[ColoradoTree]

Thanks COTree - this is super helpful and provides all the context necessary to understand these choices (both ours and the teams picking their Kick-Off Weekend sites).

It's frustrating that tennis doesn't have some way for the prior season's results to at least partially influence rankings, at least at the beginning of the season. Unlike true team sports like basketball or football, where even losing one starting PG or QB can really alter the entire team's trajectory, tennis is a sport of individuals, so I would think the prior season's rankings would have more predictive power. Certainly, it would be odd if on the pro tour they ranked players for the French Open based only on results since January. Having some sort of 52-week rolling ranking would make much more sense.

To that end, why couldn't the NCAA just tally up the team's collective UTR as calculated from singles matches for 1-6 and from double matches for the three doubles teams? Or at least have that play some role in the rankings? It just seems absurd that so much of the first half of the season is spent managing matches to establish a suitable ranking. 52-week UTR would do much to mitigate that.

Good question and thoughts.  I suppose the answer is that while there's no formal relationship between a team's prior-year performance and the new season's rankings, there's a very strong informal relationship, since the early-season rankings are done by committee and they naturally factor in roster makeup, last year's performance, and the quality of incoming recruits.  For example, here are the first team rankings of this past season, done by the ITA committee: http://www.itatennis.com/ResultsAndRanki...Jan__3.htm

They helpfully show the final 2017 rankings, and with a few exceptions, they're pretty close to how things ended up the prior year. Some of the particular cases with bigger drops/rises are interesting, though:

* The committee thought Texas would be improved, with a preseason #12 ranking compared to a final #22 ranking in 2017. Good prediction, but underestimated how big the jump would be (UT finished #6). 
* The committee thought Ohio State would take a dip due to losing some seniors (#3 final 2017 ranking to #13 preseason ranking), but underestimated how much of a dip they took, as the Buckeyes finished #33.  
* They thought Ole Miss would improve (#32 to #23), but underestimated how big of a jump they would have (Ole Miss finished #8).
* They thought UCLA would improve (#23 to #17), and they were right but again underestimated how much (UCLA finished #10). 

Interestingly, the committee gave Cal the #11 preseason ranking after a #13 2017 finish, and that looks like their biggest swing and miss (finished #52). They thought Arkansas would be better (#18) than they ended up being (#35 plus shenanigans), and they underestimated Duke and FSU, but in each of those cases, the problem was that they stuck too closely to the previous year's rankings, not that they diverged too much.  All in all, they did a pretty decent job of predicting how things would end up.  

I suppose the bigger concern isn't the early-season committee rankings, but how they're affected by the switch to the automated rankings in late February after Indoor Nationals.  But aside from the teams opting out of the tournament like Stanford has the past several years, that methodological switch doesn't create too much havoc, though it obviously puts excessive weight on Indoor Nationals, since that's the only time where good teams have played other good teams at that point in the young season.  I'd be OK with your change for some of the season, but at some point, you have to judge a team based on what it's done in the current season, and that would have to happen before NCAA selection at a minimum, and probably conference tournaments.  I'd also be fine keeping a committee approach longer into the season and not switching over to the automated system until later--or, even better, keeping a committee approach all season long.  That's what I really think should happen, which could take the ITA automated ranking as just one piece of data to measure quality of wins, and then factor in UTR, the USTA experts' poll, and other data to come up with a better ranking.  Last year, the committee should have considered how Lord's injury impacted Stanford and how UTR showed Stanford was a superior team to many higher-ranked ones, but they obviously didn't.  (That said, by playing Indoor Nationals and hopefully with a resurgent Pac-12, the ranking/seeding complaints of the last few years should be a thing of the past, as the gap should close between (1) our resume when compared to other teams and (2) how good we are. If we go into the 2019 NCAAs with the best UTR ranking but, say, the 5th or 6th seed, it'll likely be because we lost to some teams ranked above us, and that's as it should be.)  

There are obviously many differences between college tennis and pros, but even though both have a defined season, there is just one tournament that matters in college tennis, and it concludes the season.  You don't have the same problem with seeding players for four Grand Slams that take place at different points throughout the calendar, so you need some sort of rolling average to fairly seed players for the four most important tournaments of the year.  (A closer analogue to the NCAA tournament is the season-closing events, the WTA Finals and ATP Finals, where only points earned during that season count towards qualifying for those year-end events.)

I think UTR and the prior year's results are the main components of the committee's preseason rankings, which are in turn used for seeding the fall individual tournaments.  UTR fades out over time as more current-year results are available, I believe.  

And an important point to keep in mind is that your resume is recalculated every week to take into account changes in the rankings, so an early-season win over a struggling Stanford team may only count as a win over the #25 or #46 team at the time, but by the time the NCAA field was determined, we were #15, so Vanderbilt, for example, was given credit for beating the #15 team in the final rankings before NCAAs.  Losing to Florida when we did was really damaging to our ranking, because they were unranked then, but as they turned it around in SEC play, that loss became less and less of a blemish, and by the final pre-NCAA rankings, it barely affected our ranking at all because the Gators were back in the top 10.  So, as the season proceeds, some of that early-season noise should get smoothed out.

