Well... Week 10:

http://www.avca.org/divisions/division-three-women/poll-11-10-15/

I'm just wondering: What justification does AVCA use for "fudging" their rankings? What I mean is, only 37 of their 63 voters actually turned in a ballot this week. They ranked the teams based on those ballots, and added 26 "ghost" ballots filled in according to that ranking. The actual poll, without the ghost ballots, was:

1   Washington-St. Louis (24)   905
2   Emory (1)               861
3   Juniata (9)            860
4   Calvin (3)               827
5   Christopher Newport      765
6   Wittenberg            757
7   Claremont-Mudd-Scripps      652
8   Colorado College         627
9   Cal Lutheran            604
10   Hope               584
11   Wisconsin-Whitewater      580
12   Southwestern (Texas)      502
13   Carthage               462
14   Texas-Dallas            427
15   Wisconsin-Stevens Point      390
16   Eastern               351
17   Elmhurst               291
18   Clarkson               267
19   Wisconsin-Oshkosh         140
20   La Verne               135
21   Gustavus Adolphus         123
22   Chicago               120
23   Illinois Wesleyan         106
24   MIT                  82
25   Carnegie-Mellon         76

The actual poll was a lot closer than what they published. For example, Emory garnered only 1 more point than Juniata, not the 27 that the published poll shows. And Carnegie-Mellon barely edged out Hendrix (72) for the last spot.

Quote from: 7express on February 27, 2015, 10:37:05 PM
I think it was Hoopsfan that mentioned this on another board but you should start from the bottom and work your way up for the tiebreaker.  if 2 teams tie for first with 12-2 records (1-1) vs. each other while team A lost the 7th place team and team B lost to the 3rd place team you should reward team B with the championship for losing to the better team and not the other way around.

Except - the first tie-breaker should be head-to-head, which implies you should start from the top. My point was that anything past this is arbitrary since, by virtue of the tie, any such set merely trades wins in what is included, for loses in what is excluded.

Quote from: WUH on February 22, 2015, 07:14:38 PM
I thought about it for a minute and then gave up because I am unsure what happens if we have a three way tie at 7-7 if Rochester defeats Emory, Case Western Reserve defeats Carnegie Mellon and Brandeis defeats NYU.

Then we have a potential two way tie at 8-6 if Washington loses to Chicago and NYU defeats Brandeis.

Do we simply look at all the head-to-heads or do we move on to the next team in line?  Or maybe I am missing something.

If WashU beats Chicago: Emory and WashU will be tied in the "top 2," and they split. So the first tiebreaker, head-to-head, doesn't do anything.

If NYU also wins: Emory, WashU, Chicago and NYU will constitute the "top 4." WashU is 3-3 (split all around) and Emory is 2-4 (lost twice to NYU).

If NYU loses, the second tiebreaker adds in at least Rochester, requiring at least the top 5. WashU's advantage increases to 2 games since they won 2 and Emory (would have) split. If CWRU wins, we need the top 6. Emory won twice and Wash U split, offsetting the increase from Emory. So it is the Emory-NYU record still establishes it.

+++++

If Chicago wins: they get +1 for two wins over WashU, and Emory gets -1 if NYU adds into the top 4.

Quote from: WUH on February 22, 2015, 04:18:24 PM
Here are the latest standings with just one game to go.  Emory has clinched a share of the championship and will host Rochester with a chance to win it all.

Emory University 9-4
Washington University 8-5
University of Chicago 8-5
New York University 7-6    
Case Western Reserve 6-7    
University of Rochester 6-7    
Carnegie Mellon 4-9
Brandeis University 4-9

If I understand the tiebreakers correctly, Emory still has to win against Rochester to receive the Pool A bid. If they lose to Rochester, they will be co-champions with either WashU or Chicago. Emory loses the tiebreaker with WashU because Emory lost twice to NYU, and with Chicago because Chicago would have beaten WashU twice.

While these tie-breakers sound validated because they require beating the "better teams" more, they are equally invalidated because they require losing to the "worse teams" more.

Quote from: Hoops Fan on February 18, 2015, 06:45:17 PM

Quote from: WUH on February 18, 2015, 04:31:00 PM
I asked every student I talked to today and though my sample is relatively small and included only one athlete who is not on the basketball team, no one I talked to knew Pi past 3.1415.

