|
发表于 2009-12-13 04:04 PM
|
显示全部楼层
本帖最后由 Bayside 于 2009-12-13 16:07 编辑
即 0.7777777
A 的胜率 X, B的胜率 Y
X*Y/(X*Y+(1-X)*(1-Y))
老黄 发表于 2009-12-13 14:35 
Baysian Approach
P(Market=Up | Sys1= Up & Sys2 = Up)
= P (M=Up, S1=Up & S2= Up) / P (S1=Up & S2=Up)
= [ P (M = Up) * P (S1=Up & S2=Up|M=Up) ] / [ P(M=Up)*P(S1=Up&S2=Up|M=Up) + P(M=Dn)*P(S1=Up&S2=Up|M=Dn) ]
Assume P(M=Up) = P(M=Dn) = 0.5, and S1 and S2 are independent,
then we obtain
P(S1=Up|M=Up) * P(S2=Up|M=Up) / [P(S1=Up|M=Up) * P(S2=Up|M=Up) + P(S1=Up|M=Dn)*P(S2=Up|M=Dn) ]
= x * y / ( x* y + (1-x) * (1-y))
comments:
final formula symmetry for x and y
final formula symmetry for x_i and (1-x_i)
if multiple systems, then P = X1* X2 *... *Xn / ( X1* X2 *... *Xn + ( 1-X1) *(1-X2)*...(1-Xn) )
a hiddlen assumption is that corrective rates in origianl question( 60%, 70%) are the same for up and down market |
评分
-
1
查看全部评分
-
|