ColinWright
Posts: 2604
Joined: 10/13/2005 Status: offline
|
quote:
ORIGINAL: JAMiAM quote:
ORIGINAL: ColinWright quote:
ORIGINAL: a white rabbit ..British figures give 10% losses on bombing raids as heavy, from all causes.. ..British estimates are for most of WW2 it takes 18,500 shells to down one bomber, divide confirmed AA kills by shells used (pre proximity fuse) ..given the deterrent option i'd say the toaw figures are in the right area, for everything except bridge attacks and airfield attacks. It can't be hard to program AA to know when its in a bridge/airfield hex, other unit types know where they are, why not AA ?.. ..oh and 2 Swordfish downed at Tarranto, despite the available AA.. There you go. TWO Swordfish downed at Taranto. OPART losses of four. Now, go ahead and set up some AA guns at 10% proficiency. See how few you'll need to shoot down down four. It'll be something like ten barrels. Say, maybe a fiftieth of the flak the Italians probably had, between the ships and the port defences. See the Seelowe thread at TDG. The figures are way out of line. If one reran the 1940 strike on the Meuse bridges, for example, it wouldn't be a matter of a third of the strike aircraft being shot down between the fighters and the flak -- the fighter pilots can go get coffee. In OPART, the German flak would suffice to shoot down every attacking aircraft ten times over by itself -- and I don't think that's an exaggeration. For the time being, I've divided the AA values by anywhere from three to five, and losses are still suspiciously heavy. The essence of the problem is that OPART doesn't reflect the primary effect of flak -- it plays hell with the aim. So we get flak that either (a) has little value at all, or (b) achieves its effect in entirely the wrong way. Flak in OPART is never going to work right if it simply functions by shooting down planes. Some things to keep in mind for this debate... TOAW III does abstract the aiming issue that Colin raises, by the different proportion applied to "return to inventory:losses" in the disabled results for airframe equipment. That higher proportion of returned equipment, relative to other equipment, represents damaged airframes, as well as those pilots who drop their loads off target. Are you stating that more aircraft are returned to the pool if losses are inflicted by flak than if by other aircraft? More than the half that is returned for losses suffered by other types of equipment? It did just occur to me that if reducing the effectiveness of aircraft by the amount of flak in a hex imposed insuperable programming problems, that increasing the proportion of 'losses' returned to the pool might offer an alternative approach?quote:
quote:
Much testing needs to be done to get the numbers "right". The "right" numbers are themselves in question. The cases being raised are, in themselves, statistical "outliers" which do not represent normal operations. The attack on Taranto achieved a good level of tactical surprise. Saying that the results, in losses to ships and or aircraft is somehow typical is like claiming that the Pearl Harbor attack was typical of WWII Pacific naval operations. How to model these eminently historical, but atypical tactical results in an operational game might generate an entire new subsection in the Wishlist.quote:
I don't think the results at Taranto are especially atypical. The message seems the same as that in the results obtained at the Sedan bridgehead and everywhere else I've looked: flak losses are way too heavy. Just set up the Schweinfurt raid or something. I'll bet the Germans won't need to put a fighter into the sky: I imagine the flak at Schweinfurt will blow all three hundred B-17's out of the sky.quote:
Further regarding the right numbers, the testing must be set up rigorously and take into account a variety of AA values, proficiency, readiness, environmental, attrition divider, and hex-scale (density) settings. They must take into account, losses and return to inventory. They must be run with enough repetition to generate a good base of results. Averages and the range of the data is important, as well as the distribution of the outliers. The funny thing is though, the test data would likely be more comprehensive than the historical results of the last century, and as above, interpretation of what should be considered a "good correlation" is subject to some debate. Yeah. I'm certainly not arguing that numbers should just be pulled out of a hat. However, I'd insist that the results we're getting now are totally out to lunch: we've got nowhere to go but up. Anyway, I see three approaches. 1. As you seem to imply currently happens, a gretly increased proportion of all 'losses' could be returned to the pool. 90%? 95%? The number would need to be researched. 2. Divide the attacking strength by the amount of flak in the target hex. As I discussed, the effect should not be linear: even a few flak pieces should seriously attenuate the effectiveness of bombers, or to put the same point differently, troops without any flak at all are really at the mercy of attacking aircraft. 3. If this is impossible, it should be practical to make the presence of any unit with a flak icon have the same effect on bombing effectiveness as clouds. Obviously, this would be a rough approximation, but it would be an improvement over the current situation, and the programming requirements would seem to be minimal.
< Message edited by ColinWright -- 10/4/2007 12:20:07 AM >
_____________________________
I am not Charlie Hebdo
|