WBB: Michigan Intercollegiate Athletics Association

[Ryan Scott (Hoops Fan)]

New regional rankings are out.

In spite of the fact that Hope is 22-2, Ohio Northern (20-4) and Baldwin Wallace (20-4) are still both ranked above Hope. Both teams do have slightly better S.O.S. than Hope, so I accepted them being ranked higher last week. But both lost last week and now have a two-game worse record than Hope.

It's a pretty significant SOS difference.  .042 and .048 are big disparities - certainly enough to overcome a two win difference in most circumstances.  I don't think Hope's in much danger of missing the tournament, though.

[scottiedawg]

New regional rankings are out.

In spite of the fact that Hope is 22-2, Ohio Northern (20-4) and Baldwin Wallace (20-4) are still both ranked above Hope. Both teams do have slightly better S.O.S. than Hope, so I accepted them being ranked higher last week. But both lost last week and now have a two-game worse record than Hope.

It's a pretty significant SOS difference.  .042 and .048 are big disparities - certainly enough to overcome a two win difference in most circumstances.  I don't think Hope's in much danger of missing the tournament, though.

Loras over UW-Oshkosh, +109 points of WP beat out -85 points of SOS. (Loras has worse Common too)

Hope vs BW: + 71 points of WP, -44 points of SOS.  Similar with ONU.

I think ONU and BW over Hope largely has to do with best win vRRO. BW has a #1, ONU has a #2. Hope's best is #4.

[scottiedawg]

I know scheduling:
(a) takes two to tango
(b) requires some luck

but if Hope had anyone of: (Chicago, Millikin, Wartburg, Loras, Whitewater, Oshkosh) on their schedule instead of Finlandia, and beat them, not only is Hope #1 in Region 7, they are likely also Top 10 in the Top 16.

[pointlem]

I'd love to have someone explain the weighting system used by Massey's computer algorithm. Any prediction system is imperfect, with outcomes subject to random variations. Nevertheless, in real life and clinical judgments, statistical predictions, based on accumulated data, tend to surpass human prediction.

Massey's estimate of Hope having an 82% chance of beating Baldwin Wallace on a neutral floor, and an 86% chance of beating Ohio Northern, means the outcome would be uncertain . . . but do indicate something meaningful that is not captured by the DIII regional ranking system.

My observation is that Massey does predict outcomes with an accuracy that approximates its own confidence estimates . . . but some enterprising person could objectively assess its many predictions. Across teams and games, does a team that's said to have a 60% chance of winning actually win about 60% of the time . . . and likewise for 70%, 80%, and 90% predictions? For all I know, perhaps Massey has done that analysis, which is incorporated into its current predictions.

All that said, the only outcome that matters is what happens on the floor, and this season we've seen some weird results (albeit mindful that, if Massey is right, a team that it gives a 90% chance of winning should lose 10% of the time).

[TUAngola]

To me the best way to rank teams is thru the 'eye test'.  I have watched Hope, ONU and Baldwin Wallace several times this year, either in person (Hope and ONU) or on live stream (all three).  There is no way ONU or Baldwin Wallace is better than Hope when using the 'eye test'.  I get that the powers that be love their SOSs and RROs and what not, but swap out ONU to the MIAA and Hope to the OAC and Hope is still going to destroy most teams in that conference.   

[Dutchfan]

Does anyone know what the formula used to rank teams in the regional rankings is?

I know it is based on Win%, Win% over Regionally ranked opponents, and Strength of Schedule. But what is the exact formula used?

[HOPEful]

I know scheduling:
(a) takes two to tango
(b) requires some luck

but if Hope had anyone of: (Chicago, Millikin, Wartburg, Loras, Whitewater, Oshkosh) on their schedule instead of Finlandia, and beat them, not only is Hope #1 in Region 7, they are likely also Top 10 in the Top 16.

I agree with this. My biggest problem with the process actually isn't the process, but the SOS methodology itself. The formula needs changing. It too severely punishes teams for scheduling a Finlandia. There are many ways to tweak the formula to get a better representation of the best team and not just the best resume.

[Flying Dutch Fan]

Does anyone know what the formula used to rank teams in the regional rankings is?

I know it is based on Win%, Win% over Regionally ranked opponents, and Strength of Schedule. But what is the exact formula used?

No such formula exists.  Those criteria (along with some secondary things like head to head results) are what are used by each regional committee to rank the teams in the region and eventually by the national committee in selecting the Pool C teams and seeding the tournament. 

It comes down to how the committees interpret those items (like how much "weight" they put on SOS vs W/L).  Folks like scottiedawg and knightslappy attempt to mimick the committees with math and formulas (and are both really good at that), but you will see them constantly tweaking their formulas from year to year (or week to week) to try to align with the committees.

If there was an acutal formula used, we wouldn't need the committees at all (and we can argue whether that is a good thing or not)

[scottiedawg]

Does anyone know what the formula used to rank teams in the regional rankings is?