I think we would fare better on GPA.

Wait, so I know one more place of pi than the average WUSTL student?  That makes me feel a lot better about myself today.

I can't resist pulling up a 40 year old memorization: 3.1415926535897932384626433.

Quote from: WUH on December 02, 2014, 11:27:37 AMI need to double check the years, but at one time, it would have been possible for a WUSTL basketball to live closer to the Fontbonne gym than any other building on the WUSTL campus.

Quote from: WUH on December 02, 2014, 04:56:41 PM

Quote from: jeffjo on December 02, 2014, 03:14:48 PM
You don't need the rental agreement for that. From Google Earth, a center-to-center measurement of Wheeler dorm (SW corner of the South 40) to:

I was referring to the fact that WUSTL students used to live on the Fontbonne campus in a dorm they rented from then Fontbonne College.

I really don't mean to sound testy, but you were referring to the possibility that "a WUSTL [student athlete could] live closer to the Fontbonne gym than any other building on the WUSTL campus." You used the possibility of these rentals as evidence, not the subject. And you don't need that evidence - it was possible entirely within the South 40.

But yes, I do know several WashU students who lived at Fontbonne. I sometimes ate dinner at Wohl Center with them, so they were under the umbrella of the South 40 but lived across the street. But they were Law students. Other kinds of students may have been possible in other years, but these are the ones I know of. And it possible, but unlikely, that they could have been on a team. But not basketball, as we had no basketball team when I attended.

Quote from: WUH on December 02, 2014, 11:27:37 AMI need to double check the years, but at one time, it would have been possible for a WUSTL basketball to live closer to the Fontbonne gym than any other building on the WUSTL campus.

You don't need the rental agreement for that. From Google Earth, a center-to-center measurement of Wheeler dorm (SW corner of the South 40) to:

• Fontbonne's Dunham Student Activity Center is about 250 yards.
• WashU's Simon Hall (closest on campus) is about 650 yards.
• WashU's Field House is about 675 yards.

For my first dorm, Rubelmann in the center of the South 40, these numbers are 390, 450, and 550. Still closer to Dunham. I believe Gregg, a new dorm, is now the closest to campus. Its numbers are 570, 265, and 450; but about half of campus is further than 570.

Quote from: WUH on December 01, 2014, 01:37:17 PMThe craziest thing about Fontbonne playing Washington University in California is not that they are separated by a two-lane road, but that Fontbonne traveled all the way to California for just the one game.  At least Washington University played two games.

Incidentally, the NCAA mileage calculator states that the institutions are 8 miles apart when the gyms are probably a 8 minute walk away from each other.  Maybe 1/4 miles.

It's about 1/2 mile - but that's not the most interesting thing. (A) The gym is one of the closest WashU campus buildings to the WashU dorms, but (B) almost all of them are closer to Fontbonne's gym, in the center of their campus!

Did we ever get a reason for the California trip? Did they just want a beach for Thanksgiving?

Quote from: AmherstStudent05 on March 07, 2014, 04:24:24 PMI respect that there is a felt need to have a computer provide us the comfort of an "objective" ranking system, but I am genuinely curious: Is there any actual evidence of Massey's wonderful predictive powers?  I say this because I recall a post from Pat at the end of last year's Tournament.  He had created separate brackets for the 2013 bracket challenge: one where predictions were based on Massey rankings and the other where the d3hoops poll dictated.  The d3hoops bracket performed incredibly well.  Massey......not so much.  Obviously only one year and one tournament, but still interesting I think.

Admitting that I have no knowledge of this comparison, I have to jump in. Massey predicts all 61 games, including all the nebulous ones between unranked teams. D3T25 predicts about half of them, but most of those are the cherry-picked ones of ranked vs. unranked.

Did Pat compare just the same set of games? A comparison isn't valid otherwise, and your description (bracket vs. bracket) sounds like he didn't. Regardless, one year isn't a valid sample space. I've done comparisons in the past, with computer programs that weren't as good as Massey, and they do marginally better than polls. But both are based on regular-season results, and the NCAA tourney is a different animal, so I don't like the comparison.

