clock menu more-arrow no yes mobile

Filed under:

NCAA Basketball Box Score Analysis Dos and Don'ts

Though the mathematics of analyzing college basketball is increasing in complexity, the parlance used among fans and media has remained largely static. Here's some simple tips on how to read stats the right way.

No one appreciates Mike Scott as much as we do :(
No one appreciates Mike Scott as much as we do :(

For those interested in statistics, the box score is the most concise way to determine the answer to the simple question, "what happened in the game?" Fans and media commonly read and cite stats like points, rebounds, assists, and turnovers to explain who did well, who did poorly, and to create a narrative of the game. However, there are some pitfalls in using and comparing these stats haphazardly, as it's easy to draw conclusions that appear reasonable, but don't hold up statistically.

Here are a few simple dos and don'ts to keep in mind this season to impress your nerdy friends and not incite the wrath of the internet (See: Jason King). Hat tip to Burnt Orange Nation for the idea, as they had a great "stat geek" post last week. There are a TON of complex formulas out there to help us understand basketball, some of them highlighted in that post. What follows are more general guidelines that require no more than a glance and maybe some "back of the envelope" calculations to easily see through the numbers.

Do NOT ever cite, care about, or think about rebound margin

This article says it all. Mr. Teel (who usually does a fantastic job covering UVA), asserts that the Hoos' poor rebounding margin (52-32) caused a loss to UNC. To evaluate this statement, one must consider when a team gets a rebound - after a missed shot. Remember, defensive rebounds occur at a much higher rate than offensive rebounds. Thus, a team who misses lots of shots will likely also give their opponents lots of rebounds. Additionally, teams who never force turnovers, rarely commit fouls, or just play really fast will also have inflated margins. Therefore, while it is possible, and indeed likely, that an out-rebounded team did a poor job rebounding (in the case above...we got crushed), we are really just looking at a proxy for shooting. So overall, UVA's rebound margin didn't necessarily cause this loss - the loss likely affected a poor rebound margin.

Do use rebounding percentages

A better way of determining which team did a better job of rebounding is to use percentages, which normalize for number of missed shots. A team's defensive rebounding efficiency is equal to (# of defensive rebounds) / (number of opportunities), or (# of defensive rebounds) / (# of defensive rebounds + # of opponent's offensive rebounds). Offensive rebounding efficiency is equal to (# of offensive rebounds) / (# of offensive rebounds + # of opponent's defensive rebounds).

Note that team A's defensive rebound % is the same as 1 - team B's offensive rebound %. So, how bad did we do against UNC? The Hoos had a defensive rebounding efficiency of 56% (24/43), while UNC's was 81% (29/36). In other words, we rebounded 56% of UNC's misses, while they rebounded 81% of ours. Remember to consider strategy when thinking about rebounding too; Tony Bennett prefers to get back on defense rather than committing to the offensive boards, which will lead to higher opponent defensive rebounding efficiency.

Do NOT compare pace-affected stats across box scores

A common practice from last season's ACC Player of the Year race was to compare how Mike Scott and Tyler Zeller did in individual games against different teams. "Scott scored 16 points last night[...] while Zeller put up 23." Last season, UNC averaged 72.2 possessions per game, while UVA averaged 60.6. Thus, all of UNC's stats should automatically be about 15% higher. Ignoring this is PACISM, a terrible crime to commit.

Do think about offensive efficiency (Team)

A team's offensive efficiency is the number of points scored per possession. If Virginia scores 65 points in a 60 possession game, it's efficiency is 65/60 = 1.08. Conversely, defensive efficiency is points a team allows per possession. As a baseline, teams average scoring 1 point per possession (ppp). Some, such as Ken Pomeroy, measure this in terms of "per 100 possessions;" last season, the Hoos finished with offensive and defensive efficiencies of 102.9 and 87.7, respectively, adjusted for quality of opponent.

Thus, to determine how "good" or "bad" the team played, check out the number of possessions (sites like will show it), and compare to points scored/allowed. Last season, our best defensive efficiency our first game, against SC State (38 points allowed in 65 possessions); our worst was our final game, against Florida (71 in 63).

Do think about offensive efficiency (Player)

Player A scored 19 points. Player B scored 23. Who had a better game? In this case, Player A did. That was Mike Scott, who on February 8th scored 19 points on 9 for 9 shooting against Wake Forest. That same day, Tyler Zeller had 23 on 8 for 15 shooting (and gave up the game-winning shot) against Duke. Remember to look at how efficient each player was with the shots he took by checking FG% too.

Even better, look at effective FG%. That takes into account the fact that a 3-pointer is worth 50% more than a two and is calculated by (FG Made + .5* 3PT FG Made) / (FG Attempted). So, if Player A took four two-pointers and made three of them, he has an EFG% of 75% (3/4). If Player B did the same, but one of his made shots was a three-pointer, his EFG% is 87.5% (3 + 1(.5) / 4)).

Do NOT worry about game-by-game +/- ratios

I'll leave this to Ken Pomeroy - here is his treatise. The "Too long, didn't read" version - there's too much randomness for this to be meaningful at all.

Have questions? Or other handy-dandy shortcuts? Let's hear it in the comments.