[Consim-l] Analysis of Coral Sea

[Brian McCue]

> I'm not sure how you'd do it, but I would be more impressed by building a 
> model based on how you expect things to work, then running it and seeing if 
> the historical outcome ever appears.

This might be possible in some type of combat (e.g., WW I battleships (_Seekrieg_
represents such an attempt), or WW II tank warfare (AH _Tobruk_)),  but with pilot
skill counting for so much and with aerial dogfights being so complicated, I wouldn't
pick Coral Sea as a battle to try modeling that way.

That said, I wanted my group to try playing _Flat Top_ and enacting the combats in
_Dauntless_.

> For instance, I'd assume historical Midway was not 
> an "average" outcome at all.  It might not be truly a "miracle", but certainly 
> it was a better than expected victory for the US

The cited "miracles" are usually above the attrition-model level; variously either
the preparedness offered by the decryption, or the multiple instances of lucky timing
that worked out just right.

An additional problem in modeling attrition combat is that most attrition combat
models, and very possibly the combats themselves, are dynamically unstable in the
sense that (espeicaly if the sides are evenly matched), any slight random variation
in favor of one or the other will snowball into a larger advantage. This would occur
in the Salvo Equations if they involved rolling dice and if they were applied
repeatedly. Another example, perhaps familiar to many in this conversation, is the
ground combat in XTR Victoy at Midway.

Brian


 
____________________________________________________________________________________
The all-new Yahoo! Mail beta
Fire up a more powerful email and get things done faster. 
http://new.mail.yahoo.com