Wednesday, September 29, 2021

NFL Picks: Holding My Own ... Sort Of

In week one of FiveThirtyEights's NFL forecasting game, I picked 11 of 16 NFL matches correctly, racking up 69.9 points and establishing myself at 62nd place (99th percentile) out of 4,764 players. FiveThirtyEight's own model picked six games correctly and scored -36.3 points.

Last week, I again picked 11 of 16 games correctly, but scored only 66.6 points and fell to 98th percentile and 93rd place out of 5,654 players. FiveThirtyEight picked nine games correctly and scored 40.8 points.

This week, I picked 11 of 16 games correctly for the third week in a row, scoring 88.3 points and maintaining my 98th percentile performance, but falling to 132nd place from among 6,231 players. FiveThirtyEight's model seems to be adapting well to whatever reality it measures. The model picked 10 games correctly and put up 68.7 points.

Naturally, the biggest hit I took was on the Kansas City Chiefs. I gave them a 93% chance of beating the San Diego Chargers. Their loss cost me 61.5 points. FiveThirtyEight also blew that pick, but only lost 31.2 points as they set KC's chance of winning at 75%. That was the model's worst outcome as well.

Of course, FiveThirtyEight is betting its data-driven model, while when it comes to the Chiefs I'm betting my religion. So I think I've got a better excuse where they're concerned.

The picks that I got right and FiveThirtyEight got wrong: I picked Green Bay to beat San Francisco and Cincinnati to beat Pittsburgh.

The single pick that FiveThirtyEight got right and I got wrong was the Washington Whatchamacallits to beat Buffalo.  But I bet small on Washington and only lost two points. FiveThirtyEight bet big on Buffalo and won 22.8 points.

I did change one pick last week, before the games were played. I initially followed my "never bet against Tom Brady unless he's playing against Patrick Mahomes" rule and picked the Tampa Bay Buccaneers to beat the Los Angeles Rams. Then I happened to notice a news article mentioning three positive COVID tests on the Tampa squad that might result in players missing the game. Adding that to the Rams beating Tampa Bay last year, I went for the Rams with a 53% chance of winning, and of course they did.

ADDENDUM: I'm changing my strategy this week, and that change is: BIG bets. A big bet wins more points if I'm right, but loses a LOT more points if I'm wrong. So far I'm right more often than I'm wrong, so I'm going big where I'm sure or where my religious beliefs mandate. Among the big ones, I've got four 100% bets in -- the Bills to beat the Texans, the Buccaneers to beat the Patriots, the Packers to beat the Steelers, and of course the Chiefs to beat the Eagles. Each of those bets will knock down 25 points if I'm right and lose me 75 points if I'm wrong. Wish me luck.

No comments: