Why should Va. Tech get into the NCAA, based on its record to date? It doesn't have 4 conference wins; it has 2. It's conference record is 2-3-1. It is currently the 10th ranked team in the conference. Its two wins came against the current no. 11 team in the 14-team ACC and the no. 12 team. It has played three good teams--and lost by a cumulative 0-9 score. It is hanging its hat on a tie with Clemson, which is tied for 5th in the conference. Your argument seems to be that if you are in a good conference and can manage to get a couple of results--even against weaker conference teams, or even a half result against a decent team--you should get in the NCAA tournament. I don't buy that. RPI is skewed too heavily to strength of schedule; it you play good teams but lose, you benefit--perhaps more than teams that are playing and beating solid teams that are not on a level with unc and uva. I think the principle of not giving bids to teams with losing conference records makes perfect sense as it keeps you from rewarding mediocre teams just because they've played a good schedule.
Thank you for the response. I look forward to your Tuesday updates and have read through how the ratings are calc'd. I agree 100% with this statement "Realistically, however, there always are going to be some turnaround teams one won't be able to "predict" in advance, at least not using a mathematical system." With young student athletes and the crazy early recruiting of those athletes we will see more parity across the nation and some wild swings year to year for programs that strike out or find a hidden gem amongst all of these young players. Overall I think the US is seeing a broad increase in soccer skills due to the explosion of youth soccer clubs and leagues.
Kazoo.....V Tech is starting to "warm up" and the tough part of their acc schedule is over. Realistically, they should get 4 wins, maybe even 5. I think they get in based on the amount of total wins, and the fact they will get 4 to 5 wins in the toughest conference in the nation. That seems legit to me, as opposed to having a 2nd place team from a conference where they play nobody of consequence and rack up 15 wins. Each conference is not on equal talent level, which is why not letting a 3-5-2 team from the ACC in the tourney because they have a losing conference record seems absurd to me.
Of course we have to see how every team finishes the season. But pundits are projecting that Va. Tech would be in the NCAA if the selection were today. It is a mediocre team--10th in its conference-- whose RPI benefits greatly from being in a good conference. Here's a team that played va., unc and duke--and didn't score a single goal in those three games. They've beaten two bad ACC teams, two bad SEC teams, and tied Clemson--and everybody is saying, 'oh, well, of course Va.Tech belongs in the tourney.' Really? Could the hokies beat some upper level teams from lesser conferences? Sure, and they could lose, too. Team talent and conference talent are not the same thing. There are a ton of teams that could play with, and probably beat, the hokies but won't get in the tourney because they're not in the ACC and didn't get a chance to get shut out by va., duke and unc. The system is skewed to favor lower level ACC and maybe Pac12 teams at the expense of upper level teams from other conferences. The idea that 10 ACC teams should get in just because it's the ACC is also absurd.
Thanks to a couple of suggestions, I've figured out how to use each week's new ARPI ratings as a basis for simulating future games' results, including the end-of-season conference tournaments. I've just posted two new reports and a summary at the RPI and Bracketology for D1 Women's Soccer Blog, to show new simulated ratings based on actual results through October 9 and simulated results thereafter using teams' October 9 ARPIs, plus simulated NCAA Tournament automatic qualifiers, at large selections, and seeds based on those ratings. From here on out, each week I'll use that week's new ARPI ratings as the basis for my simulations. This will make the simulations increasingly more realistic and useful.
It is unlikely that the ACC will get either 9 or 10 teams into the tournament, as it hasn't happened over the last 9 years, so I wouldn't recommend suffering angst about that possibility unless and until it actually happens. The statement, however, that the system is skewed to favor lower level teams from the strongest conferences is incorrect. I've demonstrated, in great detail, that when you compare teams' ratings to their results, on average teams from stronger conferences outperform their ratings and teams from weaker conferences underperform their ratings. This is a function of how the ARPI is structured. If you want to see for yourself why I say this and see the documentation for it, go to the RPI for Division I Women's Soccer website and trudge through the pages.
I've just posted three new reports and a summary at the RPI and Bracketology for D1 Women's Soccer Blog. The three reports are (1) Weekly RPI Report: Games Through October 16; (2) 2016 Season Simulation: Week 9 Update (Games Through October 16); and (2) 2016 NCAA Tournament Bracket Simulation: Week 9 Update (Games Through October 19).
