plane loss rate in general in this game

[Golden Bear]

No problems here, Nick.

I decided to test this out for myself by running raids under different conditions, trying to keep as much constant as I could. This is from the Allied side.

All units involved are well rested following a period of weather shut down on bombing. Presumably the interceptors are equally rested. No escorts in any case.

The first set of tests was a single B-26 unit flying straight to Vaines (sp?) and back. All raids launch at 6 a.m. They arrive over target at 6:41, get attacked from ~6.55 until 7:15. It looks as if 6 units reach them during this time. Others are launched but don't get there. All these raids were run 10 times to get an idea of variability, etc.

B26, 69-99: 12.2/7.8 (69 exp, 99 morale): average 12.2 losses vs. 7.8 for LW
B26, 62-54: 10/6.4
B26, 59-66: 12.4/7.0

There is a lot of variability within each of these numbers. The 59-66 test for example had a result at 4:7 and three at around 17:4.

Then I did a triple raid to the same site:
B26(3x): 11.5/9.5

Bomber casualties did not increase but losses for LW did. Same number of attackers keeps bomber losses even? ...more bombers shooting back drops more interceptors?

Next, a long raid with a single unit to Chaumont. TOT 7:11, first attack at 6:53. The attacks before bombing did not contribute much to bomber losses except in one case that caused 5 casualties:

Long raid B26, 59:66: 28.5/15.5

More interceptor units got involved and for a longer period. Three tests had 31 bombers shot down. None had 32, curiously.

Then, try it with B17s. First to Vaines: TOT 7:03, first attacks 7:04:
B17(63-53): 16/11.8

Greater losses for both bombers and interceptors but the attacks lasted for 50 minutes instead of 20 for the B26s. Note that I tried to match the Exp/Mor with one of the B26 tests.

Finally, B17s going to Creil to shorten the trip. TOT= 6:51.
B17Creil(63-53): 13.2/7.6

In summary, repeat post of above results:
B26Vaines(69-99): 12.2/7.8
B26Vaines(62-54): 10.0/6.2
B26Vaines(59-66): 12.4/7.0
B26(3X)Vaines: 11.5/9.5
B26Chaumont(59-66): 28.5/15.5
B17Vaines(63-53): 16/11.8
B17Treil(63-53): 13.2/7.6

And what does it mean?
First I should note that the B17 results showed less variation from test to test than did the short single unit B26 raids.
The 3 B26 tests, High Exp/Mor, Low Morale, Low Exp all look similar. Do the EXP and Mor numbers of bombers have any effect on combat?
The B17 raid to Treil, the most comparable in terms of combat duration, look similar to the B26 results.
Increasing the number of bombers involved didn't change losses for bombers but raised them for interceptors.
Over a long raid a Medium unit can get virtually wiped out but the interceptors pay a price for it. The ratio of Bomber/Int. losses doesn't improve very much for the interceptors in this condition.

Are these "realistic?" I cannot decide without thinking about it some more.

At least it is some data!


Carlos

[kaybayray]

Hey Bear, Great tests man. You put together a pretty good Matrix my friend. I dont really have time or inclination to do a real work up on your data but I will share with you what you can do to see if any results are significantly different from any other of that cell of your Matrix.

The true test of variation or differing between groups is to analyze the deviation between the Mean of each group. Simply put.... if you analyze your data and determine the Mean and the Standard Deviation of each test group you have all you need.

Once you have calculated those values from standard canned statistics formulas you can compare them in the following manner.

Take the Mean of each data set and add/subtract 3 standard deviations to the Mean. That will give you a form of a comparison range to compare each group with.

The test is if there is ANY overlap between the Range of the Mean +/- 3 Std Dev of ANY one group to ANY other group then both groups are NOT significantly different from each other and can be considered statistically as the SAME population.

Yeah there are a multitude of statistical test you could apply to make your point one way or the other but this test is just about the most simple and it cuts through the BS real quick.

