This message was deleted.

Hi Graeme, excellent name. I'm not sure its in our docs - we probably have it in the pull request. We detected a slight bias in the splitting results if the experiment names are very similar. v2 corrects this.