Sunday, December 29, 2013

Chris Bosh single handedly beats the Trailblazers (so to speak)

NBA Win Probability graphs have been updated with games through last night (December 28). With that update, Chris Bosh amassed the largest single game Win Probability Added (WPA) of the season: 104.4% (an entire game's worth, with 4.4% to spare). Here is the win probability graph and box score: MIA 108, POR 107.

Of Bosh's 104.4% WPA, 68% came on a single play: A three pointer to put the Heat up by 1 with less than a second to go in overtime.

See yesterday's post for additional info on the graphs and the calculation of Win Probability Added.

Saturday, December 28, 2013

NBA Win Probability Graphs and Box Scores

Box Score from
Wilt Chamberlain's 100 Point game
I'm excited to announce a new feature to this site: NBA Win Probability Graphs and Box Scores for the 2013-2014 season. I first published these graphs for each game of the 2013 NBA Finals (relive the drama of Game Six...in chart form!). In November, I rolled out an NBA Win Probability Calculator tool that generates win probability as a function of game state (quarter, time remaining, margin, and possession). And over the past few months, I have worked on refining and enhancing my methodology, as well as building an easily accessible database to store win probabilities for all NBA games.

Model Enhancements

But in order to fully roll out this latest feature, I needed a model with the ability to calculate win probability at a finer level of game state. For example, after a missed shot, but before the rebound. Or, when a player has been fouled and has been awarded two free throws. With this level of detail, I can properly apportion Win Probability Added (WPA) at a player level (the key feature of my box score).

Saturday, December 21, 2013

Playoff Implications - Week 16

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background).

Ranking Week 16 Games by Leverage

The table below ranks the week 16 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Turnover Index - Week 16

Here are the Turnover Index picks for Week 16. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 15 results

Our sole week 15 bet against the spread was successful (just barely, the Jets were 10.5 point underdogs and lost to the Panthers by 10). Here is the season-to-date performance:
  • Against the Spread: 12-15-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $814 (-19% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741
12 1 1 $741 $37 (5.1%) $34 $775
13 4 2 $775 $46 (5.9%) $21 $797
15 1 1 $797 $18 (2.3%) $16 $814

Sunday, December 15, 2013

Playoff Implications - Week 15

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background).

Ranking Week 15 Games by Leverage

The table below ranks the week 15 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Saturday, December 14, 2013

Turnover Index - Week 15

Here are the Turnover Index picks for Week 15. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 14 results

There were no week 14 results as no bets met our criteria. Here is the season-to-date performance:
  • Against the Spread: 11-15-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $775 (-20% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741
12 1 1 $741 $37 (5.1%) $34 $775
13 4 2 $775 $46 (5.9%) $21 $797

Saturday, December 7, 2013

Playoff Implications - Week 14

HereHere is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background).

Ranking Week 14 Games by Leverage

The table below ranks the week 14 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

NBA Rankings Now Available

NBA team rankings for the 2013-2014 season are now available (they've been available for a few weeks now, but I've been trying to see if I can fix the issues where some users can't view the table). My rankings are an attempt to reverse engineer what the betting market "thinks" are the best and worst teams. See here for more background, or my Methodology page for a broader overview.

Just mouse over the column headings for a description of what they mean. Each metric has an associated sparkline showing the season-to-date trajectory.

I've also updated The Ticker as well, which displays each team's progress over time in chart form (there are also alternate versions with additional detail).

The Pacers crash the party

For much of last season, the Heat, Thunder, and Spurs dominated the top 3 spots in the rankings. Owing to their strong 17-2 start, the Pacers have displaced the Heat in that top 3 (for now).

The market is still playing "wait and see" on a few teams this season, despite stronger than expected starts. The Lakers are above 0.500, with the 12th best record in the league, but are ranked #22 according to the market. It will be interesting to see how the rankings move in response to Kobe Bryant's return.

The 76ers have also performed somewhat above expectations (although 7-13 is nothing to brag about), but the market has stayed firm in its estimation of Philly as a bottom dweller this season.

