Top 25 discussion

[scottiedawg]

Thanks!

Do you know if the "Primary Criteria" are ordered, or merely enumerated?

[PauldingLightUP]

Thanks!

Do you know if the "Primary Criteria" are ordered, or merely enumerated?

Just up on page 16 it says "all criteria listed will be evaluated (not listed in priority order)"

[scottiedawg]

Probably best to keep the exact "weights" (whether conscious or unconscious) used between the primary criteria.

If a team knew those, they would know exactly how to schedule. Or could at least model it.

My hunch is that a sparkling record carries more weight than a strong SOS, unless the difference between win% is small, AND the difference between SOS is large. (e.g. a 27-2 team would quickly be slotted ahead of a 24-5 team, EVEN IF the SOS disparity was great). Of course, I have no way of knowing that for sure. I bet lots of you who have scrutinized seeding, hosting, and Pool C selections for years could she a lot of light.
BUT, if my hunch is true, it's best to schedule a bit easier, to give yourself the best chance at hosting and not needing to travel much.

Of course the flipside is getting your team ready and tested for top level competition, regardless of how it impacts your record, seeding, hosting, etc.

[VT-Alum-NOVA]

do members on selection committee use Massey ratings as data for their deliberations?  I have heard they request data and wonder if massey ratings used?

[Roundball999]

And Hope's SSF is 48 vs. 54 for Depauw.  Depauw has a 2 point squeaker at home vs no-vote Denison yet Hope's wins vs. Albion and Trine are brought up as negatives?  Not sure I see any discernible objective reason to say this year's Hope schedule is not as strong. 

It's an interesting math question to think about whether it's better to beat one team that was maybe favored and pick up a loss against a weak team vs. being favored and beating five teams that all had some reasonable chance to beat you.  SoS could turn out to be the same in those scenarios; which tells more about a team?  A question for better brains than mine.

Don't get me wrong, I don't see Hope as a juggernaut a la Thomas More.  Maybe next year since virtually everyone returns.   They don't have anyone that will be mentioned in All America discussions.  They are exceptionally balanced, go 15 or more deep every game and get 50% of their scoring from their bench, but sometimes you really do benefit from having that go-to star player at crunch time.  All I suggested was that the scenario mentioned - an undefeated Hope not hosting - would be an unfortunate effect of the NCAA's regional criteria.  Spoken of course as a Hope fan

[FCGrizzliesGrad]

In regards to Transylvania, while the HCAC isn't the strongest, they've won all 8 games by double digits as you'd expect. Their non-con includes wins over 15-2 Berea, 12-4 Wittenberg, 12-4 Redlands, 11-4 Monmouth with their lone loss against 13-4 Maryville. I think their non-con schedule was tougher this year than last when they went 27-3 and were a Sweet 16 team losing to the champs by 15 which is basically what everyone the last 2 weekends did against Thomas More.

[scottiedawg]

Absolutely.

This is where human minds have failings though. Those things can all be impressive, but that's not what we're debating. We're debating "impressiveness" relative to other teams. We're not even debating if a certain team is "good", "bad", or "other." Just if they're "better" than another team, and how many teams are they better/worse than.

I'm unable to mentally grasp that or calculate that at scale. But it's exactly what something like Massey is trying to do. (or ELO (soccer, football, others), DVOA (football), Pythagorean (anything with score differential), and a million more. Each has failings, but each is trying to calculate relative "impressiveness" in a way that humans without computers can't.

It's also hard to analyze Winning Percentage and Massey SOS, because those numbers themselves have limitations; they were simplified for humans. A single game is either a win or a loss. You don't get 97% of a win for playing amazing, but losing right at the end.
Similarly, The team with the #1 SOS may be waaaaaay harder than #2. Those "bucketizing" of statistics are ways of presenting data cleanly, for ease of human consumption.

A computer model gets down to all the nitty gritty, and spits out an order, based on that model.