I know it is based on Win%, Win% over Regionally ranked opponents, and Strength of Schedule. But what is the exact formula used?

No such formula exists.  Those criteria (along with some secondary things like head to head results) are what are used by each regional committee to rank the teams in the region and eventually by the national committee in selecting the Pool C teams and seeding the tournament. 

It comes down to how the committees interpret those items (like how much "weight" they put on SOS vs W/L).  Folks like scottiedawg and knightslappy attempt to mimick the committees with math and formulas (and are both really good at that), but you will see them constantly tweaking their formulas from year to year (or week to week) to try to align with the committees.

If there was an acutal formula used, we wouldn't need the committees at all (and we can argue whether that is a good thing or not)

and with that post, FDF will be taking over all my PR opportunities from now on!

[sac]

[sac]

To me the best way to rank teams is thru the 'eye test'.  I have watched Hope, ONU and Baldwin Wallace several times this year, either in person (Hope and ONU) or on live stream (all three).  There is no way ONU or Baldwin Wallace is better than Hope when using the 'eye test'.  I get that the powers that be love their SOSs and RROs and what not, but swap out ONU to the MIAA and Hope to the OAC and Hope is still going to destroy most teams in that conference.

Long story short, the "eye test" was how it used to be done and this led to the "smokey back rooms" with conferences scratching your back if you'd scratch ours.  There were conferences that could never get a team in the tournament.  And it was even worse when there were only 48 teams in the tournament.

I like the ideas of our current criteria except the use and application of RRO's, its double counting and with regions being uneven in strength you get teams ranked in some regions counting as RRO that wouldn't be ranked in others.  Then using that same regional data to select a national tournament.  Square peg, round hole.

[scottiedawg]

To me the best way to rank teams is thru the 'eye test'.  I have watched Hope, ONU and Baldwin Wallace several times this year, either in person (Hope and ONU) or on live stream (all three).  There is no way ONU or Baldwin Wallace is better than Hope when using the 'eye test'.  I get that the powers that be love their SOSs and RROs and what not, but swap out ONU to the MIAA and Hope to the OAC and Hope is still going to destroy most teams in that conference.

I also think it's not really a problem of how you value SOS. The SOS formula is pretty misleading.

Currently WP + SOS does not separate these teams very much:
Hope - 917 / 525
St. John Fisher - 920 / 500

(Massey gives Hope 85% winprob. Massey thinks these teams are VERY different)

If Hope and ONU/BW had played a mostly overlapping schedule, then the SOS difference WOULD be quite meaningful.  (SOS spreads are much tighter than WP spreads)

Hope has played largely good teams or bad teams.  ONU/BW have played teams all along the spectrum. If Hope gets to play 8-17 teams instead of 3-22 teams their SOS jumps up and they still have the same WP (I bet).

SOS lumps it all together. I think the distribution of strength of opponents played is a much bigger factor in "strength of schedule" than lumping it all together. But there's no way to tease that out currently.

[HOPEful]

To me the best way to rank teams is thru the 'eye test'.  I have watched Hope, ONU and Baldwin Wallace several times this year, either in person (Hope and ONU) or on live stream (all three).  There is no way ONU or Baldwin Wallace is better than Hope when using the 'eye test'.  I get that the powers that be love their SOSs and RROs and what not, but swap out ONU to the MIAA and Hope to the OAC and Hope is still going to destroy most teams in that conference.

I also think it's not really a problem of how you value SOS. The SOS formula is pretty misleading.

Currently WP + SOS does not separate these teams very much:
Hope - 917 / 525
St. John Fisher - 920 / 500

(Massey gives Hope 85% winprob. Massey thinks these teams are VERY different)

If Hope and ONU/BW had played a mostly overlapping schedule, then the SOS difference WOULD be quite meaningful.  (SOS spreads are much tighter than WP spreads)

Hope has played largely good teams or bad teams.  ONU/BW have played teams all along the spectrum. If Hope gets to play 8-17 teams instead of 3-22 teams their SOS jumps up and they still have the same WP (I bet).

SOS lumps it all together. I think the distribution of strength of opponents played is a much bigger factor in "strength of schedule" than lumping it all together. But there's no way to tease that out currently.

100% agree. The problem isn't that SOS is used, but the current way it's calculated.

It's just incredibly harsh to punish Hope as hard as they are for scheduling Geneva, Finlandia, and Great Lakes Christian instead of better D3 teams. Wisconsin Eau Clare, Wisconsin Lutheran, Illinois Wesleyan, Wittenberg, Carnegie Mellon, and Otterbein is a very decent slew of a non-conference opponents. Their SOS is just anchored down by the very bad teams on their schedule.

[scottiedawg]

Minor point but the Great Lakes Christian game doesn't affect Hope's SOS since it's not a D3 game.

[Flying Dutch Fan]

Minor point but the Great Lakes Christian game doesn't affect Hope's SOS since it's not a D3 game.

True - although it does impact their W%