Freelance Friday: Home-Court Assist Biases

Oct 30, 2015; Phoenix, AZ, USA; Two-time NBA Most Valuable Player Steve Nash smiles during his induction to the Suns Ring of Honor at Talking Stick Resort Arena. Mandatory Credit: Jennifer Stewart-USA TODAY Sports
Oct 30, 2015; Phoenix, AZ, USA; Two-time NBA Most Valuable Player Steve Nash smiles during his induction to the Suns Ring of Honor at Talking Stick Resort Arena. Mandatory Credit: Jennifer Stewart-USA TODAY Sports /
facebooktwitterreddit
Freelance Friday 2
Freelance Friday 2 /

Freelance Friday is semi-regular series here on The Nylon Calculus where we accept solicitations from the broader basketball community. The intent is to give as wide a range of voices as possible a chance to be heard as well as to encourage new perspectives and experiments. Today’s feature concerning the dreaded “hometown assist” comes courtesy of Bryan Grosnick. Bryan is a baseball writer and analyst. He’s a contributor to Baseball Prospectus, former Managing Editor of Beyond the Box Score, and co-host of The Four-Man Rotation podcast.


Oct 30, 2015; Phoenix, AZ, USA; Two-time NBA Most Valuable Player Steve Nash smiles during his induction to the Suns Ring of Honor at Talking Stick Resort Arena. Mandatory Credit: Jennifer Stewart-USA TODAY Sports
Oct 30, 2015; Phoenix, AZ, USA; Two-time NBA Most Valuable Player Steve Nash smiles during his induction to the Suns Ring of Honor at Talking Stick Resort Arena. Mandatory Credit: Jennifer Stewart-USA TODAY Sports /

As a human being who lives and breathes sport statistics, I tend to trust numbers over the “eye test” because data is concrete. It has rules, and I can’t watch every game, every play, every moment of a sport. I rely on the stats to tell me what happened, and I count on them to be accurate. I think most people do.

This becomes problematic at times, because in a number of sports a recurrent theme emerges: just because a number is assigned to something, doesn’t make it an absolute truth. You can see it in all sorts of areas where an official scorer assigns a number to an event. In my primary sport of investigation, baseball, you can see it in how official scorers assign errors to fielders. But in basketball, there’s perhaps even more room for nuance in assigning credit for certain events.

A basketball fan, I became interested in assists and how they were assigned to the home team versus on the road.[1. Ed – a topic we’ve covered in broad detail here.] Figuring I could test how much “extra credit” home scorers give to their squads, I decided to run a simple study. Scouring the internet to see if had been done in the past, I found Ken Pomeroy’s work at Basketball Prospectus. Ken examined how home-court affected the assist statistics for different players and teams in college basketball, using Acie Law and Texas A&M as an example. All I had to do was crib the methodology and apply it to the NBA, a fairly simple (if time-consuming task). It was worth the time.

Basically, I started by looking at the simple assist rate (assists divided by field goals made) at home for each team during each season in my sample. For example, the 2010-2011 Los Angeles Clippers. The Clips made 1,525 field goals at home, and recorded 990 assists. That gave them a 64.9 percent simple assist rate–that’s pretty high! Props to Baron Davis, I guess.

Then I tabulated the assist rate on the road for the Clippers. The team’s assist numbers were quite a bit different, even though to total field goals made were about the same. The Clips made 1,490 field goals at home, but only recorded 823 assists. The result was a very different percentage than at home: a 55.2 percent road assist rate.

There was a difference of almost 10 percent between those two Clippers assist rates. In essence, the Clippers during that season recorded 17.6 percent more assists while playing at home than they did on the road.

Of course, there are a lot of reasons why a team may have a higher assist rate at home rather than on the road. The first that comes to mind is team success–NBA teams have more success (and wins) at home than the do on the road. It’d be fair to speculate that all teams would get a small assist rate bump at home, versus on the road.

Here’s a picture of that assist rate difference, between road and home, for each of the past 10 seasons.

