Many bloggers and tweeters around the NFL have started to notice that Home Field advantage in the NFL this year has seemed awfully large. Home teams won 14 of 16 games last week, and depending on which spreads you use, home teams have covered at nearly a 60% clip. Jason Lisk over at The Big Lead has a story up about the possibility that the replacement officials are causing the issue. The Associated Press, in the same vein, has a piece up titled “NFL replacement officials affecting Vegas bets.”

They both paint a picture of replacement officials, overwhelmed by the crowd noise, making more favorable calls for the home team. Lisk writes:

One of the areas where we might surmise that inexperienced officials might be affected is by crowd influence. This could be particularly true for those coming from situations with much smaller home fields, and we know that these officials by and large are not coming from the FBS level of collegiate officiating.

This is a pretty interesting, and definitely sounds plausible. In fact, this was the basis for many of my Week 1 and Week 2 picks.

Thanks to Toby Moscowitz and Jon Wertheim, authors of the book Scorecasting, we have a pretty good idea that officiating bias can be an incredibly strong effect in producing home field advantage. Moscowtiz and Wertheim, using data across many different sports, found strong evidence that home field advantage does not derive from a travel effect, scheduling bias, or simply better home team performance. Instead, they found that the effect was almost exclusively the result of officials making more and better calls in favor of the home team. In particular, they noted that the closer a crowd was to the field, the bigger the home field advantage became.

Assuming Moscowtiz and Wertheim are correct (a claim I’ll examine more in depth later), then the idea that the less experienced replacement referees will be more affected by crowd noise, and thus call an even more pro-home team biased game, seems to make sense. To wit, Mike Kania, at the always excellent, has noted that there have in fact been an extra four penalties per game this year over 2008 averages.

So what does this all mean? Should you rush to put all your money on the home teams before the books adjust (assuming they haven’t already?). There’s another possibility here however. Couldn’t all this just be noise? After all, while the NFL season does seem to be in full swing, we are just looking at 32 games. What did the data look like last year through 32 games?

It turns out, in terms of penalties, it was very similar. Last year, through 32 games, there were 203.5 penalties called per week, before tailing off for the rest of the season to a 193/week average. In 2010, the first two week average of 197 penalties contrasted strongly with the season average of 178.5. The first two week average in 2009 204, far in excess of the season average of 177.8. Finally, in 2008, there were 195 penalties per week the first two weeks, before dropping off to a seasonal average of 167.8.

So there seem to be two simultaneous trends going on here. First, there are more penalties called in Weeks 1 and 2 than over the season as a whole (by a significant amount). Second, there seems to be more penalties being called per season generally in the NFL, having gone from a figure in the high 160s in 2008 to the low 190s by last season. Between the general trendline towards more penalties in the NFL and the tendency for penalties to be overabundant in the opening weeks, there doesn’t seem to be all that much cause for concern with the impact of replacement refs so far.

How about those point spreads though? Well, so far this year, home teams have been favored by 2.56 points per game, and they’ve beaten their opponents by a whooping 5.9 points per game, or 3.34 points per game better than the spread. That’s a fairly incredible advantage. Any experienced gambler would tell you a better than three points per game edge is essentially free money. But how do those numbers compare to last? Last year, there was an almost identical effect, with home teams outperforming the spread by 3.06 points per game. In 2010? Larger still, home teams in the first two weeks beat the spread by 4.89 points per game. There was no similar effect in 2009 however, with home teams underperforming the spread by 3.04 points.

In fact, the current figure of 3.34 points per game is not unusual at all on a historic scale for the first two weeks of the season. That’s before getting into issues of multiple endpoints. There have even been 32 game stretches where home team have beaten or failed to beat the spread by 8 points per game, or better than a touchdown off of expectations.

So why didn’t we hear about it then? Sadly, I’d suggest marking this one down as confirmation bias so far. We all know there are new officials in the NFL, and we know officials play a major role in determining the outcome of games. Doesn’t it stand to reason that if something unusual is happening, they could be the cause? Perhaps. And if this trend continues, it may warrant further evaluation. But home teams vs. the spread is a number that fluctuates greatly, even without the presence of replacement officials. Getting overly excited by a 3.34 point per game variation is akin to a baseball fan getting excited by leaguewide BABIPs being up by a few points. It’s just not that unusual. For now, don’t go nuts with this.

Coming up: A look at the last time we had replacement officials in the NFL.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s