 |
|
| A10 BY THE NUMBERS Recent Articles 02/11/10: 2/11 A10 Postseason Odds Report 01/09/10: 1/9 A10 Postseason Odds Report 01/05/09: The Effect of the Unbalanced Schedule on the 2010 A10 Conference Season 12/30/09: A10 Postseason Odds Report 12/23/09: The Effect of Melquan Bolding's Absence on Duquesne 12/22/09: A10 Postseason Odds Report 12/18/09: A10 Postseason Odds Report 12/11/09: Quantifying the Rebounding Problems of St. Joe's Basketball Links NIT-ology Ken Pomeroy Mid-Majority Yoco Hoops Bracket Board College Hoops Net A10 Forum Storming the Floor College Chalktalk Villanova by the Numbers Valley Ledger March Madness All Season Contact mikebscullin@gmail.com | The Effect of the Unbalanced Schedule on the 2010 A10 Conference Season January 5, 2010 by Mike Scullin As we enter into the Atlantic 10 conference schedule once and for all, the focus has turned to projecting where teams will end up in the standings and what they can do in the postseason. A number of factors will combine to determine the teams' fates this conference season: how well they play on any given day, how well their opponents play on any given day, injuries, coaching, plain dumb luck, and countless others. One of the most important factors in that last category that sometimes goes unnoticed: the unbalanced schedule. With the addition of St. Louis and Charlotte into the conference in 2005 bumping the total number of teams from 12 to 14, the A10 was forced to switch from its 2-division format to one in which all 14 teams are thrown into one big mess. The old alignment allowed for a certain amount of balance in teams' schedules, at least among those within the same division: every team would play their divisional opponents twice, and the teams from the other division once. The new alignment, of course, results in an ugly mathematical problem -- each team must have a schedule of 16 games with 13 opponents to go around. It's no secret to A10 fans that the unbalanced schedule has played a factor in creating some weird outcomes in years past: for instance, Fordham finishing in the top 4 and qualifying for a bye in 2007 and Richmond doing the same in 2008 -- two teams that few if any would argue were actually among the top 4. What follows is an attempt to quantify the effect of the unbalanced schedule on the upcoming season, and to determine the potential winners and losers created by it. In addition to its obvious effect on A10 seeding, I will also take a look at the equally important but less obvious effect it can have on a team's chances for an NCAA at large bid. The Seeding Effect Here is a quick look at what the data through Sunday's games from the Pomeroy Ratings predicts concerning the 2009-10 A10 conference schedule:
For those who don't know, the first two columns represent the expected win-loss record for each team. "Pyth WP" refers to a team's adjusted pythagorean win percentage, which is determined by the following two columns -- Adj OE (Adjusted Offensive Efficiency, points scored per 100 possessions) and Adj DE (Adjusted Defensive Efficiency, points allowed per 100 possessions). The final column refers to the Home & Home Strength of Schedule for each team -- reached by calculating the mean adjusted pythagorean win percentage of the 3 opponents they face twice. This calculation gives a quick idea of what teams are dealing with in their 3 in-conference home & home series -- for instance, Charlotte gets thrown to the wolves against Xavier, Richmond, and George Washington. St. Bonaventure has a cakewalk consisting of Fordham, St. Joe's and Duquesne. In order to see exactly how much the ease and difficulty of these schedules is hurting or helping each A10 team, we can do a little experiment: the control will be a world in which every team plays the 13 possible opponents exactly once, then plays 3 games against a theoretical squad that represents a perfectly average Atlantic 10 team, which will be given a pythagorean win percentage of 0.6668 (the mean of all 14 A10 teams). Using the pythagorean win percentages and the log5 method, we will then compare each school's expected number of wins in 3 games against the teams they actually play twice in the 2010 schedule vs. their expected number of wins in 3 games against average A10 competition. Subtracting the second number from the first number will give us the expected number of wins above expected with equalized schedule, or the catchy "WAwES." A more positive number means the schedule is more favorable for a team, while a more negative number means a schedule is less favorable.
