I’m a big fan of the Tuesday Morning Quarterback column by Gregg Easterbrook, now on ESPN’s Page 2. It’s fun to read and it’s full of interesting commentary on football, politics, and television. TMQ also goes on inspiring tirades against things like Direct TV’s monopoly on expansive NFL television coverage in the USA, the media’s habit of reporting long-term contracts with exorbitant dollar-figures that will never be paid, flaws in Star Trek plots, and the unfair advantage that Detroit and Dallas get from always playing at home on Thanksgiving.
This last tirade was the title of an article last year, Pass the turkeys out of Dallas, Detroit, and TMQ mentioned it again last week, arguing “As TMQ has written before and will surely write again, the league’s system of allowing Dallas and Detroit to host on Thanksgiving every year is one-sided. Hosting every Turkey Day both confers an annual half-game advantage on the Cowboys and the Lions in the standings, and totally screws whatever team is sent on the road with just three days’ rest.” In the article from last year, he supported this position with these stats: “Consider that since 1936, Detroit has won 46 percent overall, but 53 percent of games played at home on Thanksgiving. Since 1966, Dallas has won 60 percent overall, but 65 percent of games played at home on Thanksgiving. This works out to roughly a half-game annual bonus for the Cowboys and Lions, a windfall at the expense of all the league’s other teams.”
It seems plausible that the short rest and the holiday crowd in the opponent’s stadium would put the visiting team at a disadvantage. There’s just one problem with this view: the data don’t support it. For some reason, TMQ chose to compare Detroit & Dallas’s home Thanksgiving records with their overall records, instead of comparing them with their home records. After all, the Thanksgiving game replaces another home game for Detroit and Dallas, not just any game.
Carrying out my duty to fight ignorance wherever it is found, last fall I took the time to look up the home records of Dallas and Detroit. I found that Detroit had a 252-189-12 home record. That comes out to a 57% winning percentage, which is better than their Thanksgiving winning percentage of 53%. I couldn’t find the stats for Dallas. I sent a letter to TMQ with this information. I also mentioned in the letter that, even if his stats were correct, that would only give Detroit an advantage of 7% of a game (53% - 46%), not half a game. TMQ did not print anything about it. My guess is that he did not see it, since he gets so many reader comments.
This year, TMQ repeated his claim so I spent some more time looking up the data. I found more complete numbers from a variety of sources, put it all together, and wrote him a letter saying:
Again, TMQ made no comment in his column. Since I wanted to spread the word somehow, I decided to post this open letter to TMQ. I will also continue to write him. I hope that I can get through to him.
If anyone is interested, here is how I did the statistics for my argument.
My sources for the numbers are:
- Lions’ home & Thanksgiving record since 1934 from the Official Detroit Lions website
- Dallas’s home record 1974-2001 from the Denver Post
- Dallas’s Thanksgiving record from Greg Dempson Sports (1966-2002)
- Dallas and Detroit home records for particular seasons from Football-Reference
From these assorted data, I had to calculate:
The Lions had a home record (1934-2002) of 256-200-12, a home record for 1939-1944 (when they didn’t play on Turkey Day) of 17-17-1, and a Thanksgiving record of 32-28-2. So their record for non-Thanksgiving home games in seasons with a Thanksgiving game is (256 - 17 - 32 = 207) wins, (200 - 17 - 28 = 155) losses, and (12 - 1 - 2 = 9) ties. Ignoring ties, that’s 207 wins in 207 + 155 = 362 games, or a .572 record.
The Cowboys had a home record (1974-2001) of 144-69-0, a home record (1966-1973, 2002) of 49-14-1, a home record for 1975 and 1977 (when they didn’t play on Turkey Day) of 11-3-0, and a Thanksgiving record of 22-12-1. So their record for non-Thanksgiving home games in seasons with a Thanksgiving game is (144 + 49 - 11 - 22 = 160) wins, (69 + 14 - 3 - 12 = 68 losses), and (0 + 1 - 0 - 1 = 0) ties. That’s 160 wins in 160 + 68 = 228 games, or a .702 record.