Matrix Games Forums

Matrix was either right - or close - in its calculation of operational attrition. All of us critics were wrong - or substantially wrong - in spite of our ability to be very sincere and precise in our measurements of it. There was close to a consensus on the board, and actual agreement by imperical tests, that operational attrition was way less than 1% of historical values: probably less than 1/10 of 1%: almost certainly less than half of that: not less than 2000 times too low, probably more.

And, by our lights, we were quite right. Turns out those numbers were conservative: it should be more like 4000-5000 times too low the way we were thinking about it - for Japan - and worse for the Allies.

What we got wrong - aided and abetted by the peculiar statistics WITP gives us - is that operational attrition is usually stated in losses per 1000 aircraft per unit of time (week, month, year) - NOT in terms of losses per sortee. Matrix algorithm turns out to be based on the only thing possible - it must decide if THIS flight is attritted? - on the latter. It duely reports sortee count and operational losses. And we all miscalculated them, not realizing that we were comparing golf balls and beach balls.

For example, JNAF has a total of 7619 planes in stock (2457 at start not in pools, 3557 reinforcement, the rest in pools); CHS155 had 7717 and RHS 7839. Similarly, JAAF has a total of 5822 in stock, 5760 in CHS155 and 6153 in RHS (starting with 1627, 1898 and 2249 on the board, respectively). Initial operational attrition rates must be multiplied by the number of planes to yield the actual loss % over a unit of time: that is 4084, 4615 or 5432. Eventually the number of planes in service is probably twice as high, although it depends how many units are lost or disbanded how many are really operational: these are what Joe calls "plane slots" - places for planes in operational units. The theoretical maximum (if every unit exists) are 13441 for stock, 13447 for CHS155 and 13992 for RHS. [Remarkably close data - the differences being mainly that pools are larger in stock and some types do not exist except in RHS - e.g. gliders]. Allied numbers are much higher. IF all these planes are generating sortees, the operational loss rate is the number of planes lost compared with the total number of planes flying - not compared with the number of sortees generated. The distinction is a factor in the thousands.

Operational losses may indeed be (have been) too low - but it will take a different perspective to figure that out: we need to understand what we are measuring - and use very large samples - since it depends on things like the size of the field you are using - so small data samples will not yield overall values.

Sid,

That's very good to read. Without being schooled in how ops losses are measured IRL, I did wonder how mutiplying what we had by 2,000 would be sustainable. This new info makes sense.

Hey Elf, your PM box is full...

Interesting. But I don't belive it yet. Have we not seen reports of loss rates with units of percent? Loss rates reported in percent are inconsistant with rates expressed in parts per thousand per unit time.

Another interesting tidbit. My sources suggest US pilots had about a 7% chance of not returning from their first combat patrol, with a decrease to just below 1% after about 10 missions. Operational losses were a significant fraction (at least 1/4) of that 7% and declining risk. This is entirely consistant with our prior expections. If the loss rate for accidents was really per 1000 sortie weeks rather than per 100 sorties, almost 100% of the risk to pilots would be from being shot down. Operational losses would be extremely rare. Thats just not right. I'd guess the operational loss rate per sortie must have been around 0.1% to 1% (0.1 to 1 per hundred sorties) and the risk of being shot down was near 1%. In other words I don't belive the ratio of combat losses to operational losses could be between 10:1 and 1:1 perhaps close to 4:1.

One last thing. The Japanese ferried aircraft to island bases because the operational losses were to high to fly them. Again consistant with 1% loss rate but NOT 0.0001%.

Me editing is horrible.

"In other words I don't belive the ratio of combat losses to operational losses could be between 10:1 and 1:1 perhaps close to 4:1."

SHOULD SAY ....I DO BELIVE....not DONT BELIEVE

Interesting. But I don't belive it yet. Have we not seen reports of loss rates with units of percent? Loss rates reported in percent are inconsistant with rates expressed in parts per thousand per unit time.

No we have not. IF loss rates APPROACHED 1% - fliers would not fly.
THAT we have seen.

You may be confusing loss rates due to enemy action with operational losses. Different subject. This is PURELY an operational loss discussion. And IF you lose, say, 10% per month, and fly 2 times per day with a multiple per cent loss rate per flight- you will have no fleet left! You don't mean to say that. 60 flights per plane times 100 planes = 6000 flights times, say, 3% losses = 180 planes lost - but you only had 100 planes to lose!

My sources suggest US pilots had about a 7% chance of not returning from their first combat patrol, with a decrease to just below 1% after about 10 missions. Operational losses were a significant fraction (at least 1/4) of that 7% and declining risk.

Not really. It never appraoches 1%. So it is not a major component. Further, if it was, it would mean bad training policy.

