Friday, September 10, 2010

Dayton in Crunch Time, 2010

Dayton and Pythagoras
One of the interesting story lines in the Atlantic-10 Conference last season was the Dayton Flyers' "failure to launch". Through much of the out of conference schedule the Flyers won the games they were expected to win though they were not dominant in those wins. Playing a 37 game schedule, the Flyers participated in 13 games decided by five points or less, and logged an underwhelming 4-9 record in those games. This viewer's lofty expectations were shared by an array of 2010 preseason predictions, and justified (to a modest degree) by the Flyers' 0.971 Pythagorean Winning Percentage (PWP), as computed by Ken Pomeroy, who ranked the Flyers #26 in Division 1. I wondered, based on Pomeroy's log5 predictive model, whether the Flyers met expectations with those results (ie if they were expected to win, did they win, and if they were expected to lose, did they lose?). The table below suggests that the team in fact "under performed", as many felt they had...


Of the nine close games they were "supposed" to win (Georgia Tech -- yes, Pythagoras looks at points for and point against, not conference affiliations -- Towson, George Mason, Rhode Island, Saint Joseph's, both St. Louis games, Duquesne and Mississippi), they only won four (Tech, Towson, Mason and Ole' Miss), and of the four they were "supposed" to lose (New Mexico, both Xavier games and Richmond)...well, they did not exceed expectations. The results become even more interesting when broken out by (rough) chronology -- Out of Conference (early season), A-10 Conference games and the post season (A-10 Conference Tournament & NIT). The Flyers' record...

Out of Conference310.750
A-10 Regular Season070.000
Post Season110.500

What Happened to Dayton?
Staff Writer Doug Harris from the Dayton Daily News penned a long and thoughtful analysis piece that settled on three explanations...
1. The Seniors Did Not Step Up -- for whatever reason (injuries perhaps?) seniors Rob Lowery, Marcus Johnson, London Warren and Kurt Huelsman in particular, did not see their games progress from their junior to the senior seasons. And that brought the team up short in tight games.
2. The Dynamic Was "Off" -- Saint Joseph's coach Phil Martelli wondered if the loss of A-10 All-Sixth Man Charles Little may have thrown the dynamics on the squad off just enough to come up short in close games.
3. The Coach Got It Wrong -- Coach Gregory's substitution pattern and game management had caused distractions that ultimately undermined the team effort. Ok, there were four reasons, sort of...
4. Dumb Darn (Buzzard) Luck -- Harris seemed to settle on this one, just stupid individual game circumstances coupled with random bad breaks.

While all were plausible (ok maybe not the luck one, at least not the variation Harris posited), how many were verifiable using stats and the play-by-play logs? I decided to look at all 13 games decided by five or fewer points. I took note of the score at the 5:00 minute mark, and then charted the plays (and the Flyer players) from that point to the end of the game (note in the first St. Louis game the Flyers and Billikens had to play two overtime periods to decide the contest)...

A-10 Only241

I looked at Oliver's Four Factors for the last five minutes (and the two OTs) of the seven close A-10 regular season conference games, and teased out a few hints on what did and did not happen. But I also discovered a paradox that begged to be resolved...