Here's a quick look at the projected A10 standings with actual schedules vs. the projected A10 standings with theoretical equal schedules. The teams in red are those who have siginificantly unfavorable schedules (-.2 wins or more); the teams in green are those who have significantly favorable schedules (+.2 wins or more)
Among the top tier of A10 contenders, the only team significantly affected one way or the other by their schedule is Dayton, with their home & home schedule accounting for -0.35 wins (two-time opponents: Xavier, St. Louis, and Duquesne). They're in a particularly precarious position, because there's a decent chance they will be in an all out war with Richmond and Rhode Island to secure the 3rd and 4th byes. 0.35 wins may not sound like a lot, but in what projects to be a very tight race this creates a real disadvantage. On a related silver lining note: Dayton having by far the toughest conference schedule of the 5 at large bid contenders is the reason I think they can sneak into the NCAA tournament with a 10-6 conference mark. Where the unbalanced monster will REALLY rear its ugly head this year is in the middle of the conference, where seeding is newly hugely important. Starting with the 2010 A10 tournament, seeds 5 through 8 will host seeds 9 through 12 in the opening round at campus sites. To have virtually any shot of winning the tournament this year, a team is going to have to come in with a top-8 seed. For this reason, St. Bonvaventure fans should be thanking their lucky stars. Not only do the Bonnies have by far the most favorable schedule in the entire conference (Fordham, St. Joes, Duquesne) at +0.68 wins; all of their direct competitors for the 6 through 8 seeds have wholly unfavorable schedules: Charlotte (-0.66 wins), St. Louis (-0.44 wins), Duquesne (-0.22 wins), and George Washington (-0.21 wins). With equal schedules, the data would project Charlotte and GW to both finish ahead of the Bonnies for the 6th and 7th seeds, with less than 1 win separating St. Bona from the 10 seed. With the reality of the 2010 St. Bonaventure Dream Schedule Initiative in effect, the Bonnies are projected ahead of Charlotte and GW for the 6 seed with a 0.6 win cushion and are separated from the 10 seed by 1.6 wins. The High Quality Win Effect The second way that we'll judge the equity of the A10 schedule will relate to how it affects the chances of an at large bid for the five teams in contention: Temple, Xavier, Dayton, Richmond, and Rhode Island. Since the committee is smart enough to adjust their perception of how a team performed during the conference schedule based on how difficult said schedule was, the first factor we looked at will probably not hurt a team here. What makes a conference schedule more favorable for an aspiring NCAA at large qualifier? The ability to get as many high quality wins as possible. For the purposes of this experiment, we'll call a "high quality win" any win within the aforementioned group of 5 teams. These wins are tremendously valuable for many reasons: 1) They will almost certainly end up being Top 50 RPI wins. 2) There is a good chance they will end up being wins against other NCAA tournament teams. 3) They are wins against teams that you are in direct competition with for at large bids. 4) They are wins against teams that you are in direct competition with at the top of the conference. Having said that, there are two important factors to consider. First, a win against a high quality team in conference is rewarded at a much higher level than a loss to a high quality team would be punished. Thus, for the purposes of creating an NCAA at large resume, more chances for quality wins is always better. Second, while a road win against a high quality team in conference will be judged somewhat more favorably than a home win against the same team, the disparity in reward between these two situations is not nearly enough to trump the advantage of a greatly increased chance of victory due to home court advantage. It is important to contrast this with out of conference games, where the reward disparity is likely sufficient incentive to schedule road games. In my experience, the disparity is less when talking about conference games because the venue is not under either team's control, and thus a quality conference win is generally weighted relatively similarly by the committee regardless of venue. Thus, for the purposes of creating an NCAA at large resume, having in-conference quality win chances at home is better than having them on the road. In order to figure out which of the contenders have the most favorable schedules for the purposes of piling up high quality wins, our control will be a world in which each of them simply plays the other 4 contenders once on a neutral court. Using the pythagorean win percentages and the log5 method, we will then compare each school's expected number of high quality wins against their actual 2010 schedule vs. their expected number of wins in 4 games on a neutral court against the other four teams within the group. Subtracting the second number from the first number will give us the expected number of high quality wins above expected with equalized schedule, or "HQWAwES". A more positive number means the schedule is more favorable for a team, while a more negative means a schedule is less favorable.
Temple (+0.79 HQWins) and Xavier (+0.73 HQWins) are the clear winners here. While having 5 opportunities for high quality wins is key, the important advantage is that due to the luck of the draw 3 of them are home games. Rhode Island (+0.26 HQWins) and Dayton (+0.25 HQWins) have the advantage of 5 opportunities, but are below Xavier and Temple in schedule favorability because they were unfortunate enough to end up with 3 of them on the road. Poor Richmond (-0.03 HQWins) is forced to live in the control world, getting only 4 opportunities. It's important to note that the committee is not blind, and they obviously will not immediately grant Xavier a bid upon seeing that they have 3 high quality wins in conference without looking at their full body of work in context. However, one of the most important lessons I've picked up in predicting tournament fields over the past five years is that wins against other tournament teams can erase a lot of sins, and losses to other tournament teams are not thought of as sins. St. Joe's got a bid in 2008 despite an underwhelming out of conference performance and a 9-7 A10 record mainly because they got 3 conference wins against NCAA tournament teams -- Xavier twice, and Temple once. The fact that they had 6 chances to do this, and the fact that only one of these wins came on the road did not seem to bother the committee. Looking back on that season, I strongly believe that they would not have have received a bid by putting up the same rate of quality wins in fewer chances (i.e., going 2-2 instead of 3-3 against Xavier and Temple). Thus, quantity and not rate is the most important thing when it comes to amassing high quality wins for a tournament resume, and a higher quantity of high quality wins is the likely result of: a) more chances and b) chances that are more likely to be wins (i.e., home games). |