detail of figureing out "what is actually better in a sense we can use it?" - something easy to say but not so easy to do.

EDIT: It appears that RHS is closer to NikMod values than to others. Nik used a result driven model while RHS uses a theory driven model corrected to result ranges. There are advantages to both models and I wish to say that NikMod air combat results are quite good and far better than what came before.

[Terminus]

You call this "quick and dirty"? You should see the aircraft I put in my mod; there you can talk about quick and dirty...[:D]

[Herrbear]

What is the rationale for dividing by the number of engines. You may have stated this before in other posts, but I don't remember it at the moment.

Thanks

[el cid again]

It appears that this may be similar to the original formula used by Matrix. Actually, they seem to have divided by the number of engines for 1E and 2E planes UNLESS the plane was a 2E fighter - when they ignored the 2 - and it appears they multiplied by 2 again for 4E planes. Since RHS values are much lower, our 4E planes could have similar values if we just divided by 4 (vice 8) - and since I like consistency - I made all 2E planes the same. That resulted in a single denominator for all planes - which is good practice in equation building. The previous system has serious anomolies if you compare IDENTICAL planes used as bombers and fighters/night fighters. Either they have to be inconsistent, or one has to be out of sync with other 2E bombers or 2E fighters. On the other hand, fans of 2E fighters (P-38 in particular) don't like the impact of this consistent principle.

The idea is simple: multiple engines are always an advantage in terms of survivability (that is, durability rating) but always a disadvantage in terms of maneuverability rating. Wether or not it is directly proportional to number of engines is open to debate - and I am willing to (say) make it square root of number of engines (for example) - and then measure the impact and compare it with actual. Right now it appears that 2E fighters are relatively too effective in other mods but not effective enough in RHS - although I believe their overall value is well modeled in RHS (due to the increase in durability and also their long range - which is related to engine count).

Anyway - the compromise is not a happy one - and I want a better one - if it can be found.

[Nemo121]

Hmm, dividing by the square root of the number of engines will certainly help twin-engined fighters and bombers BUT if applied to four-engined bombers it is going to result in some four-engined bombers having quite reasonable manoeuverabilities which would fly in the face of their use en masse in formation flying. Still, if that's the way to get the P38 to be "more reasonable" in the opinions of various users then I say go for it. Mark my words though, it won't increase the P38s effectiveness enough for most and you'll find crys for the modelling of other issues as 2nd tier efforts to "improve" the P38s.
 
Also don't forget that what works for the P38 will also effect the Betty and make that a bit more survivable too. I think this is reasonable from reading some accounts from F4F4 attacks on Bettys and Nells during the Guadalcanal campaign but prep yourself for howls of outrage the first time someone gets a message that a Betty evades attack from some Allied plane.

[el cid again]

The nice thing about these equations is that a bomber usually is lacking in speed and particularly lacking in ROC - so a 2E fighter may do better because of its high values in that regard. Then too, the night fighters (which typically ARE bombers) are sort of modest in these figures - placing them between high performance fighters and true bombers.
This is actually pretty good modeling.

But it may be that we should go back to the Matrix concept of doubling the value used for 4E bombers (and I assume any 3E plane that ocmes along). [Transports are the same as bombers]. And we might rationalize it as a formation flying factor - although clearly a 4E plane might not fly in a gigantic box formation. If we did that, RHS 4E values would not change at all - neither would any 1E plane - and only 2E planes would change. This change would benefit those with high performance most of all - and might be a happier compromise in terms of relative ratings. Wish I had thought of it before now.

Even so, I am proposing to go one step farther, by increasing the proportion of the upper section of the equation devoted to non-speed functions. Since power and weight are known in all cases, we can easily include wing loading and power loading - and these will help separate dogs from acrobats. Perhaps we might give 60% to speed (vice 80%), 20% to ROC (as now) and 10% each to Wing Loading and 10% to Power Loading - for the quick and dirty fix.

Longer term I would like to give 10% to roll rate - if it can be determined what that is to a proper standard for all planes - and reduce ROC to 10%.

And I STILL want any better ideas out there.

