1

The point of this question is only to check that I still know how to do statistics.

So in Assassin's Creed Rebellion, there are 9 character classes. And each class has a unique resource which can be turned into a weapon or armor etc.

In the previous event round, I noticed that the second to last treasure (a resource) you could collect matched the class of the character for whom you could collect "DNA" tokens (essentially steps to the next level for the character), and it made me think there might be a correlation. Now in the current round I have noticed the same thing.

As I understand it, the chance of this correlation happening twice randomly is 1 / (9 * 9), or 1.23%.

So from this I think I can conclude with 2 sigma certainty that there is a correlation. One more round which again matches and the random chance drops to 0.137%, or 4 sigma, right?

All comments welcome as to how I am using statistics and whether my conclusions are valid

ControlAltDel
  • 175
  • If you roll two 9-sided dice, the probability that they match is $1/9$, not $1/9^2$. – Karl Oct 26 '23 at 15:39
  • 1
    @Karl Thank you for your comment. I have updated my question. You are correct (as I understand it) in what you state about rolling two 9 sided dice. But what I'm saying is this was true for two events / 2 trials, so that's why 1 / (9 * 9) – ControlAltDel Oct 26 '23 at 15:49
  • I see. Another thing: when you say "the second to last treasure", that suggests there are other events where you haven't noted a correlation, which you aren't taking into account. See https://en.wikipedia.org/wiki/Multiple_comparisons_problem – Karl Oct 26 '23 at 15:56
  • 1
    @Karl I mean "the 2nd to last treasure on that specific event". Does that clarify? Yes, there have been other events... But I wasn't looking/checking for this correlation. I only started looking for it with the previous event. It could have been true or it might not have been true. All I have at this point in terms of inference is the last two events. I hope this clarifies what I am asking. – ControlAltDel Oct 26 '23 at 16:04
  • I am missing a few words in my last comment. I meant "[Previous to when I started looking for the correlation], it could have been true or it might not have been true" – ControlAltDel Oct 26 '23 at 16:21

0 Answers0