It would be interesting to build a model according to the NCAA's selection criteria, even just taking Win% and SOS, baseline both, decide on weights between them, rank teams off that, and see how closely the selections end up being.

Really, this is close to what @knightslappy does on the men's side. http://tomaroonandgold.blogspot.com/p/division-iii-mens-regional-rankings.html

Specifically what I'd be interested in would be:
Winning Percentage, standard deviations above the mean winning percentage
SOS, standard deviations above/below the mean SOS

average them (which assumes equal 50% weight on each), and see what you get. You could mess around with the weights if you think the committee favors one.

[gordonmann]

And Hope's SSF is 48 vs. 54 for Depauw.  Depauw has a 2 point squeaker at home vs no-vote Denison yet Hope's wins vs. Albion and Trine are brought up as negatives?  Not sure I see any discernible objective reason to say this year's Hope schedule is not as strong.

I didn't say Hope's wins over Albion or Trine were a negative, and the comparison to DePauw's narrow win over Denision is a good point. The difference is that DePauw has beaten teams better than Denison and Hope hasn't played anyone better than Albion or Trine. So I have some evidence that DePauw's ceiling is higher than "two points better than Denison" this year, and I don't for Hope. I have to use something else for Hope, like "the eye test" or success in recent tournaments, which is more subjective or dated.

Look at DePauw's non-conference schedule and look at Hope's.

https://www.d3hoops.com/teams/DePauw/women/2019-20/index
https://www.d3hoops.com/teams/Hope/women/2019-20/index

DePauw's best non-conference wins are No. 6 UW-Whitewater, Illinois Wesleyan (in-and-out of Top 25) and either Wash U or Oshkosh. Both of those teams were Sweet 16 teams last year and have recently won their conferences. That's what DePauw signed up for when it scheduled those teams or went to events with those teams.

What are Hope's best non-conference wins by comparison?

Keep in mind that the NCAA has an SOS measure as a primary criteria, but it also uses "results versus regionally ranked teams" as a tie breaker for tough decisions. If Hope is 0-0 and DePauw is 3-0 or 3-1, maybe DePauw gets the advantage. I don't know for certain.

[gordonmann]

I'm very intrigued by this.  So would you assume that "27th most voting points" when voting for only the top 25 teams, would look different than "27th most voting points" when voting for say, the top 40 teams?

Here's a hypothetical.

Let's say 15 voters all think that Pacific is one of the 30 best teams in the country. The Boxers have beaten George Fox and Whitman who are on most people's ballots and are in first place in the very respectable NWC. But, when it comes time to take Pacific or, say, Emory or Gettysburg for the last slot on their ballot, most of the voters go with Emory (better losses) or Gettysburg (one loss).

In contrast maybe there are only a handful of people who are really high on No. 27 Texas-Dallas. They are from the same region as UTD, lost to UTD, want to order them ahead of someone they have high on their ballot like Mary Hardin-Baylor, etc. So UTD appears on fewer ballots, but they are ranked high enough on those ballots to finish 27th.

Note: I haven't examined whether this is actually the case. UTD happens to be sitting No. 27 right now.

The absence of opportunities to pick a 27th team doesn't hurt Texas-Dallas because they are only on the radar for the few people already voting for them. Everyone else has decided, "Nah, that's too many losses."

Extend the ballot to 30 teams and Pacific picks up all those near miss votes. It pushes them up the ballot and now they aren't "No. 32" anymore.

Maybe I'm thinking about this wrong. If so, I'm teachable.

[Baldini]

If DePauw and Hope hold serve, are they destined for a 2nd weekend matchup or is it more likely that they get broke up?

[gordonmann]

Good question. It depends who else is in the mix to host that second weekend and whether Hope or DePauw is 500 miles from them.

It's conceivable (but by no means even likely) that you have something like Bowdoin, UW-Whitewater, DePauw and Scranton as hosts and Hope gets sent to Whitewater. Or flip it, let Hope host, and send DePauw to Whitewater. But, in that scenario, DePauw says, "Wait, we already beat Whitewater head-to-head. We shouldn't have to go there."

