Finding the best positions for Boston's most versatile players, based on last season's data

Brad Stevens will have an endless amount of lineup combinations to play around with this season. There are so many different pieces that can fit in multiple places. Other than Isaiah Thomas and Al Horford, whose positions should remain constant regardless of who is sharing the floor with them, Stevens will be facing numerous decisions as to who to play where.

Golden State has set the pace for small-ball lineups, and the rest of the league has been taking notice. Playing Marcus Smart at shooting guard? Avery Bradley at small forward? Jae Crowder at the four? Could these Celtics play one spot above their typical position in a small-ball lineup?

It isn’t easy to differentiate who excels at which spot on the court, but thanks to statistics from, it becomes a bit more clear. For the sake of this topic, we will be focusing on Boston's five most versatile players who can be used at multiple positions interchangeably. Let’s take a look at what their 2015-16 Player Production by Position calculator spit out for these five Celtics:

Before we digest these numbers, it's important to remember two key factors that could skew this data.

First, some of these results may not accurately reflect a player's ability to play a position because of a small sample size. Avery Bradley's net rating was three points higher when he was considered a small forward instead of shooting guard. However, he only saw action from the three spot 12.6% of the time, meaning that the sample size isn't substantial and may not perfectly portray his abilities as a small forward. With this in mind, here is how much time was spent in each position last season:

The second factor to consider is differing lineups that are played together--something that these statistics do not measure. For example, perhaps Jonas Jerebko was mostly used at small forward when the second or third units were playing with him, and used at power forward when playing alongside the starters. This would have a huge effect on his net ratings, regardless of how he performed.

Considering how evenly distributed Kelly Olynyk and Smart's playing time was for their two respective positions, they are perfect examples.

Kelly Olynyk

Olynyk's 4.4 net rating increase at the power forward spot is backed up by his shooting--His effective field goal and true shooting percentages jumped up by 7% when playing the four instead of the five, suggesting that he is more effective from this spot in the lineup. This could be because he isn't worrying about the interior game with a center playing by his side, letting him focus on doing what he does best and stretching the floor.

In terms of the debated final starting lineup spot, maybe this indicates that Olynyk is the best fit next to Horford (once Olynyk is healthy, that is). To put it into perspective, Amir Johnson was considered a center 97.5% of his time on the court last season, so he is accustomed to playing that position. And that isn't to say Johnson would be worse as a power forward, but that Olynyk wouldn't be as valuable as a center coming off the bench.

Marcus Smart

I believe Smart's results are attributed to the second factor of differing lineups. If you recall from earlier, his net rating was only 0.5 as a point guard, but rose to 5.8 when he played a spot up. However, a lot of his time at shooting guard was spent paired with Isaiah Thomas in the backcourt and with many other starters, which threw off the results.

This is evident when looking at his shooting and overall production from each of the two spots--not only did Smart shoot a hair worse as the two, he also averaged fewer points and assists compared to when he was point guard. In other words, Smart as an individual played better at the point, but his team played better when he was at shooting guard.

Despite lopsided minutes, these statistics are still notable for the remaining players.

Jae Crowder

Crowder's negative net rating from the power forward position is probably jumping out to most people. This is a cause for concern given that he has a decent sample size from last year after logging 505 minutes and is the essence of the Celtics' small-ball lineup. Sliding Crowder up to the four spot due to his defensive versatility is what allows Coach Stevens to go small, yet the team doesn't seem to do that well with him there.

Have no fear. While the negative net rating is a bit of a mystery, Crowder actually excels as an undersized power forward. also displays a Daily RAPM estimate (DRE) stat, which is a box score plus-minus measure and represents an estimate of net points contributed, and Crowder recorded the top DRE on the team in 2015-16 from the four spot. In addition, virtually all of his traditional stats saw a slight bump at power forward compared to small forward.

Avery Bradley

When Bradley is playing the three, the Celtics tend to do well, albeit it only occurred for 202 total minutes last season. We can't take too much from this (his stats didn't change too much between shooting guard and small forward), but it seems like Bradley is capable of sliding up a spot for small-ball lineup combinations.

Jonas Jerebko

Jerebko was clearly most comfortable at the power forward position last year. In the 17.4% of his time as a small forward, his efficiency and production dropped off. The Swedish Larry Bird should not have to worry about slipping down a position anymore with Jaylen Brown and Gerald Green coming off the bench.

On top of these versatile athletes from last year, Jaylen Brown will be another piece that Coach Stevens can toy around with in multiple spots throughout the 2015-16 campaign. It will be fun to see how Stevens solves the complex puzzle of maximizing his lineups.

Follow Erik on Twitter: @erikjohnson32

Photo of Smart, Crowder, and Bradley by AP Photo/Michael Dwyer