Drift traces look strange

[aboharbf]

This seems obviously incorrect. What might be going on? I Imagine the resultant sorting can't be correct if it thinks the probe is shuffling back and forth like this.

[rajatsaxena]

another possibility could be that there are very few or no spikes in those batches. I would recommend increasing the batch time and rerunning to see if that helps.