And I will always caution to take care with Statistics because you can very easily tell lies by how you group and sample your data. For example in this thread, the one thing that I have not yet heard mentioned is that the AI, unlike a Human opponent or unlike the waring Military Commanders of the real Air War, does not learn, adapt and innovate. It is a Program with a lot of "If ---> Then --->" lines of logic. Which can be easily mislead.

Also being a game you can absolutely exterminate your supply of Pilots and keep on flying at the rate the game is slotted to deliver Aircraft and Crews. No fear of National Rebellion over extreme losses here. So if you are getting hammered and continue with that strategic / tactical path nobody is going to shut you down. IMO its kind of a Negative Feedback Loop. You can ignore it if you choose to and then say the game is Whack.

Not meaning to smak anybody here. So please if you think I am flaming you, you are mistaken. Just want to point out a few things that I have not seen spoken of that I think are relavent to this thread and the game. Keep in mind this game does not model Olympic Synchronized Swimming. It models one of the bloodiest sectors of one of the largest conflicts of our world.

Just some things to think about.

Later,
KayBay

[Golden Bear]

Hey, Kaybay, yes, I could do that - I actually am experienced in this from my job - but chose not to for a couple of reasons. One, it wouldn't be fun for me. The other is that it is such an impure experiment with so many dependancies that I don't feel it would really hold up statistically. I'm content to look at the results and just sort of size them up. Also, I would need to type all the individual numbers into a spreadsheet and then put it into Minitab, etc. Just not inclined to do so.

You bring up a lot of good points. I've also wondered what would have been the response of the Allies if things weren't going their way when they needed them to. Would the B-29s have wound up in Europe? Could there have been different research/development paths that led to useful jets in Europe sooner? As it is the game is set up to explore the possibilities of different German productions but not for the Allies.

...and I'm not complaining. Just musing.


Carlos

[kaybayray]

Bear, [8D]

I agree. I didnt bother for the same reasons. I am an Engineer and I dance between my calculator and the Real World all day every day so I am not incliined to go there either. [:D]

I took a brief overview of your matrix and my take was... not much difference, or at least nothing I saw jumped out at me.

Yeah your thoughts about the Allied R&D were what I was thinking. The Axis has the ability to adapt and innovate with equipment as well as strat and tactics. The Allies have to rely solely on Strat and Tactics.

But I dont want to hijack this thread. IMHO you Reap what you Sew with respect to losses in this game. At least that has been my observation. I can, by design of missions run the LW round and round and take few losses and eat them alive. Which would only happen once or maybe twice before a Real Air Commander would adapt and bushwhack me. I mean if you and I were playing this out I dont think you would let me get away with any Sneaky Peaky crud more than once. You would figure out a way real quick to smack me upside the head. [8D]

I have only been playing the Allies with this game. Personal reasons. Mostly this game is so complex that its all I can do to wrap the 3 brain cells I still have around the Strategic Bombing Campaign of the 8th AF for the most part. I spend weeks building each days set of missions. For me that is where the fun lies. In the planning, building and adapting missions.

If there was one thing I could add to this game it would be to be able to actually fly AC in missions. Like you can in the Battle of Britain II Wings of Victory. [8D]

Sorry, I just couldnt pass up the opportunity to plug my Dream Game.

Later,
KayBay [8D]

[Deckard777]

Lower Aircraft Kill Rate mod v1.0 (Long Campaign 1943)

see here [;)]

[Golden Bear]

ORIGINAL: kaybayray

I am an Engineer...

KayBay [8D]

Hey, I am too! Except that right now I am a laid off engineer after 22 years with Hewlett-Packard. Not too many openings here in my area of Oregon right now but I expect I'll be OK when the economy shows a little more strength.

Carlos

[Nicholas Bell]

I decided to test this out for myself by running raids under different conditions, trying to keep as much constant as I could. This is from the Allied side.

