|
|
Post by noreaster on Jun 12, 2024 12:38:05 GMT -5
I get the NCAA being sensitive to schools arguing they shouldn't have to spend money to go play teams from every other region to qualify for a Pool C bid (at-large bid). When they looked at different sports and see a Mary Hardin-Baylor going 27-2 with one weekend in a hotel and getting in over Williams who was under .500 in their own conference, I think they probably took that as a sign that this DOES work instead of proving the opposite. Thing is, the data they had was applying the new rules to schedules that were designed to maximize opportunity under the old rules. Looking at the coaches on the committee, MIT has Calvin coming in and is going to play fellow committee member Santa Cruz at NYU, Berry is going to Minnesota to play Northwestern and Whitewater, Haverford is going to play a double header vs Christopher Newport - as I understand the new system, none of these look like great decisions and these are the coaches that knew what was happening. If anybody has an NPI friendly/smart schedule this year, it's probably an accident. So I think they'll point to the first year and be like "see, it works" and then as coaches get smarter about what is in the math and how to exploit it, this will break down more and more until we try something else. Also, it's worth mentioning, I think Wash U was a bubble team in '16 with 10 losses who got hot and made it to the final match against Calvin who was a (1 loss) Pool C that year. But since then, is Trinity '22 the only at-large team to make a final since 2016? And I think they were a pretty safe selection that year. So whatever they do with the selections it shouldn't have too great of an impact on the naming of a champion aspect of the national championship tournament. There is something to be said that while these decisions by MIT, Berry, Haverford create risk on getting selected, the coaches may feel a strong schedule is required to be ready to get wins in the tournament if they win in/get an at large. Schedules are done in October and November. Aside from the doubleheader nothing looks atypical for what you would have seen under the old system. If they had thought about it, I assume they bought into the line that this will only change maybe one team each year and to keep doing what you have been doing. And honestly, many coaches didn't really understand the old system, even after so many years, so it's entirely possible that most schedules won't change right away. |
|
|
|
Post by coahc21 on Jun 14, 2024 8:47:57 GMT -5
|
|
|
|
Post by bumpsetspatch on Jun 14, 2024 11:05:46 GMT -5
UWO opened two new satellite campuses in neighboring cities five years ago and due to the overall enrollment struggles that are hitting the Wisconsin state system the Fox Cities and Fond du Lac locations are both closed. Enrollment is down at the main campus too, but there does not seem to be an end in sight when it comes to UW-Oshkosh closing. |
|
|
|
Post by d3follower on Jun 15, 2024 7:29:04 GMT -5
UWO opened two new satellite campuses in neighboring cities five years ago and due to the overall enrollment struggles that are hitting the Wisconsin state system the Fox Cities and Fond du Lac locations are both closed. Enrollment is down at the main campus too, but there does not seem to be an end in sight when it comes to UW-Oshkosh closing. To be clear, the two-year colleges within the Wisconsin System were merely restructured to be under control of the nearest four-year university within the system. Oshkosh just happened to be nearest to two of them. The Fox Cities campus has served Northeast Wisconsin since 1933. There were 13 colleges and 13 universities when the restructuring occurred in 2018 with plenty of controversy. There will be seven colleges after the closing of four more. Five of the remaining seven fall under the umbrella of DIII universities. |
|
|
|
Post by noreaster on Jun 22, 2024 0:37:56 GMT -5
NCAA announces the webinar to explain the NCAA Power Index will be Friday afternoon, during the semi-finals and finals of 17s at AAUs.
|
|
|
|
Post by noreaster on Jun 26, 2024 9:17:26 GMT -5
Eastern Nazarene in Boston is closing immediately. whdh.com/news/eastern-nazarene-college-in-quincy-announces-closure/Looks like Gordon will get a bunch of players from this, and a bunch of schools are losing a match on their schedule. William Smith sent out an e-mail blast looking for someone to replace them on September 13-14. They were also headed to the opening weekend event at Cedar Point. |
|
|
|
Post by ned3vball on Jun 28, 2024 16:29:14 GMT -5
The webinar on NPI today was not very informative. They repeated many statements from the earlier YouTube video that do not really explain the calculations involved. They again did not provide the correct "NPI" definition of win% or SOS. They did not mention the "game" NPI formula at all. The main fact to come out was that they are counting ALL losses, even those that improve your NPI. The "standard" formula does not use losses that help you. So feel free to add Juniata to your schedule if they are looking for another match to fill things out.
They did say they will not be providing NPI data till about the 3rd week of October, mimicking the old ranking schedule. Nor did they say what data they will be providing. I take that to mean they will only be producing a rank list by NPI numbers with no detail data. Myself and others will be doing it from day 1.
|
|
|
|
Post by noreaster on Jun 29, 2024 6:41:34 GMT -5
Did anyone else on the NPI Webber get the feeling that they were answering from talking points/the PowerPoint and not an actual understanding of what is going to happen?
And they said we aren’t weighting home and away because with 5600 matches they were not confident that information would be accurate, did anyone else think that was kind of damning of the entire venture and not just the home/away factor?
|
|
|
|
Post by salsolomon on Jun 29, 2024 14:36:10 GMT -5
I felt the webinar was more an explanation of how they took the formula/list of criteria from the NCAA and then applied that to volleyball.
It's like they felt their audience was NCAA administrators and they were answering questions like "How did you decide to weight the home/away factor?"
In reality the audience was coaches and fans who wanted to know not how did they set up the formula but "how will this impact my team?" or "How should this impact how my team schedules in the future?"
