Writers of Pro Football Prospectus 2008

14 Aug 2012

The Effect of Circadian Rhythms on Monday Night Football

UPDATE August 15: Hey, so the internet is vast, and things often get overlooked. Turns out our buddy Scott Kacsmar (@CaptainComeback) already did the update you see below back in April, albeit only looking at win-loss records. He also summarizes straight-up results for the other non-Sunday-afternoon games. I'm a stickler for giving proper credit to previous work, so check out his piece if you're interested. As far as I know, though, this XP is the only place with an update of winning percentages against the spread.

Calling all gambling degenerates! Sleep science says that, when a west coast (i.e., Pacific time zone) team plays an east coast (i.e., Eastern time zone) team on Monday Night Football, the former benefits from humanity's late-afternoon energy boost while the latter is searching for a soft place to lay their heads (no tuck, please). Therefore, you should pick the west coast teams. And it turns out -- this being science and all -- west coast teams do in fact win straight up 63.5% of the time by an average of 14.1 points, and beat the Vegas spread 67.9% of the time. Well, at least that's what sleep science said in 1997 during a conversation about the NFL circa 1970-1994. Sounds like we need an update.

Even though the pre-2000 internet was mostly Lycos, I was able to cobble together MNF betting lines from various sites going all the way back to 1995. Here are the 21 east-west MNF matchups since then (winners against the spread in bold):

09/25/95 -- San Francisco (-10.5) @ Detroit. Final score: Lions 27, 49ers 24.
11/20/95 -- Miami (-1.5) vs. San Francisco. Final score: 49ers 44, Dolphins 20.
11/11/96 -- San Diego (-4) vs. Detroit. Final score: Chargers 27, Lions 21.
12/02/96 -- San Francisco (-10.5) @ Atlanta. Final score: 49ers 34, Falcons 10.
12/23/96 -- San Francisco (-9.5) vs. Detroit. Final score: 49ers 24, Lions 14.
09/29/97 -- San Francisco (-3.5) @ Carolina. Final score: 49ers 34, Panthers 21.
11/10/97 -- San Francisco (-4) @ Philadelphia. Final score: 49ers 24, Eagles 12.
09/14/98 -- San Francisco (-6) @ Washington. Final score: 49ers 45, Redskins 10.
11/30/98 -- San Francisco (-13) vs. New York. Final score: 49ers 31, Giants 7.
12/14/98 -- San Francisco (-10) vs. Detroit. Final score: 49ers 35, Lions 13.
01/03/00 -- Atlanta (-6.5) vs. San Francisco. Final score: Falcons 35, 49ers 29.
10/01/01 -- New York (-3) vs. San Francisco. Final score: 49ers 19, Jets 17.
11/25/02 -- San Francisco (-7) vs. Philadelphia. Final score: Eagles 38, 49ers 17.
12/02/02 -- Oakland (-7) vs. New York. Final score: Raiders 26, Jets 20.
10/27/03 -- Miami (-3.5) @ San Diego. Final score: Dolphins 26, Chargers 10.
11/17/03 -- San Francisco (-4) vs. Pittsburgh. Final score: 49ers 30, Steelers 14.
10/10/05 -- San Diego (-3) vs. Pittsburgh. Final score: Steelers 24, Chargers 22.
12/05/05 -- Seattle (-4) @ Philadelphia. Final score: Seahawks 42, Eagles 0.
09/22/08 -- San Diego (-8) vs. New York. Final score: Chargers 48, Jets 29.
12/05/11 -- San Diego (-3) @ Jacksonville. Final score: Chargers 38, Jaguars 14.
12/19/11 -- San Francisco (-1) vs. Pittsburgh. Final score: 49ers 20, Steelers 3.

Clearly, the MNF trend continues to this day: West coast teams over the past 17 years had a 16-5 record against east coast teams, with an average victory of 18.5 points (32.6-14.1) and a 76.2% winning percentage against the spread. They're currently on a four-game winning streak of the 37.0-11.5 variety.

