B17 Losses too heavy

[treespider]

I personally don't see a problem with large numbers of unescorted bombers being shot down in large numbers by large numbers of fighters..

IMO the problem lies with the ability for players to amass the large formations in the first place. Operational attrition is underrepresented ( as I mentioned in another thread. The designers responded that if it were increased it did bad things to the AI.)

[Kereguelen]

ORIGINAL: treespider

I personally don't see a problem with large numbers of unescorted bombers being shot down in large numbers by large numbers of fighters..

Even if the unescorted bombers are much heavier armed and flying in close formation? It is not so that I would not agree partly with statement, but Nates shooting down large numbers of B-17s at 20,000 feet? Surely, the German Luftwaffe managed to kill quite a lot of Allied heavy bombers, but this was in the ETO with fighter-planes that were a little bit more suited for that task than most Japanese planes.

[treespider]

ORIGINAL: Kereguelen

ORIGINAL: treespider

I personally don't see a problem with large numbers of unescorted bombers being shot down in large numbers by large numbers of fighters..

Even if the unescorted bombers are much heavier armed and flying in close formation? It is not so that I would not agree partly with statement, but Nates shooting down large numbers of B-17s at 20,000 feet? Surely, the German Luftwaffe managed to kill quite a lot of Allied heavy bombers, but this was in the ETO with fighter-planes that were a little bit more suited for that task than most Japanese planes.

Has anybody run the Nate vs B-17 test? If not what do I have to do and I'll run it...my speculation is a large number of damaged B-17's

[Kereguelen]

The Nate-thing happened to me in PBEM on more than one occasion. Of course losses were not as drastic than presented in the Zero test example found above in this thread. But still...

[treespider]

ORIGINAL: Kereguelen

The Nate-thing happened to me in PBEM on more than one occasion. Of course losses were not as drastic than presented in the Zero test example found above in this thread. But still...

So the question still remains was the "Nate-thing" a one off...? I'll gladly run the tests if someone tells me how to set it up...

[rtrapasso]

ORIGINAL: Kereguelen

ORIGINAL: treespider

I personally don't see a problem with large numbers of unescorted bombers being shot down in large numbers by large numbers of fighters..

Even if the unescorted bombers are much heavier armed and flying in close formation? It is not so that I would not agree partly with statement, but Nates shooting down large numbers of B-17s at 20,000 feet? Surely, the German Luftwaffe managed to kill quite a lot of Allied heavy bombers, but this was in the ETO with fighter-planes that were a little bit more suited for that task than most Japanese planes.

I don't think even the Luftwaffe managed anything like this!!![X(]

EDIT: These results are absurd! I mean total of 138 Zeros and Tonys (mostly Zeros) shooting down 119/169 heavy bombers??

[Mr.Frag]

Lets not assume bonuses thats do not exist. There is no bonus granted for a combat box formation. 1 B-17 has the same odds as 1000.

If you want to use your aircraft like this type of bonus does exist, you are going to be sorry for obvious reasons.

Large raids require escort. Flying without any means you are not sneaking in for a quick bombing and using your guns to escape the *few* aircraft that come after you.

[rtrapasso]

ORIGINAL: Mr.Frag

Lets not assume bonuses thats do not exist. There is no bonus granted for a combat box formation. 1 B-17 has the same odds as 1000.

If you want to use your aircraft like this type of bonus does exist, you are going to be sorry for obvious reasons.

Large raids require escort. Flying without any means you are not sneaking in for a quick bombing and using your guns to escape the *few* aircraft that come after you.

No, i think the results show there is a strong malus (negative bonus) for large formations.

If 10 Zeros and 3 Tonys met 17 Heavies on a raid (scaling back by a factor of 10), do you see consistently see an average of 12/17 being shot down?

I haven't in the game, (and certainly not in historical accounts.)

[Mr.Frag]

No, i think the results show there is a strong malus (negative bonus) for large formations.

It's a function of dice averaging.

If you say something has a 10% chance of happening and you only send 1 aircraft, the odds of that aircraft going down are 1 in 10. ie: you are virtually immune which you see with small numbers of aircraft. You are effectively defeating the dice by using a sample size so small.

