Results of 10 Runs of 7 Dec RHSCVO 2.42

[DD696]

This is the losses of aircraft in a seris of test runs. Losses are by air-to-air, destroyed on field, destroyed by flak, and operational losses. Ships sunk are allied ships sunk. Note that PG Wake is generally always sunk on turn one.

Allied A/c Losses Japanese A/C Losses Ships Sunk
17-399-3-0 4-0-83-0 1
8-429-2-2 0-0-68-0 2 (1 BB)
2-424-0-2 1-0-66-0 1 BB
6-320-0-0 4-0-69-0 2
5-400-0-0 5-0-96-0 1
8-399-2-2 1-0-64-0 17 (1BB)
5-411-3-0 3-0-87-0 1
13-396-2-0 6-0-65-0 2
6-450-2-2 0-0-76-0 4 (1 BB)
4-418-2-8 0-0-67-0 5 (2 BB, 1 Ca)

On run 6 Hong Kong fell resulting in the scuttling of the UK fleet there.

Japanese aircraft losses are primarily from the fleet carriers at Pearl Harbor, with a fairly consistent 2-4 loss of other aircraft at locations other than Pearl Harbor.

AI set at very hard, first turn surprise on.

The Naval Historial Center states that there were 29 Japanese aircraft lost at Pearl Harbor on 7 Dec 1941. My ten runs of testing are presented for all to see. Are the flak losses too high? These firgures indicate that the Japanese have lost the equivalent of one major fleet carrier's aircraft complement. I never came anywhere close to the actual losses of Japanese aircraft. I have no idea how the US losses stack up against the historical figures. Do I need to make additional test runs, or is this 'good enuf'?

[witpqs]

On average they are coming out higher than the historical PH flak losses. If the overall result is a better modeling then the inaccuracy in the PH attack is worth it.

[el cid again]

Note that PG Wake is generally always sunk on turn one.

Which is at least fair history. The David Adams Memorial Chapel at Norfolk Naval Base is named in honor of its captain...

[el cid again]

Are the flak losses too high?

Unless there is NO flak at all on the first strike - you are not simulating history. There was no flak from Army batteries until ammunition was provided and the guns were manned. I have not noticed that the game does this - it lets the flak fire - although presumably at reduced effectiveness.

Anyway you cannot measure flak attrition in this way. You need to include a lot of information and do a lot more runs.

First, no number of tests under 30 ever is of statistical significance.

Second, the flak losses are going to be a function of attack profiles. At Clark AFB the bombers (really) came in at 24000 feet - and fuses limited flak to 22000 feet - so they took zero losses to it. IF you come in at 24,000 feet you may find the same thing - vs land guns anyway. If you come in low, even mg's will be shooting and sometimes hitting. In between is, well, a mixed case. I have set severe limits on flak effective altitude compared to stock and CHS - effective ceilings instead of maximum ceilings. So altitude effects will be quite significant - and under player control. You can arrange to have "too high" losses to flak - or none at all - just by the attack altitude you pick. So a meaningful test needs to be done in a specific context - one where you can compare real data and simulation data for the conditions. And then it needs to be done a statistically meaningful number of times. Just saying "they lost xy planes at Pearl Harbor and we lost ab planes" is meaningless - In fact there were two raids on Pearl Harbor day and they suffered very different casualties because the second raid found defenders alerted and supplied with ammunition.

Another question is this:

there are three ways to lose airplanes: flak, air combat, and operational attrition. You don't tell us how many were lost to each cause. Do you think ALL the losses were to flak? If so, why?

Yet another is this:

a real player should not let the AI manage the attack. Depending on what you want to damage and how much you are willing to suffer casualties, you will pick altitudes and missions accordingly

You are not telling us what your choices were? Letting the AI manage this is an invitation to less than optimum results - whatever your goals may be.

To understand flak alone you need a test situation where there is no enemy air opposition and minimal cause for operational attrition. Then you need to run various flight profiles - and you will get various results - accordingly. This is a fair model - if you let it work. But it isn't just "I flew there and took that many losses"