I've just posted three new reports and a summary at the RPI and Bracketology for D1 Women's Soccer Blog. The three reports are (1) Weekly RPI Report: Games Through October 23; (2) 2016 Season Simulation: Week 10 Update (Games Through October 23); and (2) 2016 NCAA Tournament Bracket Simulation: Week 10 Update (Games Through October 23).
For NCAA tournament selection, how does it work ? Is first round based solely on geography ? Or are teams 1-16 protected (can only play a much higher seed) ?
For a detailed description of how it's done, use this Bracketing link. Scroll down that page to the "Bracketing Criteria" link.
I've just posted three new reports and a summary at the RPI and Bracketology for D1 Women's Soccer Blog. The three reports are (1) Weekly RPI Report: Games Through October 30; (2) 2016 Season Simulation: Week 11 Update (Games Through October 30); and (2) 2016 NCAA Tournament Bracket Simulation: Week 11 Update (Games Through October 30).
I'm almost afraid to click on the link after Penn State's horrible performances in the past 2 weeks. "Do they even make the NCAA tourney?" he asks with his hands over his eyes as he clicks the link. EDIT: Looks like they will, by the skin of their teeth (#6 rank in a regional bracket with an Adj. RPI of 30). I can live with that considering.
This post is so wrong I had to sign up and answer, I see it every year. Bottom feeder ACC teams beating teams in the regular season that go to the tournament. One year I saw a 1 win ACC school with 4 wins and a tie against tournament teams. VT, NC State, BC, Miami, Wake.... all would win most conferences this year and will likely not make the tourney.
I've just posted a 2016 NCAA Tournament Bracket Simulation: Update Including Games Through Friday, November 4. Preceding the Simulation, I've posted a Summary that discusses some of the specifics of the Simulation and issues the Women's Soccer Committee may face for specific teams. You can find them here: RPI and Bracketology for D1 Women's Soccer Blog.
I've just posted a 2016 NCAA Tournament Bracket Simulation: End of Season -- FINAL -- Simulation. The underlying data include all games of the season including conference tournament games. You can find it here: RPI and Bracketology for D1 Women's Soccer Blog. There are some very difficult decisions for the Women's Soccer Committee. Later, I'll post some thoughts about what I think the hard decisions are and what the issues are. But, for the benefit of those of you who follow the simulations, I've put the FINAL Simulation up there now. Reminder: The Simulation is based on this year's data and the Committee's decisions over the last 9 years. It's not really a prediction, it's just a report on what the Committee will do if it follows past precedent.
I've just posted a piece on the issues that Princeton poses to the Committee as it completes its NCAA Tournament bracket decision-making today. You can find it here: 2016 NCAA Tournament Bracket: Princeton. For those of you interested in the intricacies of applying the Committee's decision-making criteria to specific teams, Princeton this year presents an interesting case study.
As promised, I've posted a detailed analysis of the Committee's putting Oklahoma State and Texas A&M "in" and DePaul and Iowa State "out." You can find it HERE.
A few of you may have noticed that the NCAA's just-released RPI rankings and mine (and Chris Henderson's) are not exactly the same. This is because the NCAA has a data error in its system. I've posted detailed information on this at my RPI and Bracketology blog, HERE. The error may have affected seeds, but very likely not at large selections.
For those of you who followed my simulations this year, I've just finished a study to find the best way to come up with simulated ARPI ratings in order to do next year's (2017's) pre-season simulation for the entire season. As a side benefit, the study came up with some information on how long a new coach's teams' results are affected by the team's past history. If you're interested, I've reported the results of the study here: How to Simulate Pre-Season RPI Ratings for Next Year; and a Bonus on Coaching Changes and ARPI Trend Changes
Not without some merit. NC State finished 8th in the ACC, then went on huge road trips and beat the Big 10 Champion and the West Coast Conference Champion in the NCAA Tournament.
I've completed one last piece of work related to my simulations work, which is pinpointing when to shift the simulation from using pre-season assigned ARPI ratings to using current actual ARPI ratings for purposes of simulating the results of games not yet played. If you're interested, you can find the results of that piece of work here: When to Switch Simulation from Using Pre-Season Assigned Ratings to Using Actual Current Ratings? There's an interesting piece of info about the ARPI that came out of this last piece of work: So far as how well the ARPI ratings correlate with actual game results, there is virtually no improvement over the last three weeks of the season (Weeks 10 through 12) and little improvement over the last five weeks.