3 ms·

The mechanical rolling and computer vision is impressive, but wish he had noted if the results were statistically significant. Just a final touch to the article

The mechanical rolling and computer vision is impressive, but wish he had noted if the results were statistically significant. Just a final touch to the article. At 100-150 rolls per face, my gut is saying no. I guess I'm going to have to whip up a couple loops in python after work to satisfy my curiosity.

jcranmer 8y ago
The usual gut reaction for statistical significance is that n*p should be 20, or 20 rolls per face. 100-150 should be statistically significant.
- eterm 8y ago
  You can't assign significance to a sample size alone. A sample size of 10 can be significant (if it were 10 1's for example), and a sample size of 100,000 can be not significant, for example if you were to roll 4,953 1's.
  - dekhn 8y ago
    i believe the op meant "Statistical power"
eterm 8y ago
Yes, if these graphs had confidence intervals they would be a lot more informative.
- markfickett 8y ago
  As the author -- I agree! I did some preliminary work to make confidence intervals, but didn't get around to updating the writeup. These are my draft updated charts: http://www.markfickett.com/tmp/dice/chart.html http://www.markfickett.com/tmp/dice/chart.html I used bootstrapped subsamples to generate the confidence intervals: https://github.com/markfickett/dicehistogram/blob/master/summarize.py#L171 https://github.com/markfickett/dicehistogram/blob/master/sum...
Fomite 8y ago
One of the dangers, especially with statistical tests of uniformity, is that they often over detect departures from uniformity when using large samples.