# Black Litterman - numerical instability

I am trying to work out the formula for the posterior mean in Black Litterman's model assuming 100% confidence :

(Equation 9).

My problem is that the covariance matrix Sigma of the daily returns i'm using is very small such that when I take the inverse of the required expression the values are exploding...can anyone suggest what to do in such a case?

Thanks!

• Yes, the daily returns covariance matrix will be small, but that should not necessarily be a problem if all of your inputs are on the same scale ($\Omega$, $\Pi$). It's probably easier to convert everything to an annual scale. – Tim Wilding Aug 27 '18 at 12:14