Criticality not negative

[dsmic]

I had an idea to use Criticality in playouts to avoid playing in sekis.

I recognized, that it is forced to be >=0, and this since 2011 (or probably ever)

Is there a reason? Was this in the paper?

I think it is usefull information-> if I play my opponent owns the point...

[Francois van Niekerk]

I had a quick look at the literature and the code. I agree, it seems like this is might be a bug.

Something to bear in mind is that the criticality is equivalent to the covariance (just a constance factor of 2 difference).

Either way, the configuration that results in better performance should be used (but further investigation should be considered if it is unexpected).