The Hazards of Being Free: Defending Transition Basketball

Despite all the positive aspects of transition offense, taking a high percentage of your shots in transition does not have all that positive of an effect on a team's offense. Take a look at this chart :

The numbers in the table are average offensive ratings taken from every team from the last 3 years, ranked according to % shots taken in transition. The difference between the highest 10% - the 105 teams or so with the highest percentage of shots taken in transition - and the lowest 10% was a mere 3 points per 100 possessions. While this is a noticeable difference, it's really not all that significant.

When the same exercise is applied to defensive ratings and chances allowed in transition, however, there's a noticeable effect.

The 105 teams allowing the highest percentage of transition chances have an average defensive rating of 108.11; conversely, the bottom 105 teams in the same category have a defensive rating of 98.06.

When we're looking at transition defense, most of the things that make transition offensive chances juicy apply in the opposite sense. The defense wants to prevent the easy, effective chances afforded by transition opportunities, and force the offense to work harder for their points.

There is a significant correlation between playing good transition defense and good overall defense; much more so than the converse, at least in NCAA Division I basketball. This essentially makes sense; while taking as many transition shots as you can on the offensive end will have an overall positive effect on the offense - and it does - many of the teams that depend on them (in the extreme) have weaker overall half-court offenses. Forcing extremely high percentage of transition attempts is an attempt to cover a weakness (half-court offense, or getting to the rim), rather than a symptom of good offense.

Allowing a ton of transition attempts, however, is a symptom of poorer overall defenses. An inability to tamp down on or prevent the most efficient shot/possession available to the offense is generally a sign there are larger overall problems with the defense - though there are always exceptions. Generally, though, the worst defenses cluster around the top of the transition chances allowed table.

Villanova was certainly not one of the best defenses at preventing transition chances this year; of the 351 Division I NCAA teams, they ranked 228^th, at 22.0%. This was likely related to a relatively high turnover percentage - they turned it over on 17.3% of their possessions, 103^rd in the country - but it's not a complete explanation. Despite allowing this percentage of transition chances, Villanova finished the season with the 16^th ranked defense, per KenPom.com's adjusted rankings. How were they able to accomplish this?

Preventing transition chances must be considered the most important factor, but holding down opponent effectiveness on such opportunities is a close second. Take a look at this table, similar to what was done above for % of shots allowed in transition.

While the conclusion's pretty easy to draw - good defenses don't let teams shoot it well - the correlation with transition eFG% is extremely strong for a shot that teams take an average of 21.43% of the time. And this is where Villanova's strength in transition D was. According to hoop-math.com's stats, Villanova allowed an eFG% of 51.30%, good for 87^th best in the country - something they needed after allowing that many chances.

And, of course, I had to crunch some of my own numbers on Villanova's transition and half-court defenses. This helps highlight where the strength of Villanova's transition defense lay.

Villanova's strength on defense was obviously its protection INSIDE the 3 point line. They kept teams away from the rim, forcing them to settle for two-point jumpers if they got inside. The average % of attempts from two-point jumper range in transition and the half-court were 18.46% and 33.12%, respectively; ‘Nova's D forced opponents to take 21.74% and 34.71% of their two-point shots away from the rim. Not only that; they managed to defend the shots taken from that range incredibly well - 6 to 7% below the league average, in both transition and the half-court.

The defense also did an incredible job of protecting the rim, especially in transition - teams only got to the rim on 38% of their attempts in transition, more than 10% below the NCAA average for the same. When they did get there, the 55.42% finish rate was also far below the NCAA average (64.11%) at the rim in transition. It's an incredible season long feat, especially for a team a little too prone to turning the ball over (17.4 TO% this year). I'm not sure what to attribute this to, but the team's overall athleticism and good rim protection between Ochefu and JVP (rotating at the 5) certainly contributed. The team accomplished much the same, though to a lesser degree, with attempts at the rim in the half-court. While a slightly above-average number of attempts were allowed at the rim, ‘Nova's D also clamped down on their effectiveness, allowing only a 53.22% conversion rate on attempts at the rim (average in the half-court is 57.33%).

Where their weakness is, and has been, is behind the 3-point line. While they were actually effective at stopping opponents to shoot 3-pointers in the half-court, allowing a 30.36 3PA%, they allowed their opponents to hit an above-average percentage of 3s in both phases of defense. Allowing too many 3s, and letting their opponents hit them, has been a weakness of Villanova's for the last several years. While they're certainly beginning to clamp down (this year was much better than last), it's still a weakness - and a worry in the wake of the departures of seniors James Bell and Tony Chennault (two of the better defenders on the team, at least for portions of the year).

Then, similar to the offensive breakdown by shot clock, I went through the same exercise with Villanova's opponent's offense. The D Rtg is the same as the offensive rating of Villanova's opponents - how effectively they scored points.