But let me introduce one completely anecdotal (and completely inadequate for a comparison) piece of evidence:

Quote from: D3Hoops headlineYork sprints to upset

Massey power ranking for York: 43 (12th percentile)
Massey power ranking for Rhode Island College: 108 (25th percentile)

How is this an upset, except by the expectations in the most populous region?

Quote from: Mr. Ypsi on March 03, 2014, 09:13:19 PMThere will hopefully NEVER be a Bracket of Death to compare to 4-5 years ago - if memory serves, ONE quadrant had 7 of the top 8 teams in the d3hoops poll!!

Round #1 in 2009: 2 v. 20, 8 v. 19, and 21 v. RV; then 1, 3, 4, and 5 playing unranked teams. That's 7 round 1 games involving 9 top-21 teams; #7 Puget Sound got a bye, and then #21. So round #2 also had 2 v. 8, 3 v. 5, and 1 v. 4. Sounds like round 4 to me.

Using Massey final ratings (all I saw in their archive), it was 1 thru 11 missing only #10. That was Augustana, who didn't make it in.

Quote from: sethteater on February 26, 2014, 11:56:51 AM
Hope's losses are to...
#1 and #2 team in the West (#1 and #2 in the nation according to D3Hoops.com)
the #2 and #3 team in the Midwest (#4 and unranked according to D3hoops.com)
the #5 team in the NAIA Div. 2
and one, ugly in conference loss on the road...

Agreed: These loses (excluding the last one) can't be held against them. But they can't help them, either.

On any given day, a ranked 11-25 team can beat a top 10 team. Just not that often. One such loss counts as nothing. Three such losses are questionable. But given five opportunities, and no successes, means the team can't be put in an equivalent category.

Step up at least once, or step back. Kudos for obtaining such a tough schedule, but it means nothing if you can't do something with it.

Quote from: WUH on February 23, 2014, 02:45:17 PM
Whew!

Whew, indeed. Wash U started 10-2, but fell behind 32-21 in the first half. Battled back to tie in the second, but fell behind by 8 with 7 minutes left.

Play sequence of the game: Down 79-78 with a minute left, and the shot clock winding down. Alan Aboona drives along the baseline to take the lead AND foul out Christian Manoli (14 TP, 9 RB, 8 A, all team highs). Misses the and-one, but Nick Burt rebounds. Aboona is fouled again at 0:30, and makes both. Up 82-79, completing the comeback.

The sign of a good team is not always blowing out opponents, but being able to maintain composure and win when you don't start your best.

Quote from: Greek Tragedy on February 05, 2014, 08:10:41 AM
If my math is correct, the other 24 voters that didn't vote Cabrini #1 voted them #2.

You were close. To hit 600 points exactly, the votes have to average 24. Since one counted for 25 points, one must count for 23 to balance the average. So one #1, twenty-three #2's, and one #3.

I once wrote a program to guestimate the entire set of votes. It's not perfect (there always seem to be some anomalies - you'll see some below), but I think it is informative. Starting with #1, it says:

UW-Stevens Point, 24 1
Cabrini, 1, 23, 1
WPI, 0, 1, 18, 3, 0, 1, 1, (and a #13)
Washington U., 0, 0, 6, 8, 3, 2, 1, 2, 0, 1, 1, (and a #16)
Illinois Wesleyan, 0, 0, 0, 7, 4, 3, 3, 2, 2, 1, 1, 1, (and a #17)
Wesley, 0, 0, 0, 4, 5, 4, 4, 2, 1, 2, 1, 1, (and a #18)
UW-Whitewater, 0, 0, 0, 3, 6, 4, 3, 2, 2, 2, 1, 1, (and a #19)

These are spread out more than usual because somehow there are 31 points missing (all the totals add up to 8094, and it should be 8125). I'm guessing they belong somewhere in 3 to 5, since the differences in totals there are too big and that's where the result starts to look unusual. If I'm right, it can't change the rankings, but it will make somebody look closer to the top.

Quote from: WUH on January 31, 2014, 08:43:47 PM
I chose not to pay the $6.95 fee for the video feed, though it seems as though this game may have been worth it.  I would love a full report from someone who watched.

