[BostonCard]

Well, it is not really a valid hypothesis because the data was cherry picked.  And nobody made a serious accusation of the NFC systematically cheating.  But, hypothetically, if systematic NFC cheating had been an issue, you could test the hypothesis that the NFC disproportionately won the coin toss (but again, you would need to use non-cherry picked data, which is the data on all the Super Bowls).  [As an aside, note the tendency to cherry pick data by some people here; at least one poster wanted to consider only the data after we made the final four].  On the other hand, people are suggesting, and I am quoting, that "Monty the Traitor can't win the big games in the tourney and it's time for both parties to move on".  That is a testable hypothesis; the null hypothesis is that Monty's tournament underperformance can be attributed to chance alone; the alternative hypothesis is that "Monty the Traitor can't win big games" (or, statistically speaking, that we can disprove the null hypothesis to whatever alpha floats your boat).  "Can't win" does imply a future outcome (and "can't" refers to an indefinite time horizon), so inferential statistics are warranted.  Because if our NCAA performance really was due to something intrinsic to Mike Montgomery, then that would have been an argument for both parties moving on; otherwise, it might have been an argument for patience.

Ultimately, it doesn't matter; Monty left for Golden State, and eventually for Cal.  I am thankful for what he did while he was here, but I don't really want to resuscitate a long dead horse. I'm only interested in the statistical question (so it would be just as reasonable to substitute Mark Few for Monty), and really to answer the question as to whether you can really draw conclusions about a coach from tournament performance. And I give both a descriptive answer, "First, given our distribution of seeds, we slightly underperformed." and an inferential one, "the underperformance is not beyond what might have been observed by chance." (or, put more succinctly, "we underperformed, but I wouldn't read too much into Monty's talents as a coach based on the results")

BC

[BostonCard]

Well, it is not really a valid hypothesis because the data was cherry picked.  And nobody made a serious accusation of the NFC systematically cheating.  But, hypothetically, if systematic NFC cheating had been an issue, you could test the hypothesis that the NFC disproportionately won the coin toss (but again, you would need to use non-cherry picked data, which is the data on all the Super Bowls).  [As an aside, note the tendency to cherry pick data by some people here; at least one poster wanted to consider only the data after we made the final four].  On the other hand, people are suggesting, and I am quoting, that "Monty the Traitor can't win the big games in the tourney and it's time for both parties to move on".  That is a testable hypothesis; the null hypothesis is that Monty's tournament underperformance can be attributed to chance alone; the alternative hypothesis is that "Monty the Traitor can't win big games" (or, statistically speaking, that we can disprove the null hypothesis to whatever alpha floats your boat).  "Can't win" does imply a future outcome (and "can't" refers to an indefinite time horizon), so inferential statistics are warranted.  Because if our NCAA performance really was due to something intrinsic to Mike Montgomery, then that would have been an argument for both parties moving on; otherwise, it might have been an argument for patience.

Ultimately, it doesn't matter; Monty left for Golden State, and eventually for Cal.  I am thankful for what he did while he was here, but I don't really want to resuscitate a long dead horse. I'm only interested in the statistical question (so it would be just as reasonable to substitute Mark Few for Monty), and really to answer the question as to whether you can really draw conclusions about a coach from tournament performance. And I give both a descriptive answer, "First, given our distribution of seeds, we slightly underperformed." and an inferential one, "the underperformance is not beyond what might have been observed by chance." (or, put more succinctly, "we underperformed, but I wouldn't read too much into Monty's talents as a coach based on the results")

BC