NY Times 5th Down NFL System

The 5th Down's Brian Burke has started his win probability predictions again and I will be betting them. If I recall correctly, he has been 60% accurate against the spread for all NFL games since 2007. I will be betting every game with 2.5% of my bankroll against the spread. The predictions are posted every Thursday on the NY Times website in the 5th down section.

Wish me luck!

Post the plays

I just checked the blog. You just take the win probability and bet each game against the spread?

Yes. The win probability only predicts each teams chances to win outright, and I am picking the odds favourite against the spread. I think only 10% of NFL game winners do not cover, so I'm taking that chance.

Here are the win probabilities for week 4.

0.64

Good luck bro.

Well that didn't work. Just check the NY Times and I'll post the results after each week is over.

0.64

0.64

Cleveland at BaltimoreCarolina at AtlantaNew England at BuffaloMinnesota at DetroitSan Diego at Kansas CitySeattle at St. LouisSan Francisco at JetsTennessee at HoustonCincinnati at JacksonvilleMiami at ArizonaOakland at DenverNew Orleans at Green BayWashington at Tampa BayGiants at PhiladelphiaChicago at Dallas

0.74
0.62
0.58
0.47
0.47
0.44
0.42
0.84
0.55
0.69
0.78
0.81
0.74
0.67
0.64

I guess messages on this forum are incapable of posting tables.

Ok here is my last attempt to post the picks for this week (includes the Thursday nighter)

Baltimore 74% (already played)
Atlanta 62%
Buffalo 58%
Minnesota 53%
San Diego 53%
Seattle 56%
San Francisco 58%
Houston 84%
Jacksonville 55%
Arizona 69%
Denver 78%
Green Bay 81%
Tampa Bay 74%
Philadelphia 67%
Dallas 64%

Bear in mind these are chances to win outright and not to cover the spread. However, the winners usually do cover the spread.

I checked this a few years ago...his straight up win per cents were very accurate. That is, teams that he said had a 65% chance of winning won about 65% of the time. But this was all straight up.

When his straight up win projections were applied to point spreads, there was no real advantage.

Just an FYI teamrankings.com nfl game predictions that are 3 star or better to win SU is 85% - over 3 years.

But can imagine those are heavy favorite games - but maybe 2 team ML parlays might be a profitable endeavor

I disagree. Here is the evidence...

https://www.advancednflstats.com/2011/10/prediction-accuracy.html

He compares his accuracy to Vegas's opening lines.

I think you miss read that post. Burke list his winning percent straight up at 66.2% and then shows how often the Las Vegas opening line favorite won which is 66.1%. Burke does not discuss how his straight up picks did vs the Las Vegas line.

I think Burke does a great job...but he is not picking winners against the spread. 

Hmmm. You're right. I misread that.  Not good for my strategy.  Well, if you assessed his accuracy against the spread and there is no advantage then that puts me in a pickle.  If you have any links showing his record against the spread please provide the link. Thanks.

I don't know of any links...and I no longer have the files.

But if I reacall correctly neither money line bets nor bets on sides were profitable.

 

The analysis I did was a while back. Here is a myl write up, which does not focus on betting.

https://community.advancednflstats.com/2008/12/straight-up-winners.html

The real basic question is how can in- game box score data (pass, runs, time of possession, yards gained etc) be an INDEPENDENT variable  when the score influences that data.

Thanks for the heads up Jarhead on cautioning me against this strategy. Brian Burke went 10 for 15 straight up which is in line with his record. Against the spread is another story going 5-9 and 1 push for a net loss.

It was interesting in that article that the Vegas line gets it right almost as much as your models. It makes it discouraging to bet on NFL.

Well you can use that info with teases to avoid betting against a spread, essentially get enough point and each team to make them a Pickem or slight dog. Or even just bet his dog plays

What about going against his ml favs ? 9-5 is 64%.

ok