Carlos, do you have the data on the loss per sortie rate?  Especially the US loss/German sortie rate.  To me this is the critical statistic to compare to history.  After doing a bit more digging I've noted that the maximum German effort was normally 3-400 sorties which engage in the West & over Germany, with 400-500 sorties overall.  Typically one see's a lot more sorties than that - double or triple.

Alfred Price's two extremely detailed books The Hardest Day and Target Berlin provide a lot of interesting data on the two "best" days the Luftwaffe had during the BOB and in defending the Reich.  In both cases, despite the difference in aircraft, the number of aircraft shot down by the defender was about 1/6 the number of engaged defensive sorties.  In the Berlin raid, about 1 of 5 German planes which engaged were lost.  Price notes that only half of the German losses (32 of a total of 66) can be ascertained conclusivily.  Of these know only 2 where shot down by defensive fire, with 3 more colliding with bombers (maybe due to defensive fire?) - the rest being shot down by escorts.  If one interpolates the total number of German fighters lost to defensive fire to 10 of 66, this would support the argument that bomber defensive fire in game is too strong (although I still believe the German method of tracking damaged planes not total write-offs needs to be considered when analyzing their "losses").

[Golden Bear]

Ah, I may have overwritten the file I used. I will/would need to run test over and actually count the number of planes attacking in each combat? Or just identify all the LW units that actually stage an attack and count their planes? A little more tedious than the original test but I could do it another 10x.


Carl

[Nicholas Bell]

Carlos

No, please -  I'm not asking for you to redo your work - was just curious if you had something to compare against.  At some point this becomes work and that is not the point!  [:)]   I can do this kind of testing myself. I don't think I would be counting every interception.  I'm more interested in aggregate numbers myself since the actual historical data is only available in those terms.

Nick

[Oliver Heindorf]

Hello, I have found something odd : I just tried to mass attack by night this turn (BoB, full campaign 40, me as LW) :
I sent some 400 Bombers to a night attack against Birmigham urban (sorry !) just to see what is happening.

I lost 4 Ju 88   & 1 He-111 .

RAF lost 22 Blenheims - they all crashed on their way against my raid not while landing (maybe on their damaged fields...) they got all crashed on their way against my raid.....why ? [&:]


edit : damage urban is now 14 - is this good or bad ?

industrial damage raised +1 - because of theurban damage ?

thank you

[TechSgt]

ORIGINAL: Oliver Heindorf

Hello, I have found something odd : I just tried to mass attack by night this turn (BoB, full campaign 40, me as LW) :
I sent some 400 Bombers to a night attack against Birmigham urban (sorry !) just to see what is happening.

I lost 4 Ju 88   & 1 He-111 .

RAF lost 22 Blenheims - they all crashed on their way against my raid not while landing (maybe on their damaged fields...) they got all crashed on their way against my raid.....why ? [&:]


edit : damage urban is now 14 - is this good or bad ?

industrial damage raised +1 - because of theurban damage ?

thank you

Did you happen to notice if the interceptors were at zero altitude?
I saw this alot in the old version of TOH and since we are using the same engine for both maybe this has gotten transferred over.

TS

[Oliver Heindorf]

ORIGINAL: TechSgt

ORIGINAL: Oliver Heindorf

Hello, I have found something odd : I just tried to mass attack by night this turn (BoB, full campaign 40, me as LW) :
I sent some 400 Bombers to a night attack against Birmigham urban (sorry !) just to see what is happening.

I lost 4 Ju 88   & 1 He-111 .

RAF lost 22 Blenheims - they all crashed on their way against my raid not while landing (maybe on their damaged fields...) they got all crashed on their way against my raid.....why ? [&:]


edit : damage urban is now 14 - is this good or bad ?

industrial damage raised +1 - because of the urban damage ?

thank you

Did you happen to notice if the interceptors were at zero altitude?
I saw this alot in the old version of TOH and since we are using the same engine for both maybe this has gotten transferred over.

TS

woohooo, that would be an explanation. No, I didnt checked the altitude of the RAF planes but the test would be easy to repeat I think. All I did was a mass attack at 19:00 as night bomb raid against birmingham with +400 bombers.