[ColoradoTree]

Thanks COTree - this is super helpful and provides all the context necessary to understand these choices (both ours and the teams picking their Kick-Off Weekend sites).

It's frustrating that tennis doesn't have some way for the prior season's results to at least partially influence rankings, at least at the beginning of the season. Unlike true team sports like basketball or football, where even losing one starting PG or QB can really alter the entire team's trajectory, tennis is a sport of individuals, so I would think the prior season's rankings would have more predictive power. Certainly, it would be odd if on the pro tour they ranked players for the French Open based only on results since January. Having some sort of 52-week rolling ranking would make much more sense.

To that end, why couldn't the NCAA just tally up the team's collective UTR as calculated from singles matches for 1-6 and from double matches for the three doubles teams? Or at least have that play some role in the rankings? It just seems absurd that so much of the first half of the season is spent managing matches to establish a suitable ranking. 52-week UTR would do much to mitigate that.

Good question and thoughts.  I suppose the answer is that while there's no formal relationship between a team's prior-year performance and the new season's rankings, there's a very strong informal relationship, since the early-season rankings are done by committee and they naturally factor in roster makeup, last year's performance, and the quality of incoming recruits.  For example, here are the first team rankings of this past season, done by the ITA committee: http://www.itatennis.com/ResultsAndRanki...Jan__3.htm

They helpfully show the final 2017 rankings, and with a few exceptions, they're pretty close to how things ended up the prior year. Some of the particular cases with bigger drops/rises are interesting, though:

* The committee thought Texas would be improved, with a preseason #12 ranking compared to a final #22 ranking in 2017. Good prediction, but underestimated how big the jump would be (UT finished #6). 
* The committee thought Ohio State would take a dip due to losing some seniors (#3 final 2017 ranking to #13 preseason ranking), but underestimated how much of a dip they took, as the Buckeyes finished #33.  
* They thought Ole Miss would improve (#32 to #23), but underestimated how big of a jump they would have (Ole Miss finished #8).
* They thought UCLA would improve (#23 to #17), and they were right but again underestimated how much (UCLA finished #10). 

Interestingly, the committee gave Cal the #11 preseason ranking after a #13 2017 finish, and that looks like their biggest swing and miss (finished #52). They thought Arkansas would be better (#18) than they ended up being (#35 plus shenanigans), and they underestimated Duke and FSU, but in each of those cases, the problem was that they stuck too closely to the previous year's rankings, not that they diverged too much.  All in all, they did a pretty decent job of predicting how things would end up.  

I suppose the bigger concern isn't the early-season committee rankings, but how they're affected by the switch to the automated rankings in late February after Indoor Nationals.  But aside from the teams opting out of the tournament like Stanford has the past several years, that methodological switch doesn't create too much havoc, though it obviously puts excessive weight on Indoor Nationals, since that's the only time where good teams have played other good teams at that point in the young season.  I'd be OK with your change for some of the season, but at some point, you have to judge a team based on what it's done in the current season, and that would have to happen before NCAA selection at a minimum, and probably conference tournaments.  I'd also be fine keeping a committee approach longer into the season and not switching over to the automated system until later--or, even better, keeping a committee approach all season long.  That's what I really think should happen, which could take the ITA automated ranking as just one piece of data to measure quality of wins, and then factor in UTR, the USTA experts' poll, and other data to come up with a better ranking.  Last year, the committee should have considered how Lord's injury impacted Stanford and how UTR showed Stanford was a superior team to many higher-ranked ones, but they obviously didn't.  (That said, by playing Indoor Nationals and hopefully with a resurgent Pac-12, the ranking/seeding complaints of the last few years should be a thing of the past, as the gap should close between (1) our resume when compared to other teams and (2) how good we are. If we go into the 2019 NCAAs with the best UTR ranking but, say, the 5th or 6th seed, it'll likely be because we lost to some teams ranked above us, and that's as it should be.)  

There are obviously many differences between college tennis and pros, but even though both have a defined season, there is just one tournament that matters in college tennis, and it concludes the season.  You don't have the same problem with seeding players for four Grand Slams that take place at different points throughout the calendar, so you need some sort of rolling average to fairly seed players for the four most important tournaments of the year.  (A closer analogue to the NCAA tournament is the season-closing events, the WTA Finals and ATP Finals, where only points earned during that season count towards qualifying for those year-end events.)

I think UTR and the prior year's results are the main components of the committee's preseason rankings, which are in turn used for seeding the fall individual tournaments.  UTR fades out over time as more current-year results are available, I believe.  

And an important point to keep in mind is that your resume is recalculated every week to take into account changes in the rankings, so an early-season win over a struggling Stanford team may only count as a win over the #25 or #46 team at the time, but by the time the NCAA field was determined, we were #15, so Vanderbilt, for example, was given credit for beating the #15 team in the final rankings before NCAAs.  Losing to Florida when we did was really damaging to our ranking, because they were unranked then, but as they turned it around in SEC play, that loss became less and less of a blemish, and by the final pre-NCAA rankings, it barely affected our ranking at all because the Gators were back in the top 10.  So, as the season proceeds, some of that early-season noise should get smoothed out.