Let X = Y×T , and suppose that P0, P1 are two probability distributions given by dP0(y, t) = f(y)g(t) dµ(y) dν(t), dP1(y, t) = h(y, t) dµ(y) dν(t), where h(y, t)/f(y)g(t) < ∞. Then under P1 the probability density of Y with respect to µ is pY 1 (y) = f(y)E0 h(y, T)

f(y)g(T)

%

%

%

%

Y = y



.

pY 1 (y) = 
T h(y, t) dν(t) = f(y)

T h(y, t)
f(y)g(t)
g(t) dν(t).]
Section 2.6