Oh, and for the record, UTR did pretty well in predicting the outcome here. Stanford had its two biggest UTR advantages at Courts 2 and 6, and that's where we got our quickest wins (a nominal upset according to ITA ranking at Court 2, where #40 Lord beat #26 Sanford). The match at 4 should have been a coin flip, but Arbuthnott rolled right over Alexa Graham (more in line with what the ITA ranking would predict, with #43 Arbuthnott playing #95 Graham). The final point came from Court 5, where Stanford had its 3rd-biggest UTR lead, 0.25.
So, Stanford notched wins at its 3 biggest UTR leads (2 of which were upsets according to the ITA rankings), and got its fourth point from what UTR saw as a coin flip and which the ITA had as a solid Arbuthnott advantage. In its 4th-biggest UTR lead (0.22), Stanford was effectively tied (6-7, 2-1, up a break).
Georgia and South Carolina are playing right now--we'll play the winner. UTR and ITA rankings agree that Georgia is better (Georgia has the 3rd-highest UTR Power 6 after Stanford and Pepperdine), but South Carolina won the regular-season matchup 4-1 in Athens, so we'll see. UTR would predict a very close match in singles, as even though Georgia has UTR leads in 5 of 6 matchups, three singles matchups are basically coin flips and the other three are decent Georgia leads, but one of this big leads is at Court 1, where Georgia got smoked the first time around.
1. UGA by 0.44
2. UGA by 0.10
3. UGA by 0.07
4. UGA by 0.44
5. SCAR by 0.11
6. UGA by 0.40
South Carolina currently leads comfortably in all 3 doubles matches and looks primed to win the point. In the regular season, South Carolina
dropped the doubles point before rattling off four fairly comprehensive singles wins, so if they win doubles this time around, they have to feel great about their chances. An interesting match, with all of the metrics and rankings pointing to a Georgia win, but where the head to head result during the regular season went the other way (and it wasn't close, either). That H2H win led South Carolina to be a trendy upset pick (along with UCLA and Pepperdine, who are actually favorites according to UTR).
FYI, UCLA has close UTR deficits on the top 3 courts and big UTR leads on the bottom three, so since they're so good at doubles, they have to like their chances. And Pepperdine has solid UTR leads on all singles courts except for #1, so they should be in good shape.
(05-17-2018, 12:19 PM)terry Wrote: Computers only do what people tell them to do. The ITA designed the computer program, then accepted and published results that defied common sense. That's on them.
If there weren't enough data points, they should have waited until there were more data points.
I do feel for the ITA a bit. They've put together a system that works pretty well for everyone concerned, with two glaring exceptions:
1. Pepperdine is always underranked and underseeded because their conference stinks, even though every other team in the country is scared of their roster.
2. Stanford is always going to be a contender, but has just opted out of the rat race that leads to the ITA rankings, and thus the NCAA seeds. We play fewer matches overall and fewer Top 25 teams in particular, so any objective measure of team quality that looks to the quality of your wins and losses is going to undercount Stanford. But should the ITA change its rankings system because one team has a crappy conference and one team refuses to play the ITA's game?
What I think should happen is for the NCAA to use ITA rankings to determine the
makeup of the NCAA field, but then use a more holistic approach to determining the teams' seeds to correct for its failure to adequately account for the quality of both Stanford and Pepperdine (which has been proven through NCAA results year after year).
Stanford was a really tough team to peg all year. We were the #1 preseason team in the country (according to the ITA, no less), but without Lord in the lineup, we dropped a couple of close matches to Top 10 teams (Texas and Florida) and then got swept at Vanderbilt. At that point in the year, our resume
was trash, so I don't fault the ITA for saying that by an
objective measure of schedule, we weren't very good. We haven't lost since Lord came back, though, and by UTR, at least, we're still arguably the best team in the country, but we played in a Pac-12 having a down year and our best wins are over the #14 team in the country, UCLA. So how do you seed a team with that profile? The correct answer is "higher," but how high? I think the USTA in its rankings had it about right, and that we probably should have been around 6th or so, though I couldn't have complained with anything in single digits. (FYI, if you put us at 6th and bump everyone else down a line, we would have played #11 South Carolina today and would have played #3 Duke on Saturday--both very dangerous teams.)
By using an objective measure to determine the makeup of the field, everyone knows the rules of how the rankings are developed, and I like how fair that is. But seeding should be the NCAA committee's independent judgment, taking both resume and their opinion of how good a team is into account. Pepperdine as a 13 seed is an even bigger joke than we are as a 15. Pepperdine should be no worse than the 4 seed (who they're playing today, so there's some cosmic balance there). If you think Stanford is the best team, then we just took the place of the 2 seed in the tournament, so from here on out, we might as well have been the 2.
