Reality of AI dodge, complete statistics
I recorded every single attack in my GW tonight. And the initial results are promising. I did not count Sid or Phasma's AOE attacks. Because they do so little damage, I can't tell when they hit or...
Forum Discussion
"DjangoQuik;479119" wrote:"Krychek;479105" wrote:"DjangoQuik;479078" wrote:
I appreciate your effort but this sample size is far far too small to be considered statistically significant. You need to do this every day for a few weeks at least before it's going to be even close to conclusive.
Not necessarily. The great thing about sampling is you can get a really good idea of where the statistical mean for a set of numbers lies based upon a handful of data points.
The chance of anomalies in such a small sample is high and the effects on the results, when some of the datasets used here have only single digit figures, would be massive. The confidence interval from these results is totally unacceptable.
I agree with the smaller sets, I was simply looking at the overall numbers. The smaller sets I automatically ignored since they cannot show any significance.