Of course, we've got a few complications. First, my god, I had forgotten how much of a juggernaut San Francisco used to be on Monday nights. From December of 1996 through the end of 2001, they single-handedly carried the torch for west coast teams, covering nine straight against the spread. And in the last four years of their dynasty (1995-1998), San Francisco was football's equivalent of Death Row Records, piling up eight victims by an average margin of 20.5 points (33.9-13.4).

So, in the interest of fairness, let's set Shug Seifert, Mooch Dogg, and company aside for a moment. How have west coast teams fared since San Francisco dominated the east-coast-west-coast rivalry? In the subsequent 14 years, they've posted a 63.6% winning percentage against the spread. Still well into the black in terms of profit.

The other complications are what the study's authors controlled for in their analysis: relative team strength and game venue. I know this isn't technically ideal, but we'll use the line as a measure of the former and not bother with control groups. (I only have so much free time on my hands.) Here's how the 21 results look when split out according to those two factors:

West coast home favorites: 7-3
West coast road favorites: 6-1
East coast home favorites: 0-3
East coast road favorites: 1-0

Over the past 17 seasons, west coast favorites have been equally successful at home and on the road, and they've covered the spread more often than east coasters of similar quality (13-4 vs. 1-3). Granted, these are small groups with a heavy Death Row influence (like D12), but I don't think we can ignore the results when they conform so nicely to a previous test on an independent sample.

Whether this is really a sleep effect or some other variable explains the trend is another matter entirely. Correlation? Sure seems like it. Causation? That's for the academics to figure out. For instance, back when I dated the cousin of a former NFL offensive coordinator who shall remain nameless, said assistant regaled me with stories of caffeine binges in the locker room before games. A pot of coffee per player has to be a huge middle finger to circadian rhythms, right? (Although, admittedly, his team's results might not lend much credence to my anecdotal theory.)

Now, for the bad news, which some of you probably realized already. Although it's just a simple byproduct of geography and the whims of NFL scheduling gods, there have been only 21 qualifying matchups in the past 17 years. And if you include the study's sample of MNF games from 1970 to 1994, it's 84 qualifiers in 42 years. In other words, as much of a cash cow as this trend appears to be, NFL bettors can only take advantage of it about twice per season. To boot, you'll have to wait until (at least) 2013 because no games qualify in 2012. It's standard to judge betting systems on a long-term basis, but this is one requires a geological time scale.

With the proliferation of night games (read: advertising money grabs) over the past two decades, perhaps the MNF trend extends to other days of the week. If so, it would make for a much more useful system. Given the potential profit margin, exploring this idea sounds like a worthy mission should you choose to accept it.

Here's a link to the actual study cited in the Deadspin article:

Smith, R.S., Guilleminault, C., & Efron, B. (1997). Circadian Rhythms and Enhanced Athletic Performance in the National Football League. Sleep, 20(5), 362-365.

Posted by: Danny Tuccitto on 14 Aug 2012

31 comments, Last at 17 Aug 2012, 6:42pm by JJohnson

Comments

1
by TB (not verified) :: Wed, 08/15/2012 - 12:10am

If it's all about sleep, shouldn't all night games qualify?

2
by JohnD (not verified) :: Wed, 08/15/2012 - 12:17am

Quoting from the article:

With the proliferation of night games (read: advertising money grabs) over the past 20 years, perhaps the MNF trend extends to other days of the week. Given the potential profit margin, it sounds like a worthy mission, should you decide to accept it.

3
by bmurphyuk@mac.com :: Wed, 08/15/2012 - 7:57am

Week 1 Pitt @ Den
Week 2 Det @ SF
Week 3 Pitt @ Oak
Week 15 SF @ NE

4
by Revenge of the NURBS (not verified) :: Wed, 08/15/2012 - 8:32am

None of those meet the criteria the article is using.

I believe Denver is in the Mountain time zone, not the Pacific. Steelers/Raiders is a 4:00 game.

Det/SF and SF/NE do fit the timezone and night time criteria, but it's not Monday night. That said, if there's any truth to this theory, it should probably extend to other nights.

