So say if I have a set of data that says I have
335 successes out 587 trials
Now if I assume the likelihood of having a success is .50 (null hypothesis) my p-value for this set of data is
.0007
Obviously a very unlikely event and is considered statistically significant.
But if I reset the likelihood to being .53 (null hypothesis) the p-value then becomes
.0527
The null hypothesis is accepted.
My question is what should one's null hypothesis be when striving to create a model? P(Win)=.55? P(Win)=.56?
Null Hypothesis (Ho):Game; P(Win)=???