How could the Flyers be 0-7 when their offensive efficiency (105.8) in the last five minutes of those games was greater than their defensive efficiency, 104.5 (Dayton's defensive efficiency is their opponents' offensive efficiency)? In three of those seven games the Flyers' offensive efficiency was excellent -- good enough to bury any opponent had it been sustained for a game, or even a half -- against Xavier (it was 127.1) on 1/16, Saint Joseph's (156.4) on 1/23 and St. Louis (142.9) on 3/6. The problem in each game however, is that at the 5:00 minute mark, the hole the Flyers had to dig themselves out of was pretty deep. The Flyers were -6 against Xavier (they lost by 4), -9 to Saint Joseph's (they lost by 1) and -9 to St. Louis (they lost by 5). Among the four factors, the numbers that stand out are the Flyers' offensive rebounding rate (24.3%), the defensive turnover rate (12.8%) and the defensive eFG% (46.4%), none of which was characteristic of Dayton's overall A-10 stats. Their opponent's FTA/FGA rate (101.3%) may be well above their A-10 average (41.9%), but consider that the team that is down typically fouls to gain possession, a higher than usual FTA/FGA rate would not surprise. The Flyers' conference offensive rebounding rate was 36.8% however, over 12% higher than their rate in the last five minutes of those seven close games. Juggling the rotation to put shooters (as opposed to rebounders) might explain the significant drop, but one of the Flyers' best offensive rebounders, sophomore wing Chris Johnson, logged about the same amount of time, yet had a 7% decline in his offensive rebounding rate (see below). Rob Lowery and Chris Wright aside (see below), just about everyone else in the "5:00 minute" rotation saw their offensive rebounding decline over their numbers from the first 35 minutes.

The Flyers Crunch Time Rotation -- Who Got the Ball?
I wanted to know who was handling the ball and taking the shots at the end of close games. Given that seven of Dayton's eight conference losses were by five or fewer points, their 0.792 PWP in conference games was a striking contradiction of their 8-8 conference record. What happened during conference play? While I compiled those "5:00 minute" stats for all 13 games, I decided to focus on those seven Atlantic-10 games in particular and look at who took the possessions and shots during crunch time...

C Wright 97.722.421.51.2353.812.86.053.8
C Johnson 65.836.744.10.7733.
M Johnson 59.521.416.31.0050.
L Warren
P Williams
R Lowery 53.643.733.11.2754.59.322.345.5
K Huelsman 44.914.07.20.680.00.041.6100.0
M Perry 32.027.525.20.6030.
D Searcy 22.915.414.11.54100.00.00.0200.0

With a few exceptions, the time allocation was fairly consistent with the Flyers' overall playing time. If preseason All-American nominee Chris Wright saw his playing jump to 98% of the possible time (up from 70.8%), no one should be surprised. A few of the seniors, Rob Lowry and London Warren saw their Min% increase modestly as well, though seniors Mickey Perry and Kurt Huelsman saw their proportion of the playing time decline slightly. The staff benched the deeper rotation players in favor of the more experienced players (and Chris Wright and Chris Johnson). The reallocation of possessions and shots however, was far more drastic. The calculation I used estimates a players proportion of possessions and shots by the amount of time the player was on the court. Since Wright was almost always on the court (he missed just over a minute in Dayton's loss to Duquesne), his relatively modest allocation of touches and shots can be explained by his enormous allocation of time -- Wright pretty much played consistently within the Dayton system. While he did take the second highest number of field goal attempts during his playing time (13), he did not take over any of those A-10 games.

What About the Seniors?
If Wright was playing within the offense, much as he did during the other 35 minutes, the same cannot be said for senior Rob Lowery, nor for sophomore Chris Johnson, both of whom took (to use Ken Pomeroy's terminology) Go-To Guy level possessions and shots. Perhaps these were the result of set plays drawn up by the staff, or maybe the players were winging it in crunch time, but for Lowery, the larger role seems, on paper, justified. The senior converted his FGAs at a 54.5% rate, well above most of his fellow seniors, not to mention Chris Wright and Chris Johnson. If Perry and Huelsman's contributions were limited by coaching decision, what of Marcus Johnson and London Warren? Both received larger portions of the playing time, consistent with a strategy that would try to put experienced players on the floor when possessions mattered. Neither Warren nor Marcus Johnson however, seemed to shoulder more responsibility for offensive production. Indeed Warren seemed content to dish assists and go to the line for free throws, rather than take field goal attempts. Marcus Johnson's decision to take fewer possessions (and shots) may well have derived from an untimely high turnover rate (41.0), combined with a conversion efficiency, 50%, that was only slightly better than his overall conversion rate of 48.4%. Late game turnovers have a habit of being amplified over early game turnovers. His defensive rebounding rate, coupled with a 4.4% steal rate, no doubt made a case for keeping him in for defensive purposes. Warren had a very good 21.2% assist rate, but became invisible for shooting, taking only 2.7% of the available shots, as opposed to the very modest 10.3% he took overall.