Other random thoughts --
1) For all the appearances of the 49ers on this list, I would have thought a lot of the matchups would be old NFC West matchups. There are a few, but most are non-divisional matchups against other marquee teams.
2) Monday night ain't what it used to be. Since the switch to ESPN, the types of games that used to go on Monday now go on Sunday night. Chargers/Jaguars?! Not sure if it would have any affect on this phenomenon, but it may be worth noting that there's more of a chance of a mismatch on MNF now than there was in the old days.

6
by Podge (not verified) :: Wed, 08/15/2012 - 8:56am

Some thoughts on your thoughts:

1) Isn't the definition "NFC West" a factor here? I mean, teams in the NFC West are unlikely to be based on the east coast. To get on the list, they need to be against a team on the east coast. I'm sure the NFC West used to live on MNF back in the day (oh how the times have changed!), but it wouldn't be on this list.

2) I think that's partly to do with paucity of good teams on the west coast. Over the last 5 years or so, its generally been that at best one of the west coast teams has been good (usually the Chargers), and the other three have ranged from mediocre to terrible. Then if you bear in mind the bit in point 1 about geography, they can only be scheduled against teams that would make this list in non-division play.

Also, SNF has flex scheduling, so can move things around to have more compelling games late in the year.

9
by Revenge of the NURBS (not verified) :: Wed, 08/15/2012 - 9:44am

"I mean, teams in the NFC West are unlikely to be based on the east coast."

Perhaps I should have put more emphasis on the "old" when I said "old NFC West". Pre-2002, the NFC West was a bit geographically challenged.

11
by Podge (not verified) :: Wed, 08/15/2012 - 9:49am

Ah yes, I see what you mean. I only started following in about 2001-ish, so I was a somewhat unaware. But even of those games, only the games against the Falcons and the Panthers would have counted for this study.

I'm glad that they've kept the spirit of the NFC West alive with the AFC South and the NFC East.

26
by Shattenjager :: Wed, 08/15/2012 - 6:49pm

Just since you prefaced it with "I believe," Denver is indeed in that most ignored of all time zones, the Mountain time zone.

5
by Podge (not verified) :: Wed, 08/15/2012 - 8:44am

The other night games I can find from the last couple of seasons that would apply (both a west coast team going east and an east coast team going west, as the effect seems fairly strong both ways) are below. I've treated the West Coast teams as San Fran, Oakland (although they had no qualifying games), San Diego and Seattle.

Baltimore 16, San Francisco 6 (week 12, 2011) (0-1 [not including spread])
Seattle 34, Philadelphia 14 (week 13, 2011) (1-1)
San Diego 38, Jacksonville 14 (week 13, 2011) (2-1)
San Diego 34, Baltimore 14 (week 15, 2011) (3-1)
San Francisco 20, Pittsburgh 3 (week 15, 2011) (4-1)

Philadelphia 27, San Francisco 24 (week 5 2010) (4-2, although I suspect SF beat the spread)
San Diego 36, Indianapolis 14 (week 12 2010) (5-2)
Tampa Bay 38, Seattle 15 (week 16 2010) (5-3)

San Francisco 10, Chicago 6 (week 10 2009) (6-3)
San Diego 42, Tennessee 17 (week 16 2009) (7-3)

I've excluded playoffs because I couldn't remember the arrangements for timings, and couldn't be bothered looking them up!

My geography might be a bit off on timezones. I'm from England, and I couldn't be bothered looking up whether certain places were in East Coast time or whatever the place next to it is. I also had to assume which games were night games based off the regular scheduling (Thursday games except Thanksgiving (only the last one being night), the last listed Sunday game, all Monday games), so I may have added one or missed one.

I think the conclusion is that it does probably hold. I'm guessing that it might even be as good as 9-1 over the last 3 years, as the only 2 games that look close enough for the spread to be different from the outright winner is the two SF games, against Philly and Chicago. I'd imagine the spread there was SF+ points, given that both Philly and Chicago have been at least decent, and SF have been pretty dreadful. Even if its 7-3 its worth keeping an eye on.

The problem is that it doesn't happen often enough to let you quit your job and live on the earnings. You could extend it to the Texas teams and their ilk, but the effect doesn't (at first glance as I was going through the schedules) seem as strong.

