there are three meta-learners which
perform better on multi-class datasets and three meta-learners which perform
better on two-class datasets – as would be expected by chance. The observed
differences in performance are at most 1%. Average accuracy does not agree
with this conclusion, but it is an unreliable indicator at best.