Open girving opened 5 years ago
Nice article, well written and easy to follow! Adding on to @girving comment, I came across the same issue and I believe the next plot (for conditioning & marginal distribution) has a similar issue. The initial central plot shows a joint distribution with the off-diagonal terms of the covariance matrix as zero, in which case the mean for the conditional distribution should just be zero independent of the conditioning value, yet it shows as 0.96 Again this issue is fixed once one starts dragging. Initial:
With some dragging done:
I stared at this figure in confusion for a while since it was axis aligned but had nonzero offdiagonal covariance, then finally realized the inconsistency goes away once you start dragging.