Whatever we try we will compare to what we have and to history - in general and in relative plane type senses.

[Terminus]

2nd generation night-fighters (Beaufighter and Black Widow) shouldn't get Maneuverability ratings like bombers, though I agree with you on 1st generation (Blenheim, Havoc and Ventura).

[el cid again]

I also agree with you - I think. I have tried to say that higher performance (speed and ROC now, and other things if we add them) mean the better 2E planes end up looking a lot more like fighters than bombers - and the low performance "fighters" look a lot like 2E bombers - particularly when it is the same plane (= same specs). I think a good formula will give us our cake and eat it to - rating both high and low performnce planes more or less properly.

[Mifune]

Cid, just so I can understand. On your current proposal, how would your formula look?

[spence]

though I agree with you on 1st generation (Blenheim, Havoc and Ventura

Can't help myself when it comes to this plane which Matrix completely messed up with its take on its capabilities.  This "first generation" twin engine bomber successfully ATTACKED and shot down Japanese "second generation" single engine fighters on several occasions (Oscar IIs, Tojos), even when operating at relatively long range (like in the Kuriles after flying from Attu).  It also attacked and shot down a variety of multiengine Japanese aircraft in one on one type confrontations while on patrol.  Admittedly the number of air to air kills by PVs is relatively small but there were no confirmed losses of PVs (1 or 2) in air-to-air combat with IJ forces.

Every source I've found that is actually describing the right airplane puts its top speed over 300 and range in the 1600 mile range (because of is airliner origins and the airliner origins of its predecessor and the similarity in sillouette [though not size] I kinda think a lot of casual "authorities" got them mixed up.

[Terminus]

Yes, yes... We know you're partial to the Ventura, spence, and nobody's blaming you. It was messed up by Matrix, but can we please let this go now? It's been fixed in CHS, hasn't it?

Besides, we're talking about the night fighter Ventura here.

[herwin]

ORIGINAL: Terminus

2nd generation night-fighters (Beaufighter and Black Widow) shouldn't get Maneuverability ratings like bombers, though I agree with you on 1st generation (Blenheim, Havoc and Ventura).

The effect of having two engines rather than one is to decrease the maximum roll rate and (usually) to increase weight (relatively speaking). You end up with decreased roll rate (due to the weight of the engines out on the wings), decreased sustained climb and operating altitude (from the reduced power to weight ratio), increased dive speed (because your CD was reduced), decreased maximum speed (more parasitic drag and less power to weight), decreased power to weight (meaning you couldn't accelerate as fast), and reduced turn rate (due to increased wing loading). The only one of these that you can't estimate from the published parameters is roll rate, which you can guestimate.

4E bombers were limited to about 3 Gs rather than 6 (as in fighters), so the maximum and sustained turn rates were halved.

I hope that helps.

[Nicholas Bell]

    El Cid-

Don't forget to run the tests for yourself if you don't beleive me.  Bomber manueverability has no impact on combat resolution.  Cruise speed does.  Just set the the values to zero in a test run and you'll see.

[el cid again]

ORIGINAL: Mifune

Cid, just so I can understand. On your current proposal, how would your formula look?

[ (Speed / 20) + (ROC/250) + (Wing Loading/x) + (Power Loading/y)]

all divided by Square Root (Number of Engines)

Exception 1: Two engines on the same axis count as one.
Exception 2: 3 + engine planes double the denominator.

Where x and y are values not yet determined so that the loadings will end up in the indicated range.

[el cid again]

Spence: Black Mamba first brought the Ventura and specs in the game to my attention. We have since modified its range - and now also an up engined variant with even more range. We gave the initial version its ability to carry torpedoes and depth charges on appropriate missions as well. I don't remember what happened to speed and ROC - but it was checked - and so hopefully is also right. I saw a photograph somewhere of a plane attacking a Japanese aircraft - albiet a 2E one - with notes similar to your remarks. I think the plane may be better modeled now (in no less than 3 variants - one of them a night fighter if I remember right) - but under this proposal all versions will end up with a higher maneuverability rating relative to 1E and 4E aircraft.