Wednesday, December 4, 2013

Turnover Index - Week 14

There are no Turnover Index picks for Week 14 (no games meet the minimum criteria), but here are the results for Week 13. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 13 results

The Week 13 picks went 2-2 against the spread, but with a $21 gain to the bankroll (the two winning bets were for higher stakes). Here are the season to date results:
  • Against the Spread: 11-15-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $797 (-20% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741
12 1 1 $741 $37 (5.1%) $34 $775
13 4 2 $775 $46 (5.9%) $21 $797

The "Old" Index

Under last season's simpler betting approach (bet any game with at least a 10 turnover differential), season to date results for 2013 are 8-9-1 against the spread.

Saturday, November 30, 2013

Michigan Plays for the Win

Michigan nearly pulled off an upset with huge BCS implications this afternoon, falling just short against the Buckeyes. The Wolverines had scored a touchdown with 30 seconds to go in regulation, and could have tied the game with an extra point. Michigan instead chose to play for the win by going for two. According to a Michigan tweet:
Taylor Lewan, on the two point attempt, "Coach Hoke asked us seniors, do you want to go for it, we all said yes."
And though the two point conversion attempt failed, their seniors still made the right call. Michigan was a 15.5 point underdog at home against Ohio State. As I pointed out in last week's post, underdogs fare poorly in overtime. There have been 92 college football games in which a 10 point or more underdog has gone to overtime. They have won just 35% of those contests. The results are consistent across other sports.

I believe 2 point conversion attempts succeed a little less than 50% of the time (I could not track down hard numbers on this for college football). So let's say that the chance of success is 45%. Here is Michigan's win probability* for both strategies:

  • Go For 2: 45%
  • Play for Overtime: 35%

And the 35% is probably a bit high since it includes underdogs of 10 points or more. For underdogs greater than 15 points, those teams have won 32% of overtime games. So, unless Coach Hoke thought his team had less than a ~33% chance of making that two point conversion, going for two was the smart call.

* Note that I am ignoring the possibility of the Buckeyes driving for a quick score in the last 30 seconds of regulation. But even if that had a 10% likelihood (which seems high), that would take the 45% down to 40%, which is still better than 35%. And that assumes that they would not try to score if the game was tied.

Wednesday, November 27, 2013

Playoff Implications - Week 13

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background).

Ranking Week 13 Games by Leverage

The table below ranks the week 13 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Turnover Index - Week 13

Here are the Turnover Index picks for Week 13. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 12 results

The Week 12 pick went 1-0 against the spread, with a $34 gain to the bankroll. Here are the season to date results:
  • Against the Spread: 9-13-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $775 (-22% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741
12 1 1 $741 $37 (5.1%) $34 $775

Saturday, November 23, 2013

Playoff Implications - Week 12

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background).

Ranking Week 12 Games by Leverage

The table below ranks the week 12 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Wednesday, November 20, 2013

Turnover Index - Week 12

Here are the Turnover Index picks for Week 12. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 11 results

It's been a rough four weeks. Week 11 picks went 0-2 against the spread, with a $14 loss to the bankroll. Here are the season to date results:
  • Against the Spread: 8-13-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $741 (-26% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741

Sunday, November 17, 2013

Underdogs Shouldn't Play for Overtime


In week 9 of the NFL season, the Buccaneers nearly pulled off an historic upset playing on the road against the heavily favored Seahawks. The Bucs were a 15.5 point underdog, yet somehow found themselves leading 21-0 with just over two minutes to go in the first half. Going back to 1989, only three NFL teams have won a game as a 15.5 point or greater underdog.

When playing as an underdog, there are a couple things a team can do to maximize its chances of winning:
  1. Go for high variance plays (surprise onside kicks, going for it on 4th and short, two point conversions, deep passes, etc.)
  2. Shorten the game when possible (when you're the underdog, the house will eventually win if you stick around long enough, so cash out while you can)
Credit Coach Schiano for calling for a surprise onside kick in the second quarter. The Seahawks recovered the kick (just barely), but it was still a smart call. Late in the game, Schiano had another opportunity to make the right call....

