Interpretation of COTH results

Our laboratory is interested in PPI predictions, so we have started using
COTH. However, we have some issues with interpreting COTH results. Given
that COTH produces top ten complex threading results with associated Z
scores, it is not quite clear how many of the results should have Z score
above 2.5 to consider prediction as reliable one. For example, if the
first result has Z score 2.7 and all others are bellow cutoff value, is it
reasonable to assume that prediction is probable?

Thank you in advance!