FFHT0
FFHT0 /

In 2010-2011, the league-wide difference between home assist rate and road assist rate was approximately +2.2 percent at home. Over the last ten years, the overall rate has been about +2.4 percent, but that number has dropped a bit over the last four seasons.

The Clippers, and their +9.7 percent shift from road to home? That’s quite remarkable. And when you look at the patterns of how particular teams have had consistently high (or low) assist rates at home versus on the road, things become clear in a hurry.

FFHT1
FFHT1 /

During most seasons over the past 10 years, the Clippers have carried a +7.7 percent  to +10.8 percent bump to home assists against road assists. Their total +8.2 percent advantage in assists recorded at home is very large, and the tops in basketball over my sample.

If some teams have such a dramatic advantage in assists at home as opposed to on the road, you may ask if there are some teams that see their assist numbers drop at home on a regular basis as well. There certainly are.

FFHT2
FFHT2 /

Five of the 10 most assist-suppressing seasons came courtesy of the Miami Heat. More than any other team, Miami lost assists at home versus away over the last 10 seasons. It’s tough to imagine that the Heat, especially with their tremendous success at home particularly during the James-Wade-Bosh years, would have been potentially undercut in terms of assist totals while playing in Miami, yet it appears possible that this is the case. In every single season during the last 10, the Heat have had a smaller assist rate at home than on the road, giving them a -5.1 percent disadvantage on South Beach.

There’s tons of fun, if not eminently useful information that we can pull from these comparisons of road assist rates to home assist rates. For example, three of the top five home assist rate benefit seasons came in three consecutive seasons (’07-’08, ’08-’09, and ’09-’10) for the Denver Nuggets. There was a +13.0 percent difference during their 2008-2009 season, which amounts to the team putting up 27.8 percent more assists at home than they did on the road!

I can’t help then look at the assist totals on the year for players like Chauncey Billups, Anthony Carter, Allen Iverson, and Carmelo Anthony and think to myself this: is there any chance that their assist numbers weren’t quite as reflective of their actual performance, due to the way the home scorers doled out assists?

So let’s take this to the next level. Let’s do some quick math on a real-world scenario: Steve Nash’s 2007-2008 season. As noted in the last table, the ’07-’08 Suns were among the stingiest teams in our sample at giving out assists. The simple assist percentage at home was 60.8 percent, where the road offered a 68.3 percent assist rate. During that season, the league as a whole offered about a three percent bump in assists to players at home, versus the road. If we use the Suns’ road assist rate as a guide, we should apply that number to the team’s home assists, plus half of the league-wide home-court advantage that they should have received with scorekeepers free of arena-specific bias. That new home assist percentage would be roughly 69.8 percent.

In 2007-2008, Steve Nash personally had 415 assists at home in 41 games, good for a assists-per-game mark of about 10.1. If we jack up his assists to account for the percentage I just figured, we get something closer to 476 arena-adjusted assists on his home court. That’s a per-game mark of 11.6 arena-adjusted assists. That’s quite the difference: 1.5 assists per game!

None of this is intended to diminish or cast aspersions on the scorers who give out either more or fewer assists. Being a scorekeeper is hard–and subjective–work. But, if particular scorers tend to hand out assists on bias (conscious or no), then we should probably try to acknowledge that, and find a way to normalize those numbers.

I do think this puts in rather stark relief the difference between the numbers we take for granted and the amount of subjectivity that can come into play. Imagining, for example, that the Seven Seconds or Less Phoenix Suns–and Steve Nash in particular­–perhaps should have been credited with a greater number of assists is kind of neat. It maybe even gives us another data point in favor of Nash’s contested MVP candidacy. It also helps us identify which arenas might have a bit more of a long-term tendency, which can help in making back-of-the-napkin adjustments to payers’ expected assist totals.

Regardless, it’s interesting to know how big the swings can be between home scorekeepers. If you’ve ever wondered how much a difference a team’s scorekeeper might have on your team’s assist totals, well, now you know.


[Note: Blake Murphy of Raptors Republic and The Score wrote up some of my research back in 2013 at the now-defunct HoopData, when I only had about six years of assist numbers. Thanks to him for his assistance with this project.]