When you send 1000 aircraft, you are saying that you expect 100 aircraft to be lost (10% of 1000). This 100 aircraft being lost looks shocking, but the odds have not changed one bit.

Larger battles will be bloodier as the averages come into play and you can't beat them anymore by submitting too small a sample size.

[Sharkosaurus rex]

If 138 Jap planes shoot down 119 USA 4 engines bombers in one combat then there was to be something wrong with the model. They must be receiving tooo many dice rolls with tooo many factors helping them. Just to dismiss it as no escort was present is a joke. Maybe you should start your own TV show? You can use WitP as your dummy!

[rtrapasso]

ORIGINAL: Mr.Frag

No, i think the results show there is a strong malus (negative bonus) for large formations.

It's a function of dice averaging.

If you say something has a 10% chance of happening and you only send 1 aircraft, the odds of that aircraft going down are 1 in 10. ie: you are virtually immune which you see with small numbers of aircraft. You are effectively defeating the dice by using a sample size so small.

When you send 1000 aircraft, you are saying that you expect 100 aircraft to be lost (10% of 1000). This 100 aircraft being lost looks shocking, but the odds have not changed one bit.

Larger battles will be bloodier as the averages come into play and you can't beat them anymore by submitting too small a sample size.

Then you DO expect an average of 12/17 heavies to be shot down.

I phrased my question above carefully "If 10 Zeros and 3 Tonys met 17 Heavies on a raid (scaling back by a factor of 10), do you see consistently see an average of 12/17 being shot down?"

Yes, i understand that maybe sometimes 0/17 will be shot down. But then you would expect multiple instances of 17/17 being shot down (enough to raise the AVERAGE loss to 12/17).

This of course can be relatively easily tested, and i think it has been in the past.

But Tristanjohn could repeat the tests standing down 9/10 of the units on both sides (hint, hint). I don't believe that the results will bear out your contentions.

[timtom]

ORIGINAL: timtom

Hate to say it, but if certain stallwart defenders of the status quo were to comment, they could not unreasonably argue that what (statistically) amount to a few random results doesn't prove anything.

ORIGINAL: Mr.Frag

Larger battles will be bloodier as the averages come into play and you can't beat them anymore by submitting too small a sample size.

ORIGINAL: Mr.Frag

Predictability is a dangerous thing! [X(]

[:D]

[Mr.Frag]

ORIGINAL: Sharkosaurus rex

If 138 Jap planes shoot down 119 USA 4 engines bombers in one combat then there was to be something wrong with the model. They must be receiving tooo many dice rolls with tooo many factors helping them. Just to dismiss it as no escort was present is a joke. Maybe you should start your own TV show? You can use WitP as your dummy!

Would you care to find a single encounter during the course of the entire war where 119 heavy bombers were sent in unescorted against a very strong Japanese fighter area.

These strawman discussions with absolutely no basis in reality get rather annoying after a while. Apart from the Allied Commander being taken out and shot for treason for such stupidity, thats about the only result that would be predictable.

Taking a discussion point that the B-17 was a hardy aircraft and could handle an attack by a couple A6M2's and walk away the victor and extrapolating it into 1000 B-17's flying in a tight formation with no manuvering room with full bomb loads would be able to take on 1400 fighters is just being silly yet here we are, 7 pages worth of posts.

[rtrapasso]

ORIGINAL: timtom

ORIGINAL: timtom

Hate to say it, but if certain stallwart defenders of the status quo were to comment, they could not unreasonably argue that what (statistically) amount to a few random results doesn't prove anything.

ORIGINAL: Mr.Frag

Larger battles will be bloodier as the averages come into play and you can't beat them anymore by submitting too small a sample size.

ORIGINAL: Mr.Frag

Predictability is a dangerous thing! [X(]

[:D]

And the prophet of the hour award goes to timtom, of Aarhus, Denmark!![:D]

We are re-naming the statuette award "the Timtom" in your honor!!!

