Statistical Tests: How Important is the Turnover Battle?

There are certain basketball platitudes out there like “you have to control the glass” or “the team that gets more assists wins the game.” But in both cases, they’re either obvious (the team that got more assists probably made more shots and therefore scored more points) or questionable (if you got a bunch of offensive rebounds, it might just be because you couldn’t shoot the ball that night, so you scored fewer points, and you lost.)

With players like James Harden and Russell Westbrook setting records for turnovers even as they’ve put up MVP-level seasons, and with some teams just looking out of sync sometimes when their motion offense starts to fall apart in any given game, it bears looking into just how important it is to win the turnover battle.

We can start by establishing a baseline. After all, the question of “James Harden had a bunch of turnovers, did the Rockets still win?” has, with rare exceptions this year, been “of course they have”; Houston is 24-4, winners of 13 straight, and undefeated with Chris Paul in the lineup alongside Harden.

There’s also the secondary variable of players not playing all 82 games. Is a game where the Warriors lost not because they turned the ball over but because Shaun Livingston had to start in place of Stephen Curry really a game they lost because of turnovers?

What about bad teams? Do they win on nights they take care of the ball, or is it more simply about making their shots? After all, that’s a fair question for teams in general.

So let’s look at games this season in which a team won or lost the turnover battle by at least five. That should count as a night that there were significantly more extra possessions for one team, and with league pace what it is, a possession is usually worth about 1.1 points on the scoreboard, so that should massively skew the odds for all but the very best or very worst clubs.

From best to worst, through games of December 16:

Houston: 24-4 overall, 4-0 winning, 3-3 losing.

Well, look at that. Our first result! The Rockets don’t win the turnover battle often; in fact they decisively lose it more often than they decisively win it, in part thanks to Harden’s butter fingers (while the Beard has gotten much better after 5.7 turnovers a game last year, he’s still at a way-too-high 4.3, and his 120 total turnovers is fourth-highest in the league.)

Bear in mind, however, the first case of “missing player” comes into this; all three of those losses were without CP3. But so far, so good.

Golden State: 23-6 overall, 3-0 winning, 6-3 losing.

Two of the three losses were to the Rockets and Thunder. It’s also worth pointing out that the Warriors turn the ball over a lot for a team as ostensibly good as they are. A ball that moves around as often as it does for the Dubs is going to occasionally end up in the hands of the other team, so Golden State frequently loses this battle badly, but they make it up with their three-point shooting and more traditional on-ball defense; Dubs opponents are shooting 43.3 percent from the field, the best defensive mark in the NBA.

Boston: 25-7 overall, 3-3 winning, 2-2 losing.

This is a weird one. Boston does a good job taking care of the ball (they’re in the top ten in turnover percentage), but merely taking care of the ball doesn’t correlate in any real way with winning. This is our first contradiction of the rule; the team is 20-2 when the margin is comparable and 5-5 in all other situations. That makes no sense.

Then again, one of those losses, to Chicago, was a game the C’s just plain punted; they sat Kyrie Irving and looked disinterested while shooting 39.8 percent. Or the Bulls have secretly turned good on their five-game winning streak and we’ll watch them pull a 2017 Heat after starting 3-20. Jury’s still out.

Let’s play a lightning round here, going through the next few teams in the abstract and stopping when we hit something weird, shall we? Format is overall, winning, losing.

Cleveland: 22-8, 1-0, 2-3.

San Antonio: 20-10, 4-0, 4-2.

Toronto: 19-8, 8-1, 1-0.

Toronto has the fifth-lowest turnover percentage on offense and eighth-best on defense. They are also ninth in the league in assists per 100 possessions. Ball control and defense? What is this heresy?

Now the data’s about to start getting noisy as hell, because the following 13 teams are all right around .500; it seems like the whole league that isn’t elite or terrible is 16-13.

Minnesota: 17-13, 7-5, 4-0.

See? Noisy. Minnesota wins the turnover battle, then they lose the ballgame, and a few of those losses were to terrible teams. It seems like the Wolves can either shoot poorly and take care of the ball or play fast and loose but find a flow that lets them make shots.

Does Tom Thibodeau have this team on too tight a leash? The best that can be discerned from the numbers is when the team plays the game their coach wants them to play, they lose.

Indiana: 16-13, 4-1, 1-1.

New York: 16-13, 1-0, 2-4.

Detroit: 16-13, 4-2, 1-0.

Washington: 16-13, 4-3, 1-1.

Take a wild guess who wasn’t in the Wizards lineup for a couple of those losses. Hint: His name starts with J and he shares a last name with a key feature of a building. John Roof is a key Wizards player, you know.

Let’s stop and draw breath here and just grab a total for the 11 teams we’ve mentioned so far.

Total Record: 214-108 (.665)

When Winning TO Battle: 43-15 (.741)

When Losing TO Battle: 27-19 (.587)

Take out that wacky outlier in Minnesota and you get .675/.783/.548.

I don’t think we need to go through the rest of the league, do we? You get the point, right?

The notion that the turnover battle is a reliable predictor of winning basketball is absolutely as true as you’d expect if you essentially spotted a team 5.5 points on the scoreboard or more just from one stat alone.

Short of “who made more shots”, this one’s about as solid as these sorts of tests get.