If you want to do that analysis, I'd be very interested to see it / hear how you do it. All the data is in the GitHub repo, for example https://github.com/markf

If you want to do that analysis, I'd be very interested to see it / hear how you do it. All the data is in the GitHub repo, for example https://github.com/markfickett/dicehistogram/blob/master/data/151023d20chessexgeminicoppersteel/labels.csv https://github.com/markfickett/dicehistogram/blob/master/dat... lists the rolls for one test. I ran bootstrapped subsampling to estimate confidence intervals (see comment above), not sure how similar that is to what you're asking about.