This is why head-to-head results against NCAA tournament teams matter.

In a strange way, the team to watch here is Scranton.

In recent years, Scranton has been good enough to warrant hosting the second weekend and a lot of teams from the NE, East, Atlantic, Great Lakes and even South region can drive to them. If Scranton continues to struggle, they look less attractive as a host. Because of the concentration of teams in the northeast quadrant of the country (NE, East, Mid Atlantic and Atlantic regions), you sort of need two hosts in that part of the country. You could let two NESCAC teams host but that doesn't get you much for geographic reach since they are close to each other.  Or you could put more hosting sites west of Ohio and pay for more flights (probably not happening)

The committee needs Scranton to get hot or Baldwin Wallace to stay hot so they can build a quadrant around them.

[deiscanton]

When last we left Hoopsville, Dave McHugh (who does not vote in the women's poll-- full disclosure) wanted to have a conversation about Bethel and why the Royals are not in the top five.  BTW, the convo should also be held with the WBCA coaches, as the consensus top 7 in order are: Tufts, Bowdoin, Hope, DePauw, Amherst, UW-Whitewater, and Wartburg.  Both D3hoops.com and the WBCA have Bethel at #10, with some disagreement at who is #8 and #9.

I took some time this afternoon to watch the on demand version of Bethel's last home game vs St Olaf (don't get me started about teams using Sidearm Sports as a streaming platform, at least Bethel is not Rochester in regards of forcing a monthly subscription down your throat to watch archived games.)

Bethel is really good at stealing the ball and converting the turnovers into points.  FT shooting could use some work, as I know at least 5 UAA teams who are 70 percent or better from the charity stripe this season, and the Royals only shoot 65 percent from the line.

Still, 15-0 and 10-0 in MIAC play halfway through the season is pretty good.  Can Bethel get a perfect season and a national title?  Long way to go on that.

[Roundball999]

And Hope's SSF is 48 vs. 54 for Depauw.  Depauw has a 2 point squeaker at home vs no-vote Denison yet Hope's wins vs. Albion and Trine are brought up as negatives?  Not sure I see any discernible objective reason to say this year's Hope schedule is not as strong.

I didn't say Hope's wins over Albion or Trine were a negative, and the comparison to DePauw's narrow win over Denision is a good point. The difference is that DePauw has beaten teams better than Denison and Hope hasn't played anyone better than Albion or Trine. So I have some evidence that DePauw's ceiling is higher than "two points better than Denison" this year, and I don't for Hope. I have to use something else for Hope, like "the eye test" or success in recent tournaments, which is more subjective or dated.

Look at DePauw's non-conference schedule and look at Hope's.

https://www.d3hoops.com/teams/DePauw/women/2019-20/index
https://www.d3hoops.com/teams/Hope/women/2019-20/index

DePauw's best non-conference wins are No. 6 UW-Whitewater, Illinois Wesleyan (in-and-out of Top 25) and either Wash U or Oshkosh. Both of those teams were Sweet 16 teams last year and have recently won their conferences. That's what DePauw signed up for when it scheduled those teams or went to events with those teams.

What are Hope's best non-conference wins by comparison?

Keep in mind that the NCAA has an SOS measure as a primary criteria, but it also uses "results versus regionally ranked teams" as a tie breaker for tough decisions. If Hope is 0-0 and DePauw is 3-0 or 3-1, maybe DePauw gets the advantage. I don't know for certain.

It's a good discussion.

You seem to heavily weight nonconference; is the criteria strength of schedule or strength of nonconference schedule?  At least according to Massey, strength of schedule is nearly identical.  I do agree that Hope doesn't have a signature win of the quality of WW, best win is probably in and out of top 25 Trine.  But nor does it have a loss to a down year Wisconsin Lutheran.  Consistency or the ability to demonstrate that big win?  I don't know what's more important.  I may also be over reliant on Massey but one reason I do so is it doesn't let the past influence this year's view.  WashU and Oshkosh were sweet sixteen last year but this year they are barely above .500 and not the benchmark they were in past years.