NBA Win Probability Calculator

Classic shot of the ENIACThis past summer, I rolled out my initial attempt at a win probability model for the NBA (introductory post | the insanity that was Game 6 | graphs for all playoff games | win probability added). Since that time, I have refined the model somewhat via more rigorous cross-validation of the model parameters. And while I may still have some tinkering to do, I'm ready to share a beta version of an online Win Probability calculator.

Much like the Win Probability tool from Advanced NFL Stats, this tool allows you to input the game state (time remaining, margin, possession) and it will return the win probability of said game state.

So what's the point? From my perspective, a win probability model has three main potential uses:

Saturday, November 16, 2013

Playoff Implications - Week 11

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see this post for background). I can summarize the total impact a game is expected to have, as well as how it affects each particular team.

The main change from last week is that I am now filtering out results that aren't statistically significant. I am using the Fisher Exact Test for this purpose, with a significance level of 95%.

Ranking Week 11 Games by Leverage

The table below ranks the week 11 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Wednesday, November 13, 2013

Turnover Index - Week 11

Here are the Turnover Index picks for Week 11. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 10 results

Once again, it was not a good week. Week 10 picks went 1-2-1 against the spread, with a $45 loss to the bankroll. One callout is on the Texans-Cardinals game. The general consensus seems to have the closing line at 3.5 or 4 points in favor of the Cardinals. The Cardinals won by 3, meaning the bet on the Texans should have covered. But my official source for lines and stats is sportsdatabase.com, which has the closing line at -3. I don't want to go down the rabbit hole of double-checking every single game, so I'm sticking with the sportsdatabase numbers with the assumption that any discrepancies even out in the long run. Here are the season to date results:
  • Against the Spread: 8-11-1
  • Starting Bankroll: $1,000
  • Current Bankroll: $755 (-25% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755

Saturday, November 9, 2013

Playoff Implications - Week 10

Here is your weekly cheat sheet for NFL games with playoff implications. The purpose of this feature is to highlight games that have a big impact on the playoff picture (see last week's post for my first attempt). I can summarize the total impact a game is expected to have, as well as how it affects each particular team.

The main change from last week is that I am now filtering out results that aren't statistically significant. I am using the Fisher Exact Test for this purpose, with a significance level of 95%.

Ranking Week 10 Games by Leverage

The table below ranks the week 10 games by total leverage. Leverage in this context is a measure of both how uncertain a game's outcome is (games between evenly matched teams have higher leverage) and how much the playoff picture swings as a result of that outcome.

Friday, November 8, 2013

Could a playoff seed ever be determined by a coin flip?

Warning: Pointlessness ahead.

In order to produce my playoff seed projections for the NFL, I had to do some rather tedious coding of the tiebreaker rules. For example, here are the divisional tiebreaker rules when two teams have the same win-loss-tie record:
  1. Head-to-head (best won-lost-tied percentage in games between the clubs).
  2. Best won-lost-tied percentage in games played within the division.
  3. Best won-lost-tied percentage in common games.
  4. Best won-lost-tied percentage in games played within the conference.
  5. Strength of victory.
  6. Strength of schedule.
  7. Best combined ranking among conference teams in points scored and points allowed.
  8. Best combined ranking among all teams in points scored and points allowed.
  9. Best net points in common games.
  10. Best net points in all games.
  11. Best net touchdowns in all games.
  12. Coin toss
Most casual fans are aware of the implications head-to-head, divisional records, and conference records have for playoff seeding. But things get pretty esoteric as you move further down the list. I had to look up what "Strength of Victory" meant (it's the combined winning percentage of the opponents you beat), and frankly had to make a bit of a guess as to what "combined ranking" meant for #7 and #8. At the very bottom of the list is "Coin Toss", which I even went through the trouble of simulating with a random number generator.