What About the Coach?
On average Coach Gregory did make 9.1 substitutions in the seven A-10 losses, while he averaged about 8.3 substitutions in the other six games. A closer look however shows that in two of the seven A-10 losses Coach Gregory made 15 (Xavier on 1/16) and 17 (St. Louis 2/13). The traditional rivalry with Xavier, coupled with the fact that it was the first close conference game the Flyers played, may explain why the coach committed 15 substitutions in five minutes (there were no player DQs from fouls). The St. Louis game should be put in perspective -- there were two overtimes in addition to the last five minutes of regulation. 17 substitutions in 15 minutes should not be too disruptive for a squad's rhythm. Setting those two games aside, the Flyer staff substituted an average of 6.4 times, with a standard deviation of 1.4 (ie he subbed just over once per minute in the lasst five minutes of the other games, and did not vary the number of substitutions much from close A-10 game to A-10 game).

Chris Wright 97.75.711.91.3
Chris Johnson
Marcus Johnson 59.57.315.64.4
London Warren
Paul Williams
Rob Lowery 53.656.00.02.4
Kurt Huelsman
Mickey Perry
Devin Searcy

What About the Chemistry?
Sometimes a single incidnt can influence perceptions about a player or a team. Rob Lowery's meltdown at the end of (yet another) close Xavier game in the Atlantic-10 Conference quarter-finals may have shaped attitudes about his contributions, particularly when the game was on the line. Not only was his shot efficiency (eFG%) better when the game was on the line, he tended to take more shots (proportionally) when the stakes rose. His assist rate (Ast% in the table above -- 56.0%) and offensive rebounding rate (9.3%) suggest that he did not see himself simply as a scorer and shooter at crunch time. His two major detractors, based on the play-by-play, were a bad habit of turning over the ball (22.3%, lower actually than his overall 27.3% rate), and that technical he drew against Terrell Holloway in Atlantic City. None of the theories noted Chris Johnson's rather large, and largely inefficient, role in the offense. Johnson's proportion of the minutes increased in crunch time, clearly he had the confidence of the coaching staff. He was second in time only to Chris Wright, but his inability to convert efficiency in the clutch however seems to have been largely overlooked by fans and staff. His strongest virtue at crunch time was that he did not turn the ball over. But his tendency to shoot (he took nearly 50% of the available shots when he was on the floor in the last five minutes; he took the most FGAs, 18, of any Flyer player in those seven conference games, and 13 of those shots were three point attempts.

What About Luck?
No question the Flyers were unlucky, but was it dumb bad luck or something that can be quantified? Comparing a teams' record with it's Pythagorean Winning Percentage can provide a standard measure, especially when you do it for all 347 Division 1 teams. At -.108, the Flyers ranked #340 out of 347. They were very unlucky compared to their Division 1 counterparts (and the unluckiest team in the A-10). The good news for the Flyers -- teams with bad luck one season tend to rebound the next. With a PWP of .9170, the Flyers should win between 10 and 12 conference games next season.

Why This Matters in 2011
While it is tempting to put 2010 in the rear view mirror, the Flyers are returning the two players who logged the highest proportion of crunch time minutes. Chances are very good those two will see the largest proportion of crunch time minutes again in 2011. Chris Johnson, as useful as he was for defensive rebounding will have to do better with his shot. Paul Williams and Devin Searcy will most likely get long looks in Fall practice, but it is unlikely that the staff will give the green light to entering freshmen Juwan Staten (pg) or Brandon Spearman (sg) to take 33.1% (much less 44%) of the available shots.

No comments: