Pool C

[Greek Tragedy]

So conference tournament games are considered conference games in terms of being weighted?

[KnightSlappy]

So conference tournament games are considered conference games in terms of being weighted?

Yes, the will be weighted 0.9 / 1.0 / 1.1

[deiscanton]

Today's official NCAA DIII NPI Men's Basketball Summary Report which includes games played through Tuesday, Feb. 18, 2025 can be found at

http://stats.ncaa.org/selection_rankings/nitty_gritties/42978

You will notice that there are some discrepancies today between the D3Datacast NPI numbers and the official NPI numbers.

This is because not all of the D3 men's basketball games from last night made in into the Massey score system for today's runs, so the D3Datacast computer has to catch up to the information that the NCAA statistics computer has.

Among games of note that were entered yesterday into the NCAA statistics system, but not into Massey's system were 2 ODAC games-- Randolph-Macon's win over Virginia Wesleyan and Roanoke's win over Ferrum from last night are included in the NCAA summary report, but the data from those games has not made it yet into the D3Datacast NPI runs, for example. 

Update-- Even in the most recent D3Datacast NPI run where games from last night were entered into manually into the D3Datacast computer, there are still discrepancies in the numbers.   For example, the D3Datacast computer in the 10:50 AM ET this morning has the UChicago men with an NPI of 61.256 (#33 in DIII), while the NCAA has the UChicago men's NPI at 61.287 (NPI rank #33) NYU men's NPI is ranked #2, but D3Datacast has the NPI at 73.230, and the NCAA has the NPI at 73.240, for example.

Not sure if there were some coding or entry errors last night or early today that skewed the NPI numbers, and if so, on what end.

[scottiedawg]

The nice thing about the NPI data that d3datacast and I are calculating is that once you are able to exactly match the official NCAA numbers, you know your code/logic is all correct.

So any discrepancies HAVE to be data discrepancies, which makes finding the discrepancy a lot easier.

[deiscanton]

There are still major discrepancies in the 3 PM ET D3Datacast run and the official NPI numbers on the men's end.

For example, the NPI number for UChicago men in the D3Datacast 3 PM ET run is 61.257, while the NCAA has it at 61.287.

Today's men's basketball games will start playing within a few hours, so I will have to recheck on Thursday as to the progress.  I'm sure the D3Datacast people will keep working on it in the meantime to see what the errors are.

[KnightSlappy]

The D3Datacast people will probably wait until tomorrow morning at this point.

[bopol]

Has it ever been them that has an error and not you?  That could happen...

[deiscanton]

I am not too worried about it right now.   The NPI rankings of the teams are the same in today's D3Datacast runs and the official NCAA DIII NPI summary report.  In other words, the top 16 teams in terms of NPI ranking for men's basketball are still in the exact same order, and NPI national ranking #44 belongs to UW Stevens-Point in both the unofficial D3Datacast NPI runs conducted today and in today's summary report. 

In other words, the discrepancies are not large enough to change the outcome of the selections and/or seeds based on inaccurate data yet.  I just want to see the parallel systems reconciled sometime this week.

BTW, this only affects the DIII men's basketball side of the picture.   I don't see any problems with the women's basketball side.   

[deiscanton]

The official NCAA DIII Men's Basketball NPI Summary Report through games of Wednesday, Feburary 19, 2025 has been published.  It can be found at:

http://stats.ncaa.org/selection_rankings/nitty_gritties/42992

It looks like the scores in the Massey scoring system have caught up to the scores reported in the NCAA statistics system, as there are no major discrepancies today.  The differences in NPI numbers, if there are any, are no greater than .001 or .002 of a point.

Of note-- Officially, UChicago's NPI number today is 61.308 according to the NCAA report (#31 in the NPI rankings) and Franklin and Marshall's NCAA official NPI number today is published at 61.307 (#32 in the NPI rankings).

D3Datacast's morning run as of Feb. 20, 2025 at 6:39 AM ET unofficially had Franklin and Marshall's NPI and UChicago's NPI numbers tied at 61.309 with Franklin and Marshall getting the #31 NPI ranking and UChicago getting the #32 NPI ranking on the D3Datacast unofficial run.

However, the official ranking today has UChicago at #31 NPI ranking and Franklin and Marshall at #32 NPI ranking this morning. 

[KnightSlappy]

It looks like the UEC Tournament games from last night are all marked as non-conference in the NCAA NPI numbers. I have reached out to someone regarding this.

[ziggy]

Has it ever been them that has an error and not you?  That could happen...

Yes. There were a few games we caught in the "official" data that were incorrect with their conference/non-conference designations due to teams in the same conference playing a non-conference game. Those have been resolved on the NCAA side, though we are now seeing that happen against with UEC tournament games as Matt noted above. Our understanding is that we are correct to handle these as conference games, though the "official" NPI run through yesterday's results released by the NCAA today shows them as non-conference contests.

[KnightSlappy]

It looks like the UEC Tournament games from last night are all marked as non-conference in the NCAA NPI numbers. I have reached out to someone regarding this.

It sounds like this will be corrected (hopefully tomorrow).

[deiscanton]

It looks like the UEC Tournament games from last night are all marked as non-conference in the NCAA NPI numbers. I have reached out to someone regarding this.

It sounds like this will be corrected (hopefully tomorrow).

That probably explains the difference in NPI rankings with Franklin and Marshall and UChicago on the unofficial D3Datacast run this morning and today's NCAA DIII Men's Basketball NPI summary report.   Franklin and Marshall played United East team Lancaster Bible in non-conference play this season, so the coding of the UEC tournament games as non-conference games would skew the NPI results.  Franklin and Marshall and UChicago are really close in the NPI table this morning--enough that a NPI skew of .001 or.002 this morning would affect the NPI ranking placement for those 2 teams.

I will take a look at tomorrow's NPI summary report to see if there are any corrections.

[Ryan Scott (Hoops Fan)]

Are they conference games?  I'd always been told the NCAA did not consider conference tournament games to be conference games.

[KnightSlappy]

The intention with the multiplier is to count conference tournament games as conference games.