For 2012, the games currently scheduled that fit into this are:

Detroit @ San Fran, SNF Week 2.
Green Bay @ Seattle, MNF Week 3.
San Fran @ New England, SNF Week 15.
San Diego @ Jets, SNF Week 16.

7
by Aaron Brooks Go... :: Wed, 08/15/2012 - 9:26am

Chicago and Nashville (Tennessee) are in Central Time. Indeed, if you at all cognizant of western hemisphere time zones, Chicago is an easy one -- it's the example of a central time zone city.

Although it hasn't happened yet, an east coast MNF game involving the Cardinals after November 4 would also qualify.

8
by Podge (not verified) :: Wed, 08/15/2012 - 9:42am

I don't tend to look at timezones that often. Central is the one over from East Coast, isn't it? So I guess the effect of that would be less strong than for the proper East Coast ones, and potentially similar to an East Coast team playing someone like Denver. Presumably Detroit and Indy are also in the central zone?

That seems to basically highlight even further that while the effect does appear to be there, chances to apply it also appear to be even rarer than even initially thought.

An MNF game with the Cardinals playing, late in the season? You jest, surely. I'm curious why that's the exception. Does Arizona not do daylight savings or something?

12
by jebmak :: Wed, 08/15/2012 - 9:57am

Nope Detriot and Indy are in the EST. Over half of the teams in the NFL are.

(Probably another reason for the perceived 'east coast bias' now that I think about it.)

13
by Revenge of the NURBS (not verified) :: Wed, 08/15/2012 - 9:58am

"Presumably Detroit and Indy are also in the central zone?"

Indianapolis is on Eastern time. And that brings up another wrinkle in this theory -- time zones are very wide. Technically, Indianapolis and New York are in the same time zone, but a quick internet search shows that the sun rose in Indianapolis 51 minutes later today than it did in New York. In terms of the sunrise/sunset cycle, Indianapolis is much closer to Chicago. If this is supposed to be about natural bio-rhythm stuff, that probably matters.

14
by Revenge of the NURBS (not verified) :: Wed, 08/15/2012 - 10:06am

And yes, Arizona does not do daylight savings. That used to be true of Indiana as well, but not anymore.

15
by Aaron Brooks Go... :: Wed, 08/15/2012 - 11:32am

The history of the answer to the question of "what time is it in Indiana" is long, involved, and boring. Suffice to say, Indiana used to have three time zones in four separate pockets, which was even more confusing than it sounded, especially if you were traveling due south.

The line between Central and East basically runs south from the bottom of Lake Michigan, although it wanders back and forth as local politics and geography dictate. As much as Indianapolis is far west in Eastern, Chicago is extremely eastern for Central time. Amarillo, Texas and the outskirts of El Paso are also in Central time.

As for Detroit, Jacksonville, Florida is only about 50 miles further east than Detroit. Tampa and Detroit are about the same, and Atlanta actually lies to the west of Detroit. It's easy to forget how far east the northern eastern seaboard curves. The eastern tip of Maine really should be in Atlantic time.

21
by CBPodge :: Wed, 08/15/2012 - 4:05pm

You're not lying when you say its boring. I tried to read that. I really tried. I remember that episode of The West Wing with different time zones and hijinks ensued, so I'm at least partially interested.

And I still couldn't get past the first paragraph.

24
by Danny Tuccitto :: Wed, 08/15/2012 - 5:17pm

I like to pride myself on an above-average knowledge of geography (which isn't saying much these days), but I had no clue about this Indianapolis stuff. As a Niner fan, knew about Arizona's quirkiness, and that Detroit was (for a while) technically San Francisco's only Eastern time zone victory in ages. Indy, though, no clue.

After reading this thread, I went back and checked to see if there were any Colts MNF games that I missed out of ignorance. Luckily there were none, so what I wrote up still holds. Whew!

Thanks for droppin' knowledge y'all.

28
by DisplacedPackerFan :: Thu, 08/16/2012 - 9:15am

http://www.sportmapworld.com/map/american-football/usa/nfl/

Is a neat resource that can help illustrate some of your points. It has all the NFL stadiums and an interface that will get you the distances between (air distances based on lat/long, driving via google maps).

18
by Travis :: Wed, 08/15/2012 - 2:07pm

It works the other way - Arizona is 3 hours behind the East during summer months, but 2 hours behind in the winter.

10
by Podge (not verified) :: Wed, 08/15/2012 - 9:46am

I wonder if there's a similar effect for west coast teams playing early games on the east coast? A 1pm kickoff in New York would be the equivalent of 9am kick off for San Diego.

I'm sure I've read that one of the west coast teams is dreadful in those games, but that might be because west coast teams are often dreadful in most games.

16
by 0tarin :: Wed, 08/15/2012 - 11:39am

It's actually the equivalent of a 10am kickoff, but I'd still be curious to see if the effect carries.

17
by Joseph :: Wed, 08/15/2012 - 1:10pm

According to research that I've read, it does--in reverse of the article. In other words, west coast teams tend to do horrible in 1 PM ET starts--although, as is mentioned, west coast teams have been generally bad in general. As far as ATS #'s, I couldn't tell you--but I'd bet they're still pretty bad.

22
by CBPodge :: Wed, 08/15/2012 - 4:07pm

This is good, because now I can call my gambling "research", and feel somewhat less like a terrible person.

19
by Travis :: Wed, 08/15/2012 - 2:23pm

FWIW, the effect hasn't carried over to other night games.

Since 1985 (the furthest back the covers.com database goes), West Coast teams are 22-22 against the spread on Thursday/Saturday/Sunday nights against East Coast teams.

Notes:
West Coast teams included Arizona during Daylight Savings Time.
East Coast teams included Indianapolis 2006-2011 and 1985-2005 not during Daylight Savings Time.
The 1993 Falcons-L.A. Rams Thursday nighter has no spread listed.

20
by Travis :: Wed, 08/15/2012 - 3:07pm

The games:

