The 3 events are not truely "independent", P(A,B,C)=P(A)*P(B)*P(C) is only true when no more than 1 prophecy has an accuracy <> 1/2 (more generally, no more than 1 prophecy with accuracy <> 1/N,where N is the number of possible outcome).
To illustrate this, let's consider an extreme situation.
if P(X)=0.5, P(A|X)=P(B|X)=1, and P(C|X)=.5 as it was.
P(A) = P(A|X)*P(X) + P(A|[X])*P([X]) = 1*0.5 + 0*0.5=0.5, similarly P(B)=0.5
P(C) = 0.5*0.5+0.5*0.5 = 0.5www.ddhw.com
according to P(A,B,C)=P(A)*P(B)*P(C), P(A,B,C)=0.5*0.5*0.5
However, obviously, P(A,B,C) =0.5*1*0.5 since when A occurs, B must occur.
Now let's consider another situation.
If the accuracy of all 3 prophecies are 0.4 instead (lower than 1/2 to satisfy those who are strongly against the fortune telling profession), that is, P(X)=0.5, P(A|X)=P(B|X)=P(C|X)=.4www.ddhw.com
I think P(X|A,B,C) = P(A,B,C|X)*P(X) /P(A,B,C) = P(A,B,C|X)*P(X) /(P(A,B,C|X)*P(X)+P(A,B,C|[X])*P([X]) )=P(A|X)*P(B|X)*P(C|X)*P(X) / (P(A|X)*(P(B|X)*P(C|X)*P(X)+P(A|[X])*(P(B|[X])*P(C|[X])*P([X]))=.4*.4*.4*.5/(.4*.4*.4*.5+.6*.6*.6*.5)=.23
Again if P(X)=0.5, P(A|X)=P(B|X)=P(C|X)=.4, P(A)= P(A|X)*P(X) + P(A|[X])*P([X]) =.4*P(X)+.6*(1-P(X))=.4*.5+.6*.5=.5, Similarly, P(B)=P(C)=0.5
so when P(A|X)=.5, P(A) always = P(X) as you pointed out, however, when P(A|X)<>.5, P(A)=P(X) only if P(X)=.5)www.ddhw.com
Afterall I don't think we need P(A),P(B) or P(C) to solve this problem.
And also if P(X)=0.5, P(A|X)=P(B|X)=P(C|X)=.8,
P(X|A,B,C) >1, if we apply P(X|A,B,C) = P(A|X)*P(B|X)*P(C|X)*p / P(A)*P(B)*P(C).
www.ddhw.com
As a conclusion, whenever the accuracy of a prediction = 1/N, where N=the number of possible outcomes of a trial, then the prediction has no value and has given no additional information and hence useless. It doesn't matter if the outcomes are equiprobable or not. Actually you and 贾明 has already pointed it out.