But could the "Coin Toss" scenario ever come in to play? (doubtful) Would ESPN televise it? (probably) Should they have replaced "Coin Toss" with "Dance Off"? (definitely)

Wednesday, November 6, 2013

Turnover Index - Week 10

Here are the Turnover Index picks for Week 10. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 9 results

It was not a good week. Week 9 picks went 2-5 against the spread, with a $67 loss to the bankroll. Here are the season to date results:
  • Against the Spread: 7-9
  • Starting Bankroll: $1,000
  • Current Bankroll: $800 (-20% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800

Sunday, November 3, 2013

Playoff Leverage - Week 9

Earlier this week, I added playoff seed projections to my daily NFL rankings, which are based on simulating the remainder of the season 10,000 times. In addition to projecting playoff seeds, I can take the same simulation dataset and use it to see which games have the biggest impact on the playoff picture.

You can find a similar feature at Advanced NFL Stats, which highlights the high leverage game of the week (it's GB-CHI according to their model). This post attempts to take a more detailed view, looking at how each particular game affects all teams, not just the two participants.

A Big Week for Baltimore

My playoff simulations indicate that the defending champion Ravens have a 29% chance of making it to the playoffs at this point. This week's matchup against the Browns will have a big impact on those chances. If the Ravens lose, their playoff hopes drop to 16%. And if they win, they increase to 39%. That swing, coupled with the fact that the game is fairly evenly matched (Baltimore has a 56% chance of winning at Cleveland), makes this the highest leverage situation of Week 9.

Note that high leverage doesn't just mean a big potential swing in outcome. If I buy a lottery ticket, there is a big difference between me winning and losing. But since the probability of winning is so small, the leverage is small as well. This article from Tom Tango explains the concept (fyi - I am using the recipe #2 version of the leverage calculation).

Wednesday, October 30, 2013

Turnover Index - Week 9

Here are the Turnover Index picks for Week 9. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 8 results

Week 8 picks went 1-2 against the spread. Fortunately, the one successful bet was with the highest stake, so the overall loss to the bankroll was minimal. Here are the season to date results:
  • Against the Spread: 5-4
  • Starting Bankroll: $1,000
  • Current Bankroll: $867 (-13% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867

Sunday, October 27, 2013

NFL Playoff Seed Probabilities

Because it's never too soon to start talking playoffs.

The latest version of my NFL ranking table features some odd looking multi-colored bar graphs, taking the place of the monochrome Projected Wins bar graph. These new graphs represents each team's projected playoff seed probabilities, from #16 on the left to #1 on the right. The red bars are for seeds 16-7 (out of the playoffs) and the blue bars are for seeds 6-1 (wildcard and division winners). Mouseover the bar graph for the actual probabilities.

The probabilities are based on simulating the remainder of the season 10,000 times (see the Projected Wins post for more details). I then apply the NFL playoff seeding rules and various tiebreakers. As a reminder, the top 4 seeds belong to the four division winners, with the remaining two playoff seeds going to the two non-division winners with the best record. For a fully armed and operational NFL simulation tool, check out nfl-forecast.com.

This is virtually identical to a feature I added to my NBA rankings in February of this year. After the regular season was finished, I also set the probabilities in motion, something I plan on on doing for the NFL as well, once week 17 is complete. Some observations:

Wednesday, October 23, 2013

Turnover Index - Week 8

Here are the Turnover Index picks for Week 8. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 7 results

Week 7 picks went 3-0 against the spread. Here are the season to date results:
  • Against the Spread: 4-2
  • Starting Bankroll: $1,000
  • Current Bankroll: $879 (-12% ROI)
Here is the week by week performance:

weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879

Were it not for my pathological honesty, I would be reporting a 4-0 record against the spread this season, as the the initial two losses in week 4 occurred before I could start publishing the turnover index this season. As I was running the data for this season's version of the index, I debated whether to start betting in week 4 or week 5. But week 4 showed positive returns over the testing window (1999-2012), so I included it.

Despite the positive 4-2 record against the spread, my bankroll (managed according to the Kelly Criterion) is still underwater, having lost 121 of my starting 1000 (Zimbabwean) dollars.