[Tristanjohn]

ORIGINAL: ChezDaJez

Tris,

Try this link for info on the Tony. Pretty comprehensive for all models of the Tony. It was NOT a licensed built Me-109 though the engine is a Japanese version of the Daimler engine.

http://www.csd.uwo.ca/~pettypi/elevon/b ... ki-61.html

Note that the K-61 KAIc was a very good airplane, when the engine operated at it rated output.

The author, Joe Baugher, has also listed the references used to get this information.

Chez

Steve, I didn't mean to be abrupt with you. Let me put it this way.

I've looked at Joe's work over the years--or the past few years anyway, not sure how long I've been aware of his stuff online, but it's been some time. I've found that if it has any use at all then it's to get a "rounding" of opinion on performance statistics should I lack for these, as I've found Joe, unfortunately, to mostly quote blindly from the same sources over and over, sources I've found through experience to be wanting in one area or the other.

The following link is to his page on Zeros: Mitsubishi A6M2 Zero Fighter

I use this link only as an example. Similar problems exist for Joe's work on other aircraft, something I've had to learn the hard way.

Following is an excerpt from Joe's Zero article and it speaks volumes as to the overall reliability of his work. (That is to say, his work isn't all wrong, but it tends to simply repeat bad wisdom passed down by other authors.)

     During the first year of the Pacific War, the standard shipboard fighter serving with the US Navy was the Grumman F4F Wildcat. The A6M2
     was superior to the F4F Wildcat in speed, climb rate, and maneuverability, but the Wildcat had better firepower and was more robust. In a
     dive the two aircraft were fairly equal, but the turning circle of the Zero Fighter was very much smaller than that of the Wildcat by virtue
     of its lower wing loadings.

Do you see the same problem I do? This is Joe's work, with his familiar citations of same-same sources at the bottom of the page--William Green and Rene J. Francillon are the most familiar players, almost always present.

The thing is, I realize that anyone who could seriously write that the A6M2 Model 21 was "faster" than the Wildcat (there were many Wildcats) and could "dive" with the Wildcat, that the Zero was more "maneuverable" than the Wildcat (these people never heard of rolls, I guess, or performance at various air speeds and altitudes--perhaps we should chip in and buy them a ticket to Air Warrior) either doesn't know his subject (i.e. is just repeating what he read somewhere) or he is . . . pulling Gary's leg. Indeed, Joe's description of a Zero vis-a-vis the Wildcat sounds suggestively custom-fit for the game, does it not? And does the game's air model work properly? Are you happy and satisfied with Gary's simplistic presentation?

So, my opinion is that in Joe's case it's almost always better to look elsewhere, if necessary many elsewheres, to arrive at that distillation of truth I mentioned, unless you happen to know for sure that what he's talkiing about makes sense. And I all too seldom have confidence in that, and even if I did then why would I go to Joe in the first place?

I suggest there is better truth out there than what the likes of Baugher and Green and Francillon have to offer, though, as I say, it does take time to get at it. No one ever said it was easy.

[Kereguelen]

ORIGINAL: Mr.Frag


Would you care to find a single encounter during the course of the entire war where 119 heavy bombers were sent in unescorted against a very strong Japanese fighter area.

Taking a discussion point that the B-17 was a hardy aircraft and could handle an attack by a couple A6M2's and walk away the victor and extrapolating it into 1000 B-17's flying in a tight formation with no manuvering room with full bomb loads would be able to take on 1400 fighters is just being silly yet here we are, 7 pages worth of posts.

In the example provided by Tristanjohn the Allies lost about 70% of the committed heavy bombers. The Japanese lost about 10% of the committed fighters.

Interesting loss ratio when comparing armament and armour of the planes involved. Maybe one could expect such results if unescorted Betties would have met Spitfires in similar numbers (or Bolos meeting Zeros). But B-17s meeting Zeros?

[Tristanjohn]

ORIGINAL: Mr.Frag

No, i think the results show there is a strong malus (negative bonus) for large formations.

It's a function of dice averaging.