As did I. I refused to pay it more because I didn't want to let them justify it, than its affordability.

Wow. I didn't expect such a wide and varied response. I had intended to respond to whatever answers I got this morning before work, but there were too many.

Some background on me, so I can't be accused of hiding it. I graduated from Wash U with a Masters in Applied Math, summa cum laude, after only four years. But I attended during the interregnum, so I never was much of a college hoops fan until my kids were big enough to attend major Div I schools. I started following Wash U the same time I started following college hoops in general, at the end of the heyday (not that the last 6 years are anything to be ashamed of) of the women's program. I have read D3HOOPS ever since, but never posted here (I did think about it two years ago, when I registered). I even attended the "cruise" to the championship in Salem, 2009; much buoyed by da Bears (sorry; Chicago roots, too) surviving a true "bracket of death." I may even go again this year (fingers crossed for luck) for the quarters/semis.

My point is that I am not a newbie in any sense of the word. And I resent any implication that I was, and any inaccurate claims that I was wrong in any of the rankings (yes, I know what the different kinds are, thank you) I cited. I thought I indicated, pretty clearly, if I was talking about an 8-team or 16-team "bracket." And even though I left out the bottom half, I was accurate for what I was describing.

In fact, it was the differences in the kinds of rankings that I was trying to emphasize. Massey is an objective assessment of performance. Whether or not you agree that it succeeds, it is trying to rate each team's performance against the performance-strength of its opponents. Regional rankings - which I did not even consider accessing, since they aren't public AFAIK - are an objective assessment of merit, not performance. The difference is that merit's baseline is what is possible for a team given their schedule (or what they potentially could schedule). It's not an absolute measure.

I fully understand the difficulty the committee faces in creating this ranking, and in organizing a national tournament without the existence of anything better to base it on. I applaud their results. But each "regional" bracket is only three tiers deep; I just don't see much point in bemoaning whether the very-necessary constraints on scheduling make matchups occur one round too soon, which realistically is all that happens. I just don't place much stock in the numbers when comparing the capability of teams from different regions, so the national view is what I wanted to look at.

The D3HOOPS *poll*, on the other hand, is a subjective assessment. Whether it measures performance or merit is in the eyes of the voter, but it also is a trans-region measurement. So when I think about a "bracket of death," I think about imbalances on the national scale, not the regional scale. Both the D3HOOPS *poll* and the Massey ratings can indicate this imbalance, by comparing the actual seedings (whether or not they are based on a ranking statistic) to what they would suggest, if they had been used.

But none of that is really the reason I posted. At Wash U, we have come to expect to be in the most death-like region, and 2009 was not the only example [see note] of an extreme. My impression, from DeWitt's comments, was that the biggest concern the committee had about this issue was St. Thomas's bracket (whether 8 or 16). But the same impression I got from D3HOOPS interviewers was that it was Whitewater's (yes, I think of it as Whitewater's). And I agree with my impression of D3HOOPS. But it wasn't clear, so I asked.

[Note] Look at the women's side, in 2000 and 2011. Yes, I know the regional constraints were stiffer in 2000; but their 8-team bracket had D3HOOPS #1, #2, #5, and #12; the 16-team bracket added #3 and #6. One computer rating had it at #1, #2, #3, and #6, adding #9. (I strongly suspect there was a database error that elevated #12 to #3.) In 2011, their 8-team bracket had #1, #14, and #10 ahead of them at #12. Their path to the championship game was #10 Denison, #4 Hope at Hope, #1 Thomas More at Thomas More, #6 Chicago, #7 Illinois Wesleyan at Illinois Wesleyan, and #2 Amherst. I consider that worse than the 2009 Bracket of Death, because the pain was spread out to each game AND three were in hostile venues. I don't think any potential champion could hope to win all six such games, so in my mind Wash U's women are the true champions of 2011. Yes, this is my subjective assessment of merit.

I really am not trying to criticize the selection process. I understand the huge problems involved. But would it be too much to ask to receive some acknowledgement, especially when that acknowledgement is given to others?

+++++

But my next question, quite seriously, is this: What do you now think of the sideshow in Atlanta? I hated the idea the first I heard of it, for the very reasons exposed here.