Villanova's half-court defense, especially after the 20 second mark of the shot clock countdown, was excellent. They allowed about 0.90 points per possession on all opponent possessions that dragged past 20 seconds on the shot clock - which would be equivalent to a top 5 defense - and was where opponents scored 53% of their points.

Opponents had an assist to turnover rate below 1 for every 5 second block other than 6-10 seconds into the possession, which is excellent. Villanova did a great job preventing healthy ball movement, generating turnovers, and closing on open shooters (in the half-court) for the entire year.

Most other notes from the table have been covered in this article, or previous sections. A few quick-hitters:

• After the first five seconds, ‘Nova denied opponents the rim - the highest % of attempts at the rim, after the 51.14% in the first 5 seconds, was the 37.36% 11-15 seconds into the possession.
• Most of the good 3 point shooting Villanova allowed was concentrated in the first 15 seconds of the possession. Only the 36.07% at 26-30 seconds cleared the league average 3P%, after the 20 second mark.
• Nova's transition D was still excellent, allowing a general total of 1.05 points per possession; though their half-court was better, surrendering only 0.94 points per possession.

Transition Defense vs. Crashing the Offensive Boards

Just a last few notes on something that's interesting to me, with regard to the transition game: the traditional assumption that crashing the offensive boards leads to poor transition defense; and, further, that hitting the offensive glass does not ultimately contribute to winning.

Some very smart teams in the NBA have long punted offensive rebounding in favor of getting everyone back in transition; Gregg Popovich's Spurs and the Boston Celtics (and now LA Clippers) under Doc Rivers essentially gave up on the offensive boards long ago. When teams with that kind of brainpower behind them fold on one of the ‘Four Factors,' there's probably a reason behind it.

Zach Lowe, an excellent NBA writer currently working for Grantland, has already done a better job on this subject than I could ever aspire to, although in the NBA context. He looks at the case of the 2013 Indiana Pacers, a team who nearly led the league in both transition defense and offensive rebounding before eventually making the East Conference Finals and falling in a classic series to the defending champion Miami Heat. Essentially, he casts the Pacers as a team challenging the basic assumption that transition defense suffers if you go for offensive rebounds. There are several indicators he goes through to essentially say: no, the best offensive rebounding teams in the league do not necessarily have bad transition D's. He ends up by positing that the pendulum may have swung too far, and offensive rebounds are undervalued as generators of easy points and ‘overvalued' as transition D killers - there may be ways to extract more value from a balanced approach. It's an excellent read if you're interested in the subject, though obviously a bit dated now - it was written before this NBA season started.

In the article, he shares a paper written by MIT students and presented at the Sloan Sports Conference - a yearly gathering for sports stat nerds. They utilized camera-tracking data from the SportsVU system to determine, among many other incredibly complicated things, the value of sending extra players to crash the boards or sending them back in transition. Their essential conclusion, after developing metrics to measure player position in the various cases (again, an incredibly interesting read, though it's very technical), was that teams were leaving a net of 3-4 points per game on the board by not sending an extra player or two to crash the offensive boards. While the data is obviously raw and doesn't include all NBA games (when they collected it, the tracking systems had only been installed at 15 NBA arenas - though, now that they're at every stadium, the students involved with the study have confirmed much of their original findings with the expanded data set), it's certainly an interesting concept.

There isn't much on this in college basketball, but there was something interesting I found when going through the big data on all teams from the past couple years: teams with the highest offensive rebounding percentages have better defenses than teams with lower ones. The difference can be seen in the following table:

This could be for reasons entirely separate from offensive rebounding - teams with the personnel to be excellent offensive rebounders (good rebounding bigs, large wings, athletic wings, etc.) may correlate with personnel required to form a good defense - but it's certainly interesting to look at.

And, while there is a mild bump in % of chances allowed in transition when looking at a similar chart organized by offensive rebounding percentage, the difference is pretty minor - the top 10 percent of teams (in offensive rebounding) allow 21.94% of their shots allowed in transition, while the bottom 10 percent allow 19.72%. The 2% difference isn't groundbreaking (not ignorable, either).. who knows? It's a tough balancing act, but certainly an interesting question for coaches everywhere to pose.

That about sums up this 'advanced stats'-type look at transition basketball, and how well Villanova did on both ends (and in the half-court) this year. Hope a few of you read the whole thing, and the rest pick up at least a few interesting nuggets before clicking through to something else.

Will be writing about this kind of thing semi-regularly, and semi-officially, for VUHoops for the year. Will probably start going over the final 1-, 2-, 3-, 4-, and 5-man lineup combos for 2013-2014, like I did for most of the FPs I did last year, on a weekly to bi-weekly basis to keep some basketball 'advanced stats' content flowing. Will gladly take suggestions (though, as a pre-apology, I'll only run with the ones I like) on things you'd like to see looked at from a stats standpoint (maybe even football).

Thanks for reading, and thanks to the VUHoops managers for inviting me to write. Definitely has been fun.