Originally Posted by
FFlash
Heh, while on games, I love Bayesian probability, there is nothing wrong in believing in it!
If throwing dice and getting 9/10 times number 6, I believe the probability of getting number six is higher that 1/6!
What you are describing is actually a frequentist view. Bayesians would likely have a high-kurtosis prior for theta (assuming bernoulli trials) centered around 1/6. The posterior would thus still be around 1/6 despite the likelihood of the observations.