dwmyers
Well-Known Member
- Messages
- 2,373
- Reaction score
- 522
I was recently asked to figure out how to calculate offensive SRS and defensive SRS. I figured out a way, though the stat isn't unique.
As part of a series of articles on OSRS and DSRS for various teams, I did the Detroit Lions, and also calculated odds for the upcoming game using pythagoreans, simple rankings, and median point spread.
Those results are here: http://codeandfootball.wordpress.com/2013/10/26/2013-osrs-and-dsrs-quick-hits-detroit-lions/
An excerpt from the article:
Note: in terms of interconverting probabilities and point spreads, we give a rough rule of thumb here:
http://codeandfootball.wordpress.com/2013/01/03/the-beginnings-of-the-nfl-playoffs-some-odds/
To just state the formula
point spread is approximately 7.4*logit(probability to win). Since logits aren't found on regular calculators, the web site Wolfram Alpha can do those calculations for you.
D-
As part of a series of articles on OSRS and DSRS for various teams, I did the Detroit Lions, and also calculated odds for the upcoming game using pythagoreans, simple rankings, and median point spread.
Those results are here: http://codeandfootball.wordpress.com/2013/10/26/2013-osrs-and-dsrs-quick-hits-detroit-lions/
An excerpt from the article:
Detroit hosts Dallas this weekend and not at all surprisingly, for two teams fairly closely matched, the three predictive systems yield three different results. Pythagoreans have Dallas and Detroit as essentially even. Simple rankings would suggest that Dallas has a slight edge. Medians suggest the opposite, that Detroit will win by about 5 points.
Note: in terms of interconverting probabilities and point spreads, we give a rough rule of thumb here:
http://codeandfootball.wordpress.com/2013/01/03/the-beginnings-of-the-nfl-playoffs-some-odds/
To just state the formula
point spread is approximately 7.4*logit(probability to win). Since logits aren't found on regular calculators, the web site Wolfram Alpha can do those calculations for you.
D-