Close Games -- Who the 'Cats Relied on at Crunch Time

Scottie and Who?
I ran across an article, "Ice Water in Their Veins", by College Chalk Talk publisher Chris DiSano that looks at how a number of A10 guards performed in tight games. A terrific read by the way, one that should appeal to an audience larger than A10 fans and the five schools whose guards Chris took the time to compile statistics about. And for those fans (mostly from Xavier I suspect...) who question Kevin Anderson's A10 POY award, Chris provides a powerful statistical argument for why the Spider guard was so deserving. With Chris' approach and methodology in mind, I wondered if the same methodology, applied to Villanova's games last season, could provide some insight into what happened, especially during the team's stretch run. Games from February and March -- that's one frog I have been reluctant to dissect this off season -- but Chris gave me a few ideas on how to organize the data and focus my questions.

A Few Tweaks & Modifications On Chris' Approach
Chris wanted to know how Kwame Mitchell compared to Jordan Crawford, among others, in high pressure situations, so compiling the raw data from play-by-plays would be sufficient to see who took the shots and (more importantly) how efficiently they converted those shots to points. What I wanted to know (among other things), was who played during the crucial sequences, who among those on the floor was most likely to take the shot, and how they did. Like Chris I wanted to look at close games, but since this was about a specific team over the course of a season, I decided to look at all close games, regular and post season games, if they met my "close game" criteria. Chris looked at the end score to decide if the game was "close". I decided to look at the score at the five minute mark, and if the game was "close" (+-5 points) at that point in time, I would break down the play-by-play. Of Villanova's 33 game season, the 'Cats and their opponents were within five points (roughly 2-3 possessions) in 11 of those games (that's about 1 in 3 games)...

		@5:00		Final
Date	Opponent	PF	PA	PF	PA	Notes
Thu Nov 19	George Mason	56	56	69	68
Sun Dec 13	at Temple	59	63	65	75
Sat Jan 2	at Marquette	67	66	74	72
Sat Jan 9	Marquette	67	63	78	76
Mon Jan 11	at Louisville	77	74	92	84
Sun Jan 17	Georgetown	67	65	82	77
Sun Feb 21	at Pittsburgh	49	51	65	70
Sat Mar 6	West Virginia	50	45	66	68	OT (60)
Thu Mar 11	Marquette (BET)	57	62	76	80
Thu Mar 18	Robert Morris (NCAA)	47	51	73	70	OT (58)
Sat Mar 20	St. Mary's (NCAA)	61	61	68	75

Of the 11 games close at the 5:00 minute mark, three (Temple, Louisville and St. Mary's) were not "close" at the buzzer. Of the 22 games "not close" at the 5:00 minute mark, one, Villanova's March 2 game at Cincinnati, was a "close" game at the end (a 77-73 win), but the 'Cats held a 10 point margin with five minutes to go. A play-by-play breakdown is not necessary to see that in the six close games in November through January Villanova was 5-1, while in the five close games in February and March (notice the last four games Villanova played were all "close"?) the 'Cats were 1-4. Comparing the players' minutes, scoring efficiency and rebounding from the first set of games to those from the second set does reveal several changes over the course of the season. The first six games...

						FTA
	%Min	%Shot	eFG%	PPWS	OR%	FGA
Armwood	13.4	18.2	150.0	2.00	0.0	200.0
Cheek	30.2	8.1	0.0	0.33	0.0	400.0
Fisher	50.3	38.8	12.5	0.36	6.9	75.0
King	57.0	17.1	87.5	1.75	12.1	0.0
Pena	67.1	14.5	25.0	0.80	10.3	50.0
Redding	57.0	21.4	40.0	1.00	6.0	160.0
Reynolds	97.3	30.1	62.5	1.35	7.1	133.3
Stokes	67.1	7.3	0.0	NA	0.0	0.0
Sutton	13.4	18.2	100.0	2.00	25.7	0.0
Wayns	30.2	24.2	50.0	1.40	0.0	133.3

Scottie's 30.1% shooting percentage (he took 30% of the available shots when he was on the floor) is no surprise. Overall Scottie was taking 25-27% of the shots during the season (see Villanova's Team Page over at Ken Pomeroy's web site); taking over in tight games is consistent with "go-to guy" responsibilities, and the staff, by allowing him to take 97% of the minutes in those situations (far above his overall Min% for the season), suggest they expected him to take over in those situations. The surprse (there are a few of them) is Fisher's 38% shot rate. In his freshman season Corey had games where he took a very large proportion of the shots, about 28%-30%, depending on the game. In this first series of games however, Fisher struggled with his shot. A 12.5% eFG% is terrible, and given Corey's tendency to "keep shooting", finding he gets about 50% of the available time is hardly surprising. Wayns and Redding (whose suspension ended in mid December, making Reggie available for the last four games in the series) were servicable second options on offense. Neither posted eye-popping shooting numbers, but (check their FTA/FGA rate) each was a slasher and able to get to the line. King, Sutton, and Armwood were efficient secondary contributors. King in particular shot an efficient 3-4 from the court and 2-2 from the line. Stokes' numbers are extremely low, like Cheek and Pena, his contributions were not in shot conversion, even though he pulled in 67% of the playing time in those high pressure situations. Some of the secondary numbers...