So you will know the source, the 7% chance of a U.S. pilot being lost on his first mission is taken from Victory in the Pacific on the bottom of page 257.

As to pilots not flying if they had a 1% chance of not returning to base, they certinly did alot of that, mostly becuase of being shot down. My father was there on Black Thursday when the USAAF lost alot more than 1% of the attacking force. The 8th AAF had a period where the chances of a crew completing 40 missions were not very good. I think bomber command had a similar period.

Also, it is true that a 1% loss rate would result in lossing more than 100% of a force over a short period of time. Thats what replacements are for.

Operational loss included planes scraped due to no replacement parts. Just read anything on the flying tigers, a 100 planes didn't last very long, this didn't mean they lost a 100 pilots, most of the planes where used for parts. Operational loss in WITP most times include both the plane and pilot.

The issue is this, when playing Japan you only get 20 trained Naval pilots a month, in the game you lose 30 a month to operational losses. So if you do nothing, Japan loses all of its Naval pilots just to non-combat operations, no army or navy operated like that.

Both issues need to be addressed at the same time. Monthly operational loss should not exceed the trained replacement rate.
Japan loses 60-75 planes on Dec 7,1942 most are naval pilots. So Japan losses 1 month of naval pilots on day due to combat, then another 30 in 30 days to operational losses.

That's 3 months of trained pilots for 1 day of combat and 30 days of non-combat ops. Uping just the operational losses will completely break Japan.

This 20 Naval pilots is bogus, it's true for Japanese carrier pilots in history but in WITP this also includes pilots all of the naval flying boats and naval land based bombers.

Does the Op Losses process used (right or wrong) make players rest their units a lot more than is currently being seen.

Stand Down rather than Training, and not every rear area Bomber can do ASW all the time, also the Japanese plan of training through bombing isolated bases would also incur Op Losses.

Also in the early years the Japanese using the longer ranges of the Zero & Betty probably increase the odds of a bad Op Loss roll.

So you will know the source, the 7% chance of a U.S. pilot being lost on his first mission is taken from Victory in the Pacific on the bottom of page 257.

And so you understand my position, I do not think there is anything wrong with that figure. I note it carefully uses the qualifing phrase "first combat mission." It is nothing like a non-combat mission. Just so. Apples and oranges. Different subject for a different discussion. I have no problem with the figure - and it has nothing to do with the issue. Nothing at all.

As to pilots not flying if they had a 1% chance of not returning to base, they certinly did alot of that, mostly becuase of being shot down.

Again you are really getting topics mixed up. The thing I said - and I happen to be quoting - is that we saw reports that pilots refused to fly - consistently and in all services of all nations - if OPERATIONAL loss rates were a major fraction of 1%. This has NOTHING WHATEVER to do with losses due to enemy action. Different subject. Try to get it: operational attrition is the subject and it has NOTHING to do with enemy action.

My father was there on Black Thursday when the USAAF lost alot more than 1% of the attacking force. The 8th AAF had a period where the chances of a crew completing 40 missions were not very good. I think bomber command had a similar period.

My father also flew USAAF bombers during this period. I grew up surrounded by aircrew of that service in ETO missions. And my mother ALSO flew in B-17s!!! During the war!! Her primary job was training of gunners and bombradiers. But her emergency job skill was aviation photographic interpretation. When she joined - in 1942 - the USAAF was worried we might be thrown off Europe and Asia altogether - then invaded! They taught her how to make cameras from shoe boxes, film and developer chemicals from what you can find in any house, and how to identify things from the air - by flying. She was the only member of the first class of WAC photographers to identify a peculiar object deliberately done to be confusing: a tractor plowed in circles creating a very strange field. I have no problem respecting USAAC people - it is part of my personal heritage. And I know they were willing to take risks.
This subject is not about that. It is about reasonable flying odds. And it was demanded by pilots - in all nations - that OPERATIONAL risks be substantially below 1% per sortee. Typical figures are in the low tenths of % and figures below one tenth were not particularly unusual.

What people have got confused - you are not alone - is that loss rates per size of force - were typically measured in whole number %. 10% per month for a major unit is not wholly outrageous as a rule of thumb (although it is high) operating PTO from forward, unimproved fields. This is total operational losses - number of planes lost compared to force size.
That is, you might lose 10 planes out of 100 in a month of flying. But a typical unit would fly its planes twice a day - 6000 sortees - and tactical planes might be able to sustain 3 flights a day - a rate of 9000 sortees a month. To lose 10 planes out of 6000 is BOTH 10% of a force of 100 planes AND 0.16% per sortee: between one and two tenths of one per cent. And that is HIGH. Often it will be LESS than that. Lots of things matter: payload, were planes damaged in battle,
runway conditions, weather, and other factors. But this is the range of operational losses - from about .4% down to 0.4% - for an order of magnitude range - if I remember USAAF materials correctly.

