Steals and the Two-Way Play

facebooktwitterreddit

Feb 6, 2015; Atlanta, GA, USA; Atlanta Hawks guard Jeff Teague (0) is fouled by Golden State Warriors guard Stephen Curry (30) on a steal attempt in the third quarter of their game at Philips Arena. The Hawks won 124-116. Mandatory Credit: Jason Getz-USA TODAY Sports

Friend of the blog Danny Leroux likes to talk about the importance of feedback loops in basketball. A defensive stop can lead to an easy basket, which allows your defense to set up another good defensive possession and so on. This is the very essence of how a team puts together a run, forces a timeout, and GINO TIME! Conversely, bad offense can lead to bad defense, making the next offensive possession an uphill slog.

No single play better epitomizes this self-reinforcing of good[1. Or bad, depending on who’s doing the stealing.] play than the steal. When coaches and analysts talk about “points off of turnovers” they should really be focused more on points off of steals. Though an undesirable outcome, a deadball turnover, like a travel, out of bounds or charge[2. #BanTheCharge.] does bring the small consolation of eliminating the risk of an opposing fast break. As one of my high school coaches liked to scream “If you’re going to throw that pass, why don’t you just throw it up into the stands. At least then we can get back and play D!”

But the steal not only negates the current offensive possession, it severely undermines the ensuing defensive effort. In 2014/15 shots taken immediately after a steal[4. That is with no intervening action in the play-by-play logs, I haven’t gone back and watched all 19,000+ steals recorded in the NBA this season or anything.] had an Effective Field Goal % (eFG%) of 62.1%, as compared with an average eFG of 49.6%. Unsurprisingly, the extra damage was mostly early in the shot clock, before the now-defense had a chance to recombobulate from the sudden and unexpected change of possession:

And in fact, the steal is a great way to get early, easy offense in general:

Another way of looking at how much steals impact the ensuing offensive possession is a comparison of outcomes between the next play-by-play event following a steal and following a made field goal. While this doesn’t actually capture every possession following a steal that’s ok, and in fact a benefit! Any whistle or stoppage of play largely negates the “specialness” of a post-steal play, unless we are to believe there is something psychologically and categorically different about either attacking or defending an inbounds play if the last possession was a steal. In this analysis, anything which interrupts the “flow” enough to show up in the official play-by-play is considered to have stopped play enough for both teams to regroup.[1. “Other” includes a wide array of stuff such as technical fouls, ends of quarters, substitutions from the ball being deflected out of bounds, kicked and jump balls as well as the few extraneous entries in the play-by-play data.]

To simply it down a little, after a steal, “good” outcomes (made shots, drawn fouls) make up 58.9% of the good or bad outcomes[3. That is, ignoring timeouts and “other”.], whereas only 49.5% of plays following an opponent’s made field goal are positive. Think of any close game and how large would be the impact of 8 or 9 more “good” possessions for one team represented by that differential. It’s a big deal.

One of the chief worries for the Warriors heading into the playoffs was their propensity to throw the ball all over the court in search of highlight plays. Golden State’s overall turnover rate was slightly below league average this season, [4. 13.1% for GSW compared to 13.3% NBA-wide per Basketball Reference.] However, only Portland saw a higher ratio[6. 58.6%.] of “live ball” to total turnovers than Golden State’s 57.4%. Among playoff teams, the Warriors had the third highest rate of steals allowed at 7.5% behind only Milwaukee and Houston at 8%.

This is important because Golden State’s overall defense was so good. They led the league in overall eFG% defense at 47%. However, on those plays following a Golden State live ball turnover, this rocketed up to 57.9%. The playoffs have seen this problem emerge at times, as all three opponents have gone on large runs keyed in part by the Dubs getting a little loose with the ball. Still, it’s not as bad as could be expected, the Warriors were actually among the better teams in terms of defending live ball turnovers – that ghastly 57.9% was good enough for fourth best in the league.

By comparison, their Western Conference Finals foe in Houston were the team who had really struggled in this regard. As mentioned above, Houston was third overall and tied for second among playoff teams in terms of highest proportion of their offensive possessions ending in steals.[10. Philadelphia’s woeful offense somewhat unsurprisingly gave up steals at the highest rate with the 76ers giving 8.6% of pick-6’s.] Not only did Houston turn the ball over a lot, they defended especially poorly on the ensuing play, giving up 64.6% eFG shooting (24th in the league) which is even worse than it might look given their strong overall defense as only Utah and Sacramento saw bigger gaps between their overall field goal defense[9. Overall defense splits following steals detailed below

] and their defense following live ball turnovers. If throwing the ball to the wrong-colored jerseys and not getting back on defense sounds familiar, it’s because that’s a lot of what happened in the first four games of Houston’s previous series with the Clippers before they somehow managed to turn it around.

Despite the Warriors’ 2-0 lead, this series has been and looks to continue to be very competitive. The ability to force or prevent just one more easy basket off a steal might well be the difference between the Warriors taking an insurmountable 3-0 lead or Houston clawing their way back into it. Those few extra good (or bad, depending on your perspective) possessions could decide things.