[Golden Bear]

ORIGINAL: Nicholas Bell

Carlos

No, please -  I'm not asking for you to redo your work - was just curious if you had something to compare against.  At some point this becomes work and that is not the point!  [:)]   I can do this kind of testing myself. I don't think I would be counting every interception.  I'm more interested in aggregate numbers myself since the actual historical data is only available in those terms.

Nick

I DID have the file saved and reran it to watch and count interceptors. Let me also note in light of Sarges own tests that I made certain that none of the bomber losses were due to ground fire.

Eight units rose to the battle but only 7 were involved:
Total involved: 107 - 79 190s, 28 109s
Uninvolved: two units of D520s, 66 total

Using an average of 12 bomber losses to 7 interceptor losses gives:
1 interceptor lost for every 15 involved.
1 bomber lost for every 9 interceptors involved.
6.5% of involved interceptors were lost on a rough average.

Comparing this to 20% (1 in 5) makes the results look soft for the LW except that:
Artificial single BG battle
No escorts
No shared escorts

Still, it looks at least not out of line and possibly a little easy on the LW. What is your view, Nick?


Carlos

[Nicholas Bell]

Interesting!  Yes, I agree it does make appear easy on the LW.  I suppose the best way to look at this is in an entire day with a full plate of typical missions - done many times over.  As HS response in the other thread about the medium bombers shows clearly - it is difficult to make clear assertions given the wide range of variables involved.

I suppose had I enough time I could create day's missions and hex edit the savegame so that everyone had perfect morale and no fatigue so the results would be "best possible".  Then alter a bit here and there to see the effect.  What a chore - BTW anyone here who might tell me what a statistically valid sample size would have to be - ie how many missions do we need to run?

The impact of fatigue and morale has been mentioned as a factor in increasing German losses to bomber defensive fire.  Without knowing what is going on "under the hood" it is impossible to consider.  Seems counter-intuitive to me, however.  A pilot with high fatigue/low morale would be less likely to press an attack and therefore not come in range of defense guns.

I ran a mission using Deckard777's mod last night and the results were interesting.  Unfortunately I screwed up saving the file so I don't have all the details, but the losses looked a heck of a lot more realistic for BOTH sides (although I am still more in the camp of needing to tone down the bomber return fire a bit more).

[Jeldren]

I noticed some unusual things.
P-47 suffer more casualties to ground attacks and  enemy fighters compared to mustangs which is odd.
The german player is able to field to many heavy fighters. I tried the Schweinfurt Raid against a human
player and lost more than 200 B-17 against massed heavy fighters and the LW player lost less then
40 heavy fighters(mostly 110).
Mosquito fighter bombers are very(too?) effective against enemy air fields. One Mosquito unit had a 33:1
kill/death ratio and the best Mustang unit had a 12:1 kill death ratio. Most other allied fighter take
much more casualties compared to Mustangs and Mosquitos. A Mustang+Mosquito raid is also
effective against key industries unlike a B-17+Mustang raid with similar numbers e.g. 96 bombers + 96 escorts.

Edit: Did a test with Mosquitos @32k alt vs an air base. Result 60 dead axis fighters vs 19 mosquitos at
@32k alt and the mosquitos suffered no casualties.

[Golden Bear]

ORIGINAL: Nicholas Bell

The impact of fatigue and morale has been mentioned as a factor in increasing German losses to bomber defensive fire.  Without knowing what is going on "under the hood" it is impossible to consider.  Seems counter-intuitive to me, however.  A pilot with high fatigue/low morale would be less likely to press an attack and therefore not come in range of defense guns.

My observation has been that the interceptors do get somewhat less aggressive but not that much. They do start getting themselves shot down when I see them land and take off (and land and take off). The "for instance" for this is that usually when they attack or bounce a damaged fighter (bombers are a different issue and I don't want to put my toe in that pool right now) they shoot down the target. After land/launch they begin to get themselves shot down (not all the time) sometimes.


Carl