Friday, October 18, 2013

Turnover Index - Week 7

Here are the Turnover Index picks for Week 7. The Turnover Index is a simple betting strategy based on the theory that the market overvalues defensive turnovers when judging team strength. See here and here for more background.

Week 6 results

Week 6 picks (or pick) went 1-0 against the spread (Chargers covered against the Colts). Here are the season to date results:
  • Against the Spread: 1-2
  • Starting Bankroll: $1,000
  • Current Bankroll: $761 (-24% ROI)
There is a table at the bottom of the post which summarizes performance by week in more detail.

Early Season Power Rankings - Follow Up

This is a follow up to last week's post on early season power rankings. In that post, I attempted to measure the accuracy of various NFL power rankings by comparing the team rankings after week 4 of the season to each team's remaining wins for weeks 5-16.

The original dataset was for seasons 2009-2012, and just looked at the week 4 rankings. I have now added seasons 2007 and 2008, plus a look at how the rankings after week 8 correlated with future wins. As a reminder, higher percentages are better, with 100% meaning perfect correlation between the ranking and future wins (the metric used is the Spearman rank correlation coefficient).

Week 5 Ranking Correlation to Future Wins
ranking average 2007 2008 2009 2010 2011 2012
espn 47% 55% 42% 51% 55% 43% 39%
dvoa 50% 57% 45% 47% 46% 41% 65%
ans 40% 50% 42% 51% 15% 49% 32%
market 54% 68% 36% 67% 45% 52% 56%
srs 52% 70% 46% 47% 38% 56% 54%

Week 8 Ranking Correlation to Future Wins
ranking average 2007 2008 2009 2010 2011 2012
espn 46% 58% 42% 51% 43% 41% 41%
dvoa 53% 77% 53% 59% 44% 35% 46%
ans 49% 70% 47% 57% 38% 39% 42%
market 55% 62% 50% 62% 56% 55% 42%
srs 53% 75% 51% 55% 46% 42% 50%


Keep it Simple

The performance of the Simple Ranking System (SRS) continues to impress, either matching or outperforming the more sophisticated DVOA and ANS models. The market ranking still appears to be the most accurate.

More data makes you dumber

Interesting to note that while all of the quantitative rankings improved going from week 4 to week 8 (particularly the Advanced NFL Stats efficiency model), the qualitative ESPN ranking took a small step backward. It's hard to read too much into such a small difference, but it is consistent with Ken Pomeroy's observations on the AP rankings for college basketball. For the 1990-2010 academic years, he compared tournament performance of the pre-season AP #1 to the AP #1 immediately prior to the tournament. In general the pre-season #1 fared better in the tournament (out of 21 seasons, the pre-season #1 won it all 6 times, compared to just 3 for the pre-tournament #1).

It's counter-intuitive, but has a simple explanation. As you get further into the season, most poll-based rankings tend to become retrodictive, rather than predictive. In other words, they "reward" teams for past wins, rather than judge them on their ability to create future wins. Bill Parcells statements not withstanding, a team's win-loss record is not always the best predictor of future success.

Saturday, October 12, 2013

The Turnover Index (possibly overfit edition)

Andrew Luck fumbles at 2009 Big GameThe Turnover Index makes a belated return this season, with some additional tweaks and enhancements. The Turnover Index is a simple betting strategy based on the theory that the betting market overvalues defensive turnovers when evaluating future team performance. I laid out the evidence in this post: NFL Turnover Differential and the Point Spread. Using data from the 1998-2011 seasons, I found that teams that had generated at least 10 fewer defensive turnovers (season-to-date) than their opponents covered the spread 58.7% of the time.

In a series of weekly posts, I tracked the performance of this betting strategy (bet on the team with at least 10 fewer defensive turnovers) throughout the 2012 NFL season. Initial results were promising, but regressed somewhat near the end of the season. At season's end, the strategy had gone 18-16-1 against the spread, for a whopping 1% return on in investment.

In this post, I will lay out a somewhat revised version of the turnover index which will allow a more sophisticated betting strategy based on the Kelly criterion for bankroll management.