If you say something has a 10% chance of happening and you only send 1 aircraft, the odds of that aircraft going down are 1 in 10. ie: you are virtually immune which you see with small numbers of aircraft. You are effectively defeating the dice by using a sample size so small.

When you send 1000 aircraft, you are saying that you expect 100 aircraft to be lost (10% of 1000). This 100 aircraft being lost looks shocking, but the odds have not changed one bit.

Larger battles will be bloodier as the averages come into play and you can't beat them anymore by submitting too small a sample size.

Math doesn't vary, Ray. It does what we tell it to do. Its principles and functions, except in more rarified mathematical air, tend to be absolute, and like the classic computer saw it's garbage in garbage out. What are you trying to sell us here?

If Gary's model has it that 1 in 10 bombers should be shot down by Japanese fighters then of course the results will turn out badly. We already know what the historical ratrios were, as these are a matter of public record, and these ratios did not come out 1:10 or anything like it. Even over Germany, where the Allies faced much sterner defense, both in terms of enemy fighters and AA fire, the historical loss ratios for heavy bombers bear scant resemblance to the ratios coughed up by Gary's model. I can't wait to see what the ratios will come out as when B-29s hop onto the board and all those massive raids you keep saying didn't exist in the Pacific theater all of a sudden exist. How many B-29s will be shot down by Zeros and the super Tony?

To design a decent model one needs to work with a large enough sample to begin with--one doesn't design an air model using just one-on-one combat scenarios exclusively, does one? Did nobody bother to check on the air-combat model to see what sorts of results were given from larger aerial combats? Is this a model which will only give reasonable results if only a certain number of aircraft are present on both sides? If so, what is that magical number?

My God, how much more obvious could it be that this air model doesn't work because it wasn't designed properly? You can argue against reality all you want, but you'll only continue to fool those who aren't capable to see the light of day for themselves. And what accomplishment could there be in that? Wouldn't you rather keep more sophisticated company?

This is a serious problem. Instead of this effort expended on spin control, why not devote your energy to finding a fix?

[Erik Rutins]

Tristan,

Yes, I'd like to run your example scenario/save here for my own tests. Please e-mail it to me at erikr@matrixgames.com. Thanks.

Before we get another pile-up on Matrix/2by3 here, I'd like to point out what seems to me to be obvious, but apparently may not be for some. For any sufficiently complex design, given enough time, you can find a place where it fails to be accurate. I don't know of any complex computer wargame that passes this test and many that are much simpler fail even more spectacularly.

As a player, my view on this is that the results are out of whack, but they reinforce one key lesson which is that bombers should not fly into CAP without escort. For smaller raids, the game seems to work fine. I'll be testing to see if it works fine when Tonys are not involved too. Ultimately though, while the losses should be lower for the bombers, the point ends up the same. In an actual game, with ongoing use, chances are the bomber and fighter morale and fatigue would not be as high. When both sides are at max morale and minimum fatigue, engagements tend to be more bloody.

Anyway, please forward the test save/scenario and I'll have a go at it. Thanks.

Regards,

- Erik

[Erik Rutins]

ORIGINAL: Tristanjohn
My God, how much more obvious could it be that this air model doesn't work because it wasn't designed properly? You can argue against reality all you want, but you'll only continue to fool those who aren't capable to see the light of day for themselves. And what accomplishment could there be in that? Wouldn't you rather keep more sophisticated company?

This, by the way, is exactly what I would like to see less of on this forum. Can you explain to me how this advances your argument?

Regards,

- Erik

[Tristanjohn]

ORIGINAL: timtom

Hate to say it, but if certain stallwart defenders of the status quo were to comment, they could not unreasonably argue that what (statistically) amount to a few random results doesn't prove anything.

I'd love to see some more enterprising than myself test the hypothesis that aircraft losses increase exponentially with increased numbers involved in combat.

We can check that easily enough. I'll simply reload the turn and stand-down some of those bombers. That was originally one criterion to my suggested test.

I have an appointment this morning. I might be be able to get around to that this afternoon. (I'm a week or more behind in my game with Mogami, and Joe's turn is now three days late if I'm not mistaken, but what the hell. [:D])