How much a control variate helps

You price a derivative by Monte Carlo (payoff $X$) and use a control variate $Y$ with a known mean — you subtract off $\beta(Y - E[Y])$ using the optimal coefficient $\beta^\star$. You measure $\operatorname{Var}(X) = 25$, $\operatorname{Var}(Y) = 9$, and $\operatorname{Cov}(X, Y) = 9$. After applying the optimal control variate, what fraction of the original variance $\operatorname{Var}(X)$ still remains? Give it to four decimals.