Re: 2019 - 20 Rankings
I guess I had forgotten you had posted this. Thanks again for it (and the work it must have taken!)
I notice that USCHO likely has a bug somewhere in theirs right now.
If you go down the column of 'pairwise wins', they have the same [14,14,13,11,11,11,10] sequence of wins that you have. And that gives the right number of total comparisons (that is, it adds up to the correct number of comparisons made, same as if it had just been in normal order 15,14,13,12,11,10,9). But after that, USCHO has 7 and 7 for Union and Lindenwood; so, one fewer awarded pair win that normal sequence. But that missing pair win has not been awarded to anyone else anywhere else. So either they 'missed' awarding a pair win, or they have a comparison as a tie (!?), which I guess could happen if the RPIs were a dead heat. But of course, there are no dead heats as far as RPI goes in their table.
Originally posted by TonyTheTiger20
View Post
I notice that USCHO likely has a bug somewhere in theirs right now.
If you go down the column of 'pairwise wins', they have the same [14,14,13,11,11,11,10] sequence of wins that you have. And that gives the right number of total comparisons (that is, it adds up to the correct number of comparisons made, same as if it had just been in normal order 15,14,13,12,11,10,9). But after that, USCHO has 7 and 7 for Union and Lindenwood; so, one fewer awarded pair win that normal sequence. But that missing pair win has not been awarded to anyone else anywhere else. So either they 'missed' awarding a pair win, or they have a comparison as a tie (!?), which I guess could happen if the RPIs were a dead heat. But of course, there are no dead heats as far as RPI goes in their table.
Comment