For problems of data compression, gambling, and prediction of individual sequences x1, ···, xn the following questions arise. Given a target family of probability mass functions p(x1, ···, x n|θ), how do we choose a probability mass function q(x 1, ···, xn) so that it approximately minimizes the maximum regret/belowdisplayskip10ptminus6pt max (log1/q(x1, ···, xn)-log1/p(x1, ···, xn |θˆ)) and so that it achieves the best constant C in the asymptotics of the minimax regret, which is of the form (d/2)log(n/2π)+C+o(1), where d is the parameter dimension? Are there easily implementable strategies q that achieve those asymptotics? And how does the solution to the worst case sequence problem relate to the solution to the corresponding expectation version minq max 0 E0(log1/q(x1, ···, xn)-log1/p(x1, ···, xn|θ))? In the discrete memoryless case, with a given alphabet of size m, the Bayes procedure with the Dirichlet(1/2, ···, 1/2) prior is asymptotically maximin. Simple modifications of it are shown to be asymptotically minimax. The best constant is Cm=log(Γ(1/2)m/(Γ(m/2)) which agrees with the logarithm of the integral of the square root of the determinant of the Fisher information. Moreover, our asymptotically optimal strategies for the worst case problem are also asymptotically optimal for the expectation version. Analogous conclusions are given for the case of prediction, gambling, and compression when, for each observation, one has access to side information from an alphabet of size k. In this setting the minimax regret is shown to be k(m-1)/2logn/2πk+kCm+o(1)

### Published in:

Information Theory, IEEE Transactions on  (Volume:46 ,  Issue: 2 )