It will sort out.  I still reserve the right to be outraged should Hope win out and not host

[Baldini]

Good question. It depends who else is in the mix to host that second weekend and whether Hope or DePauw is 500 miles from them.

It's conceivable (but by no means even likely) that you have something like Bowdoin, UW-Whitewater, DePauw and Scranton as hosts and Hope gets sent to Whitewater. Or flip it, let Hope host, and send DePauw to Whitewater. But, in that scenario, DePauw says, "Wait, we already beat Whitewater head-to-head. We shouldn't have to go there."

This is why head-to-head results against NCAA tournament teams matter.

In a strange way, the team to watch here is Scranton.

In recent years, Scranton has been good enough to warrant hosting the second weekend and a lot of teams from the NE, East, Atlantic, Great Lakes and even South region can drive to them. If Scranton continues to struggle, they look less attractive as a host. Because of the concentration of teams in the northeast quadrant of the country (NE, East, Mid Atlantic and Atlantic regions), you sort of need two hosts in that part of the country. You could let two NESCAC teams host but that doesn't get you much for geographic reach since they are close to each other.  Or you could put more hosting sites west of Ohio and pay for more flights (probably not happening)

The committee needs Scranton to get hot or Baldwin Wallace to stay hot so they can build a quadrant around them.

Any chance that Transylvania could be a team a quadrant could be build around?

[gordonmann]

It will sort out.  I still reserve the right to be outraged should Hope win out and not host

Sounds good.

By the way, I focus on non-conference SOS because teams can't really control who's in their conference (unless they are in the MIAC...ZING!) but the Committee focuses on total SOS. I do think the MIAA is a better conference overall than the NCAC. So that does even things out.

Any chance that Transylvania could be a team a quadrant could be build around?

I'm not aware of any reason why they couldn't. That first GL Regional Ranking is going to be exciting, eh?

When last we left Hoopsville, Dave McHugh (who does not vote in the women's poll-- full disclosure) wanted to have a conversation about Bethel and why the Royals are not in the top five.

Fair question. I think the answer is a mix of factors such as...

1) They started the year outside the Top 25 so they had more ground to make up. No, that's not fair, but where you start impacts where you are, especially on the women's side where teams lose so rarely. Hey, we should have them play their Top 25 mirror-twin Transylvania! Maybe in the NCAA Tournament?

2) Bethel has no history of deep NCAA tournament runs. The women's side of Division III is dominated by individual programs that usually have lengthy NCAA Tournament pedigrees. Multiple Final Fours, consistent Sweet 16/Elite 8s, etc. Same could be said for Loras except they have a win over another Top 10 team while Bethel has...

3) Their best win suggests they shouldn't be ranked too far from Augsburg. I watched that game and was impressed by both teams. I thought both could reach the Elite 8 or better. That game was really, really close and Bethel won on a four point swing late. if Bethel and Augsburg are almost equal to each other and Augsburg is No. 17, then Bethel gets tethered to them. That doesn't happen as easily with a team that has a history of winning in March -- no one is tethering Amherst to Emmanuel -- but it does for teams without that pedigree. As for as the abbreviated non-conference schedule Bethel has...

4) It's not a great non-conference schedule, though I know Massey says otherwise. It's fine, but there are no major signature wins on that schedule that make a voter say, "Hold on, maybe they should be higher since they beat this team who's higher on my ballot." That's the difference between Loras and Bethel. St. Thomas would normally be that team, but St. Thomas could be out of the poll entirely if they continue to go 0-for against the other MIAC title contenders (GAC, Aug and Bethel).

For what it's worth, I have Bethel No. 9.