|
Post by Brutus Buckeye on Nov 17, 2021 12:10:53 GMT -5
|
|
|
Post by karellen on Nov 17, 2021 12:20:41 GMT -5
Centenary is a DIII team and swept Grambling a D1. Interestingly, Centenary is at the bottom of their conference and not competitive with the top teams. This does not make the point you are trying to make. Comparing any team from any division to Grambling as an example of who a team from that division may do at DI level is not relevant.
|
|
|
Post by The Bofa on the Sofa on Nov 17, 2021 13:30:22 GMT -5
Centenary is a DIII team and swept Grambling a D1. Interestingly, Centenary is at the bottom of their conference and not competitive with the top teams. This does not make the point you are trying to make. Comparing any team from any division to Grambling as an example of who a team from that division may do at DI level is not relevant. Sure it's relevant. Because it is a comparison of a D2 and a D1 team. But yes, there is a problem. Grambling is a bottom D1 team. But then again, Centenary is not a great D3, either, so you have a weak D3 team against a weak D1 team. However, there is no reason to think that is any less informative than pitting a top D1 team against a top D2 team. If, for example, it was Minnesota against CSP, and Minnesota won, would you say that is not relevant because Minnesota is top team and so beats everyone? Of course not. A big challenge in comparing divisions is that the top D2 (and NAIA) teams don't play the top D1 teams, or even the mid-range D1 teams. It's generally low-budget D1 teams that play local non-D1 teams, so that's what we have for comparison. So far this year, I have 8 matches between D1 and NAIA, 4 matches between D1 and D2 and 2 matches between D1 and D3. For the NAIA opponents, 4 of the matches are with team(s) in about the middle of NAIA, 1 from the bottom 20 - 40%ile, and 3 in the bottom 20%ile. D2 opponents are a little higher, with 2 matches in the top 20-40%ile and 1 each from the 40-60%il and 60-80%ile. The D3 opponent (I think it was the same team) is 25-50%. So no top teams in D2 or D3, and not really above top half for NAIA. The D1 teams that are playing them are all in the bottom half D1, with most of them in the bottom 20%ile. Even then, they don't win much. I see three non-D1 wins against D1 teams. But those 2 NAIA wins are very helpful, and consistent. So we can put those teams, which are middle of the pack NAIA, on the D1 scale. So Prairie View A&M lost to Huston-Tillotson on the road, but beat tham at home. To a first approximation, we can say they are probably pretty close. PVAM is ranked 331 in Pablo, with a rating at 2770. HT is 154 in NAIA, rated 3935, which is about 5000 points lower than the top team in NAIA. Now, there is no reason that the scales in D1 and NAIA should be different (although emprically there does end up being some contraction and expansion, don't ask me why), so if we assume the scales are the same, that means that the top NAIA team would should up around 7700, which would put them at the top of d1. Obviously that is wrong, but the lesson is that those wins by non-D1 teams even at the bottom end do tell you something. PVAM isn't a good D1 team at all, no, but then again, they lost to HT, who is not a good NAIA team. When I do composite rankings, I look out for these extreme outliers like the top teams and base it more on general trends, so it doesn't come out so loopy. However, it's readily arguable that the top NAIA team is 3000 - 4000 Pablo points higher than TH, and if that is in D1 that put them around 75 or better, which is where I usually find them when I am able to do composite rankings. And D2 is historically better than NAIA (and this is pretty easily converted because there is a ton of data). I usually find the top D2 to be in that 30 - 40 range. Interestingly, others have tried doing the calculation directly, and when you do that, it's not uncommon to find the top D2 and NAIA teams in the top 10 overall. I get the same result if I try a ranking where I use all the teams and all the matches at once. I figure it's an issue of crazy outliers that are unreliable. So aside from getting crazy results, that calculation for me takes for-ev-er so it's not worth doing.
|
|
|
Post by vbnerd on Nov 17, 2021 14:06:09 GMT -5
Centenary is a DIII team and swept Grambling a D1. Interestingly, Centenary is at the bottom of their conference and not competitive with the top teams. This does not make the point you are trying to make. Comparing any team from any division to Grambling as an example of who a team from that division may do at DI level is not relevant. The assertion was made that such a match is not allowed. This at least refutes that they are allowed and in fact do still happen. Yes, most of these are in the SWAC or at HBCs but if their are matches between D1 and D2 and D3 and NAIA then there are at least SOME data points. We know where Centenary and Texas Lutheran and others fits in D3. We know where Rochester, Dillard and Tougalou and others fit in NAIA, we know where the D2 fit into this, and we know where the SWAC schools fit in to D1 etc. There are more connections between D2 and D3 and NAIA than D1, especially this year it seems, but those dots are there and can be connected. Would it be nice if we still had Juniata vs Princeton every year? Or if Hawaii played a couple of the other island schools? Sure, more data isn't a bad thing, but it's not like it is a total shot in the dark. Some people choose not to believe what the data suggests, and that's fine, some people don't believe the single-division Pablo data either. But there is data to guide the thinking on this.
|
|
|
Post by Brutus Buckeye on Nov 17, 2021 14:06:33 GMT -5
|
|