Also, it is true that a 1% loss rate would result in lossing more than 100% of a force over a short period of time. Thats what replacements are for.

Yeah - but your chances were not zero of living through a tour - and your loss rates would insure that even if you never saw enemy action!

Both issues need to be addressed at the same time. Monthly operational loss should not exceed the trained replacement rate.

Why not? There is no theoretical connection at all. What if they were?
They certainly could be. There are certainly air wars in history in which NO replacements existed - but operational losses were still a risk.

IF there is a problem with replacements, it needs to be looked at on its own merits. Replacement rates are what they are. Operational losses are what they are. There is no relationship between them. And no reason either might not be higher than the other.

Japan loses 60-75 planes on Dec 7,1942 most are naval pilots. So Japan losses 1 month of naval pilots on day due to combat, then another 30 in 30 days to operational losses.

There is a good deal of Japanese material indicating the Japanese Navy had far too low a pilot training rate. I think this may be good simulation.

For fun I looked it up for you: The training rate in July 1941 (mobilization for Japan) was 84 pilots per class (actually divided into two classe). I think the classes began as 60 students (in ground school) and of these 50 were sent to flight school - of which only 42 would normally graduate. Two calsses means they were starting 100 men a month and getting 84 at the other end of the pipeline - and this only talks about basic flight training. There were two more levels of flight training - then you went to a specialist school (e.g. fighter, bomber) - and after that you had to spend two years in second line units before being considered worthy of joining a first line unit. I don't know how many came out of that stream - but the rate was too low. Too few started - and far fewer completed.

I totally agree with your argument . In fact I beleieve I asked the same question about a week ago. Perhaps I did not explain the question as eloquently as you have just answered it .

However I agree operational losses approaching 1% on average for front line pilots on the Allies side were not sustainable. Yes faults , yes problems with equipment but planes falling out of the air no way.

However I agree operational losses approaching 1% on average for front line pilots on the Allies side were not sustainable. Yes faults , yes problems with equipment but planes falling out of the air no way.

Someone, in another thread, listed what would be termed an "Operational Loss"

This theatre provided some of the worst weather on earth for flying Aircraft in Combat Missions, Weather, Terrain,(Groups of Aircraft going missing in Monsoon/Typhoons happened, The pilot of Wingate's B25 flew into a Mountain)) Distances start to scratch the surface. Maintenance on dirt/mud strips, RAAF Spitfires getting their Engines replaced in the open were common.

I Op Loss per Sqn per Mission (Assuming 10-12 per Sqn) would push 10%, so 1 loss per 10 missions is around 1% losses and IMHO, far too low

That is, you might lose 10 planes out of 100 in a month of flying. But a typical unit would fly its planes twice a day - 6000 sortees - and tactical planes might be able to sustain 3 flights a day - a rate of 9000 sortees a month. To lose 10 planes out of 6000 is BOTH 10% of a force of 100 planes AND 0.16% per sortee: between one and two tenths of one per cent. And that is HIGH. Often it will be LESS than that.

Excellent explanation of the math!

From a thread I started about a year ago...

tm.asp?m=876848&mpage=1&key=Operational%2Closses&#

I can't vouch for the source but I recently read in

Dunnigan and Nofi, Dirty Little Secrets of World War II, New York,
Quill William Morrow, 1994 p. 291

"The war in the Pacific was notable for the high number of aircraft destroyed by 'noncombat causes.' Overall, only 25 percent [emphasis added] of the aircraft lost were due to enemy action. The others were destroyed by the weather, the difficulty of operating from aircraft carriers, hastily built airfields, and the insidious effect of tropical climate had on machines. These losses were higher in the Pacific than in other theaters because of these unique conditions. And sometimes it got worse. In the fourteen month Aleutian Islands campaign in Alaska, 87 percent [emphasis added] of the aircraft losses were to noncombat causes..."

To back this quote up I checked the US Strategic Bombing Survey and found this:

p.90 Summary of Air Effort Against Kiska and Attu, 10 June 42-15 August 43
Total Avaiable US Aircraft 236
Lost to Combat - 41
Operational Losses - 184

p. 91 Summary of Air Effort for War in the North Pacific 7 Dec 41- 15 Aug 45

11th Army Air Force
A/C Attack Missions flown - 7318
A/C Lost to Enemy Action - 63
A/c lost in Theater - 409

Navy Fleet Air Wing (Land and Water based)
A/C Attack Missions flown - 704
A/C Lost to Enemy Action - 35
A/c lost in Theater - (?)

p. 331 Detailed Losses (in ref: to Okinawa Campaign)
US A/c Lost to Combat 458
US Operational A/C losses 305

IJ A/c lost to Combat 4155 (1900 suicide)
IJ A/c Operational losses 2655