One questioner asked them how prior years would have come out under this new criteria and they purported to answer that question but in fact talked right past it.
|
|
|
|
Post by miaafan on Jun 29, 2024 18:33:46 GMT -5
The webinar on NPI today was not very informative. They repeated many statements from the earlier YouTube video that do not really explain the calculations involved. They again did not provide the correct "NPI" definition of win% or SOS. They did not mention the "game" NPI formula at all. The main fact to come out was that they are counting ALL losses, even those that improve your NPI. The "standard" formula does not use losses that help you. So feel free to add Juniata to your schedule if they are looking for another match to fill things out. They did say they will not be providing NPI data till about the 3rd week of October, mimicking the old ranking schedule. Nor did they say what data they will be providing. I take that to mean they will only be producing a rank list by NPI numbers with no detail data. Myself and others will be doing it from day 1. It's not clear to me why you think that they are not using the "correct" definition of winning percentage. Is it because they are dropping games? If a 30-2 team is 12-2 after "bad wins" are dropped, their winning percentage should be 12/14, right? You wouldn't want to use their unadjusted percentage if you are dropping games. |
|
|
|
Post by ned3vball on Jun 30, 2024 8:51:42 GMT -5
It's not clear to me why you think that they are not using the "correct" definition of winning percentage. Is it because they are dropping games? If a 30-2 team is 12-2 after "bad wins" are dropped, their winning percentage should be 12/14, right? You wouldn't want to use their unadjusted percentage if you are dropping games. The commonly understood definition of 'Win %' = wins / total matches (be it the whole season or a subset) is not used in the NPI calculation. Likewise, 'SOS'(strength of schedule) = 2/3 OWP + 1/3 OOWP, is also not part of the NPI calculation. For some reason they did not want to say this. My theory is they want a comfort level that things are not really changing that much, I get it. If I type the individual game NPI formula here it would be helpful to see what those 'dials' mean in practice. But for now I am not going to do that as I don't want the NCAA to get mad at me(not sure why they would or should but you never know). Is the 'win %' dial related to win %? Yes, but in a different way. Is the 'SOS' dial related to strength of schedule? yes, but in a different way. So they just decided to not sweat the details. |
|
|
|
Post by ned3vball on Jun 30, 2024 9:19:43 GMT -5
Did anyone else on the NPI Webber get the feeling that they were answering from talking points/the PowerPoint and not an actual understanding of what is going to happen? And they said we aren’t weighting home and away because with 5600 matches they were not confident that information would be accurate, did anyone else think that was kind of damning of the entire venture and not just the home/away factor? Yes and yes. The good news might be that the data needed for this system is much simpler than the data sheets they needed to prepare for the old criteria(which always had multiple mistakes with 2023 being a glaring example, but that is water under the bridge). As they are punting on home/away/neutral, all they need is a spreadsheet with 3 columns: team A, team B, who won. Hopefully they will include H/A/N and date(lots of schools play twice) to help vet the schedules. They should make this list public as part of the data. That way people can audit and point out inconsistencies. |
|
|
|
Post by miaafan on Jun 30, 2024 9:49:00 GMT -5
It's not clear to me why you think that they are not using the "correct" definition of winning percentage. Is it because they are dropping games? If a 30-2 team is 12-2 after "bad wins" are dropped, their winning percentage should be 12/14, right? You wouldn't want to use their unadjusted percentage if you are dropping games. The commonly understood definition of 'Win %' = wins / total matches (be it the whole season or a subset) is not used in the NPI calculation. Likewise, 'SOS'(strength of schedule) = 2/3 OWP + 1/3 OOWP, is also not part of the NPI calculation. For some reason they did not want to say this. My theory is they want a comfort level that things are not really changing that much, I get it. If I type the individual game NPI formula here it would be helpful to see what those 'dials' mean in practice. But for now I am not going to do that as I don't want the NCAA to get mad at me(not sure why they would or should but you never know). Is the 'win %' dial related to win %? Yes, but in a different way. Is the 'SOS' dial related to strength of schedule? yes, but in a different way. So they just decided to not sweat the details. |
|
|
|
Post by miaafan on Jun 30, 2024 9:57:36 GMT -5
I still don't understand your claim. They are exactly using the winning percentage of the games that count. The NPI is .2*PCT + .8*oppNPI + QWB (where all three of these terms are averaged over games that count, that is excluding bad wins). The winning percentage of the games that count is definitely what the NCAA means by percentage and they definitely use it. This fact may be a little obscured by the fact that the only reasonable way to compute NPI is iteratively so that the .2 is applied before the averaging (wins are counted as .2 and losses as 0 instead of 1 and 0) but in the end .2 * PCT is what is used.
|
|
|
|
Post by ned3vball on Jun 30, 2024 12:15:23 GMT -5
I still don't understand your claim. They are exactly using the winning percentage of the games that count. The NPI is .2*PCT + .8*oppNPI + QWB (where all three of these terms are averaged over games that count, that is excluding bad wins). The winning percentage of the games that count is definitely what the NCAA means by percentage and they definitely use it. This fact may be a little obscured by the fact that the only reasonable way to compute NPI is iteratively so that the .2 is applied before the averaging (wins are counted as .2 and losses as 0 instead of 1 and 0) but in the end .2 * PCT is what is used. OK. You saw my PM. So the game NPI I used is different. I will do a run with your formula. Interesting that my method mirrored the 2021 results(from the YouTube video) very well. Note that was a version pre minimum 10 win rule and include all losses. |
|