12/8/1985 - Pittsburgh at San Diego (-5.5). Final score: Chargers 54, Steelers 44.
11/29/1987 - Cleveland at San Francisco (-3). Final score: 49ers 38, Browns 24.
9/30/1990 - Washington at Arizona (+6). Final score: Redskins 38, Cardinals 10.
10/7/1990 - L.A. Raiders at Buffalo (-3). Final score: Bills 38, Raiders 24.
9/13/1992 - Philadelphia at Arizona (+7). Final score: Eagles 31, Cardinals 14.
12/6/1992 - L.A. Rams at Tampa Bay (-1). Final score: Rams 31, Bucs 27.
10/14/1993 - L.A. Rams at Atlanta (no line). Final score: Falcons 30, Rams 24.
12/5/1993 - Cincinnati at San Francisco (-24). Final score: 49ers 21, Bengals 8.
9/11/1994 - N.Y. Giants at Arizona (-3.5). Final score: Giants 20, Cardinals 17.
9/10/1995 - Philadelphia at Arizona (-3.5). Final score: Eagles 31, Cardinals 19.
10/1/1995 - Oakland at N.Y. Jets (+7.5). Final score: Raiders 47, Jets 10.
11/5/1995 - Miami at San Diego (-1). Final score: Dolphins 24, Chargers 14.
12/3/1995 - Buffalo at San Francisco (-14). Final score: 49ers 27, Bills 17.
9/8/1996 - Miami at Arizona (+6). Final score: Dolphins 38, Cardinals 10.
12/1/1996 - New England at San Diego (-1). Final score: Patriots 45, Chargers 7.
12/15/1996 - Seattle at Jacksonville (-6). Final score: Jaguars 20, Seahawks 13.
9/20/1998 - Philadelphia at Arizona (-5.5). Final score: Cardinals 17, Eagles 3.
10/17/1999 - Washington at Arizona (+4). Final score: Redskins 24, Cardinals 10.
12/26/1999 - Washington at San Francisco (+7). Final score: Redskins 26, 49ers 20.
12/10/2000 - N.Y. Jets at Oakland (-3.5). Final score: Raiders 31, Jets 7.
12/23/2000 - Buffalo at Seattle (-3). Final score: Bills 42, Seahawks 23.
10/7/2001 - Carolina at San Francisco (-7). Final score: 49ers 24, Panthers 14.
12/2/2001 - Buffalo at San Francisco (-8.5). Final score: 49ers 35, Bills 0.
9/5/2002 - San Francisco at N.Y. Giants (+4). Final score: 49ers 16, Giants 13.
9/15/2002 - Oakland at Pittsburgh (-3.5). Final score: Raiders 30, Steelers 17.
11/17/2002 - New England at Oakland (-5). Final score: Raiders 27, Patriots 20.
9/26/2004 - Tampa Bay at Oakland (-4). Final score: Raiders 30, Bucs 20.
9/8/2005 - Oakland at New England (-7.5). Final score: Patriots 30, Raiders 20.
9/25/2005 - N.Y. Giants at San Diego (-7). Final score: Chargers 45, Giants 23.
12/31/2005 - N.Y. Giants at Oakland (+7.5). Final score: Giants 30, Raiders 21.
10/8/2006 - Pittsburgh at San Diego (-3.5). Final score: Chargers 23, Steelers 13.
9/16/2007 - San Diego at New England (-3.5). Final score: Patriots 38, Chargers 14.
11/11/2007 - Indianapolis at San Diego (+3.5). Final score: Chargers 23, Colts 21.
12/15/2007 - Cincinnati at San Francisco (+8.5). Final score: 49ers 20, Bengals 13.
10/12/2008 - New England at San Diego (-6). Final score: Chargers 30, Patriots 10.
10/19/2008 - Seattle at Tampa Bay (-10.5). Final score: Buccaneers 20, Seahawks 10.
11/23/2008 - Indianapolis at San Diego (-3). Final score: Colts 23, Chargers 20.
9/27/2009 - Indianapolis at Arizona (-3). Final score: Colts 31, Cardinals 10.
10/4/2009 - San Diego at Pittsburgh (-5.5). Final score: Steelers 38, Chargers 28.
10/25/2009 - Arizona at N.Y. Giants (-9). Final score: Cardinals 24, Giants 17.
10/10/2010 - Philadelphia at San Francisco (-3). Final score: Eagles 27, 49ers 24.
11/28/2010 - San Diego at Indianapolis (-2). Final score: Chargers 36, Colts 14.
11/24/2011 - San Francisco at Baltimore (-3.5). Final score: Ravens 16, 49ers 6.
12/1/2011 - Philadelphia at Seattle (+3). Final score: Seahawks 31, Eagles 14.
12/18/2011 - Baltimore at San Diego (+1). Final score: Chargers 34, Ravens 14.

23
by Dean :: Wed, 08/15/2012 - 4:43pm

Should have known that if anybody'd come through for us on this, it'd be you, Travis. Well done.

25
by Danny Tuccitto :: Wed, 08/15/2012 - 6:05pm

Excellent. Thanks, Travis.

So the question becomes, why is MNF different? Is this damning evidence against the biorhythm explanation or is there still a way to shoe-horn the other-night results into the hypothesis?

My reflex is to say, "MNF is probably just noise," except that the sample is large enough, and the statistical controls strong enough to suggest there's something to the trend.

27
by jebmak :: Thu, 08/16/2012 - 12:14am

That 10 minutes between 8:20pm and 8:30pm EST makes all the difference...

Maybe I'm misremembering, but didn't MNF used to start later by an hour or so? That could be something.

29
by PatsFan :: Thu, 08/16/2012 - 9:57am

You're not misremembering. For a long, long, long time MNF games started at 21:00 ET.

30
by Anonymousjk (not verified) :: Thu, 08/16/2012 - 4:24pm

I'd wonder about the influence of being a MNF game on the spreads themselves that are being used to determine this.

31
by JJohnson (not verified) :: Fri, 08/17/2012 - 6:42pm

The Dolphins v Chargers game in 03 was playing in Arizona due to the wild fires in southern California that year.