By William Ezekowitz
Contrary to popular belief, NBA players are people like you and me. But unlike you and me, NBA players are constantly travelling across the country with, as Nik Stauskas will attest, a fair amount of spending money. This inevitably leads to what Jalen Rose has termed “champagning and campaigning.” In many podcasts with Dave Jacoby and Bill Simmons, Jalen has spoken at length about this phenomenon, and even discussed each city’s relative merits in the pursuit. This got me thinking: do players play significantly worse after a night on the town? And if they do, then wouldn’t this depend on the city in which they were playing? Does the nightlife of, say, Los Angeles give the Lakers and Clippers an unfair advantage, because visiting players are more likely to take advantage of their one night in such a fun city?
To investigate, I decided to look at the performance of away teams in Sunday afternoon games over the past 8 years. I gauged performance by whether or not the home team covered the spread. In theory, because betting lines are efficient, the spreads are an accurate measure of how games should play out, and the home team should cover 50% of the time. However, if the away team is not in proper physical shape due to a late night exploring their new city, they would play worse and this would be manifested in their failing to cover the spread. I picked Sunday afternoon games because they would be the ones most likely to exhibit the effect that I am looking for.
To start I decided to look at “party cities”, which I defined as Los Angeles, New York (and Brooklyn), Chicago, Miami, and New Orleans. There was no set method for determining which cities went into the “party” category, but to me these all seemed like fair guesses. I compiled the data using spreads from covers.com and game start time data from ESPN.com. I then used a binomial distribution test to see if the data I had found significantly strayed from the half and half expectation of the cover distribution between home and away. The data is below:
While the p value falls well short of the 0.05 needed to indicate official statistical significance, the value 0.32 indicates that there may be some slight effect. Undaunted, I decided to check if the effect held across all cities, not just the ones I arbitrarily designated as “party cities”. The results are as follows:
This lower p value indicates that across the NBA, away teams do perform worse than expected on Sunday afternoons. It is not low enough to be statistically significant at the 95% level, but is weakly significant at nearly the 10% level. It then occurred to me that perhaps I had used the wrong method for calculating which cities are worse for away teams to go to. So I ranked every city (provided it had a sufficient sample size) based on how often its home team covers the spread on Sunday afternoon game. Again, team quality should have no bearing here, as it should be accounted for in the spread anyway. The table, with a low p value indicating a higher percentage of home covers just like in the other tables, is listed below.
|City||P Value||Sample Size|
For reference, the bottom four teams covered less than 50% of the time. Also, the party cities of Chicago and New Orleans had between them 6 Sunday afternoon games, for whatever reason, so the data is not absolutely perfect. Furthermore, these are more general trends than rigid rankings, but it is useful to look at them.
So do NBA teams play worse on the road in party cities on Sunday afternoons? No. But they do play worse on the road on Sunday afternoons regardless of the city. I have only charted an effect and not accounted for any causes so one could conceivably draw many conclusions from the data. It could be that the effect is due to champagning and campaigning, and that any city is a party city when you’re in the NBA. Or maybe the effect is not caused by partying at all. Perhaps the players are not used to playing in the afternoon, and the road team feels the effects more than the home team. Ideally, the data would include variables like how many miles the team had travelled, what time zone they were in the night before, and how many games they had played recently, as all these would probably have effects.
The only conclusion one can make for sure is that if you are in the business of betting on NBA games, the smart money is on the home team on Sunday afternoons, last Sunday’s Clippers vs. Kings game notwithstanding, of course.