How to Read Small Sample Sizes

10 mins read

In the spirit of Ben Falk’s impressive piece about sussing out the truth from small sample sizes, it’s worth trying to apply some of the same tests to the Raptors’ early 2017-18 season. This is a practice in methodology; how can you tell what matters and what doesn’t after only a week of basketball?

While this is a complicated subject, some parts are easy. Context can write off some trends without too much thought. Clearly, the Raptors do not boast four of the five best players in the NBA; they just played the horr-awful Bulls and led-by-all-rookies 76ers before this chart was created.

One thing is useful from that chart: The Raptors’ most successful minutes of basketball have occurred with bench players on the floor. This is statistically significant, if just to signify that the kids are ready (though decidedly not better than the starters). They play like happy weirdos, zipping the ball around to create quick-trigger 3s for C.J. Miles, who according to 538 is in some respects the 2nd fastest shooter in the league.

Delon Wright and Fred Van Vleet in particular play with an impressive mind-meld, always knowing where the other is. They play with a syncopated beat, like a jazz drummer and bassist fusing together, passing and cutting off each other, always timing their play together.

Coop had an impressive study of Wright’s pick and roll mastery, but it’s worth mentioning that Wright works best with another playmaker as his partner. Watch this dream sequence with Van Vleet:


Van Vleet knows where Wright is angling his drive – a tough thing to predict – and drifts up from the corner to open a passing lane. After Van Vleet gives it up on his drive, he floats again to an open spot behind the arc, where Wright finds him for an open triple. The improvisational driving, cutting, and passing between the two is seamless.

Having a back-up who can make your team better when he checks into the game is an invaluable asset. Just check out Manu Ginobili’s career accolades. Or look at Patrick Patterson’s vaunted +/- over the last few years with the Raptors.

Delon Wright has thus far offered the Raptors that boost off the bench. He’s even garnered praise from Zach Lowe: a sure sign of success in the NBA. Even though Wright has played few minutes in this young season, his success is no mirage. He is ready to contribute.

Jakob Poeltl and OG Anunoby are the real deal, too. They both have advanced senses of timing on both sides of the ball. Expect their positive numbers in the early season to continue, though likely not lead the NBA. But be careful with some of Poeltl’s numbers; he’s averaging 2.0 blocks per game, after only 0.4 last season. Those will come down (although his highlight, at-the-rim rejections are fun).

However, some weaknesses can be gleaned from the early season as well. Rebounding is an issue. Even though Lucas Nogueira has a +13.5 net rating on the young season, that means little. He was exposed against the Spurs, who outrebounded the Raptors by 21. Neither Bebe nor Serge Ibaka could contain LaMarcus Aldridge on the block, and the Spurs’ physicality dominated the Raptors in the paint.

Per advanced stats, Bebe corralled more than 30% of rebounds on the defensive end against San Antonio: incredible numbers. On the season, he even leads the team in defensive rebounding percentage! This misses how poorly his team rebounds when he plays. The Raps grabbed fewer than a third of the available rebounds when Nogueira was on the floor against San Antonio. This means Bebe either doesn’t block out, or he had no help from his teammates, or both. It’s likely both. Either way, he was helpless against the supersized Spurs. Point guard Dejounte Murray racked up 15 rebounds against the Raptors, and those extra possessions overcame a massive turnover advantage for the Raps to help give the Spurs the win.

So does that mean Nogueira is unplayable? No. Yet more context provides some depth. Against teams with low post scoring threats, such as Aldridge, he will struggle on defense. To compound that issue, Serge Ibaka replicates his strengths – help side blocks – but also his weaknesses – poor rebounding and on-man post defense. Without Jonas Valanciunas available against the Spurs, the Raps had no option against Aldridge beyond Poeltl, who couldn’t play 48 minutes. So Nogueira has some issues, which makes him a weapon ideally to be used in only the right situations.

Another statistic that means something through 3 games: Pascal Siakam has only averaged 6.7 minutes, all coming in garbage time. The rest of his numbers can be thrown out the window (although it is tempting to read into his negative net rating, the only one on the team less Alfonzo McKinnie). On a team that is only willing to play 10 men, with a bench unit thriving with Anunoby at power forward, Siakam seems to fit poorly. That he also cannot control the defensive glass contributes to his low minute total. I wrote in his player preview that he can contribute by improving his handle, passing, and corner 3s. You know who already offers all of those things? Anunoby. It’s tough to find Siakam minutes, which is one more reason it’s nice to have the 905 in play.

In general, the Raptors could use more positive rebounders at every rotation spot except for Valanciunas. Having guards like Terry Rozier or Patrick Beverly help because they are excellent defensive rebounders, able to limit the opponent to just one shot. The Raptors lack that when Valanciunas sits. Wright is usually an excellent rebounder, but he looked less engaged against San Antonio. Hopefully that was just a one game blip.

So how can you read small samples sizes? Watching the games helps. Golden State wasn’t in trouble at 1-2, and the Raps aren’t in trouble after a close loss brings them to 2-1. When the two teams face off tonight, there are some things you should look for to either confirm or deny the numbers.

The Raptors have not shot well as a team (32.7% from 3), but which players will improve their percentages? Ibaka and C.J. Miles are both shooting above 45% from 3. That should drop. Lowry and Norman Powell are both just above (or at) 30% on the season, so they should improve slightly. Van Vleet might knock in a few more. But the team is shooting just about where you might expect when guys like Wright and Anunoby are taking multiple attempts a game. The small sample sizes of this season concur with trends from past years.

Can the Raptors defense continue so excellently? Despite being below average for most of last year (and only picking up after the acquisitions of P.J. Tucker and Ibaka), the Raptors have had a defensive rating in 2017-18 of 92.7, good for third best in the league. But who is their best defender? 2nd year Jakob Poeltl? Lengthy Delon Wright? Lowry gets beaten easily at the point of attack, and DeRozan has never been a lock-down wing. Can a team without an elite defensive player maintain an elite defense? The Warriors will be a good test.

Watch for those and other trend lines to continue. If the Raptors are able to cobble together a stellar defense, offense shouldn’t be a problem for a team with Lowry and DeRozan. In that case, expect the Raptors’ winning ways to continue against teams better than the Bulls or 76ers, especially given their bench dominance.

One small-sample-size statistic is sure to carry on: Bruno Caboclo leads the team after three games in net rating at 67.1, per That number has nowhere to go but up.

Leave a Reply

Your email address will not be published.