Neat. Would you add the option to normalize the elo over time (e.g update the model used as an anchor for the elo computation) so the diff between labs is more

Neat. Would you add the option to normalize the elo over time (e.g update the model used as an anchor for the elo computation) so the diff between labs is more visible?