	FGM%	3FGM%	2FGM%	FTM%	Ast%	DR%
Armwood	100.0	100.0	NA	50.0	0.0	0.0
Cheek	0.0	0.0	NA	25.0	0.0	19.8
Fisher	12.5	0.0	50.0	33.3	3.3	4.8
King	75.0	50.0	100.0	NA	0.0	28.2
Pena	25.0	NA	25.0	100.0	0.0	13.8
Redding	40.0	0.0	50.0	62.5	0.0	9.4
Reynolds	50.0	75.0	37.5	75.0	1.9	8.7
Stokes	0.0	0.0	0.0	NA	0.0	4.6
Sutton	100.0	NA	100.0	NA	0.0	0.0
Wayns	33.3	100.0	0.0	100.0	1.4	0.0

A check on King's and Pena's OR% (offensive rebounding rate) and DR% (defensive rebounding rate) provide a big hint about why these two were on the floor. Neither was the focal point of the offense, but both were very strong on the boards. King's numbers in particular are impressive. While the staff clearly favored a rotation of Reynolds, Pena, Stokes, Redding, King and Fisher, the staff used 12 players total over those six games, using either eight or nine players in five of the six games. Seven players saw action in the Temple loss. Data for the second series of games shows a very clear break in player roles and the rotation used by the staff...

						FTA
	%Min	%Shot	eFG%	PPWS	OR%	FGA
Armwood	8.8	NA	NA	NA	0.0	NA
Cheek	26.3	16.9	0.0	NA	0.0	0.0
Fisher	70.2	31.7	65.0	1.41	0.0	70.0
King	11.7	NA	NA	NA	0.0	NA
Pena	67.3	13.2	75.0	1.57	0.0	150.0
Redding	84.8	13.1	0.0	0.50	9.8	120.0
Reynolds	90.6	34.3	39.3	1.23	4.6	135.7
Stokes	90.6	14.7	100.0	1.86	0.0	33.3
Sutton	5.8	38.0	0.0	NA	0.0	0.0
Wayns	17.5	12.7	100.0	2.00	0.0	0.0
Yarou	26.3	16.9	100.0	2.00	15.8	200.0

The time allocation shuffles aside, the most obvious change is Scottie and Fisher. Scottie's efficiency (eFG%) crashed, though he continued to get to the line. The number of shots he took rose very slightly, but his Shot% increased 4% because his minutes declined slightly (possession-based stats calculate proportions based on the percentage of time a player is allocated), but he continued to get to the line. Fisher's role (Shot%) declined even as his playing time (and efficiency) increased dramatically. Stokes' time also increased, his outstanding shooting efficiency (coupled with improved rebounding numbers) making a argument to increase his time. The freshmen, Yarou excepted, saw their minutes remain constant or decline slightly. Yarou, his early season protocol largely completed, was working himself back into the rotation. The offense revolved nearly exclusively around Scottie and Corey F, all other members of the roatation, though they might have slightly more time allocated, nearly disappear during crunch time. The secondary numbers...

	FGM%	3FGM%	2FGM%	FTM%	Ast%	DR%
Armwood	NA	NA	NA	NA	0.0	0.0
Cheek	0.0	NA	0.0	NA	0.0	0.0
Fisher	60.0	50.0	62.5	85.7	0.5	0.0
King	NA	NA	NA	NA	1.9	6.3
Pena	75.0	NA	75.0	83.3	0.0	19.6
Redding	0.0	0.0	0.0	66.7	0.3	18.4
Reynolds	35.7	16.7	50.0	94.7	1.3	9.0
Stokes	66.7	66.7	NA	50.0	0.0	13.5
Sutton	0.0	NA	0.0	NA	0.0	0.0
Wayns	100.0	NA	100.0	NA	0.0	0.0
Yarou	100.0	NA	100.0	100.0	1.0	5.8

Scottie's three point conversion rate crashed, the single biggest reason for his overall decline in efficiency during crunch time in those games. Rebounding also disappeared. Offensive rebounding, Yarou and Redding excepted, virtually disappeared. Scottie snagged a rebound, but the squad as a whole did not go after their teammate's missed shots. Defensive rebounding was only very modestly better. Pena, Redding, Stokes and even Scottie grabbed defensive rebounds, but the overall rate was low. Yarou showed a good instinct for offensive boards, but was strangly ineffective on the defensive boards.

Random Thoughts...
1. Marquette and Villanova played three games in the 2010 season, all three decided by four or fewer points. Villanova won the two regular season meetings by two points apiece. The Golden Eagles took the Big East Tournament game by four points, effectively evening the series (by point differentials). The reason for the close games (even those played last season were relatively close) may be due to the very similar style of play. The 'Cats may have Yarou for the next season or so, but through last season (the first two meetings anyway) neither squad had a 6-10 or taller big who was effective.
2. The assist rates were surprisingly low, even when the squad was winning the close games. The style of play values one-on-one play and lane penetration (which explains the relatively high FTA/FGA rates), but both Scottie and Fisher were top 500 (per Pomeroy) assist leaders during the season, logging rates >21.0%. Assist rates below 10% for both players suggest the squad's two strongest offensive players did little to involve their teammates at this very crucial time.