Saturday, December 29, 2012

Bowl Game Line Predictions

The power rankings I create for this site are calibrated to a single purpose: to predict, as closely as possible, point spreads for future matchups.  I am trying to understand what (and how) the betting market thinks.  If my rankings worked perfectly, you would take the difference of the opposing teams' GPF (Generic Points Favored), add a home field adjustment, and the resulting number would equal the point spread.

Of course, my rankings do not work perfectly, and the predicted point spreads based on my rankings tend to miss, on average, by about 2.5 points (for college football). In a previous post, I compared my ranking's accuracy to the rankings provided by the Sporting News Linemakers section (the only other publicly available rankings system I'm aware of that has the same purpose as mine).  For the week I chose, I found that the predictive accuracy of the two ranking systems was about the same (my rankings had a mean prediction error of 2.1 points, compared to Linemakers' error of 2.2 points).

Thursday, December 27, 2012

Expected Points and Relative Team Strength

This is a continuation of a series of posts on in-game cover probability.  The purpose of this post is to show how relative team strength (as reflected in the point spread) can be factored into the concept of Expected Points.

In my kickoff post on the topic, I laid out two enhancements to the Advanced NFL Stats Win Probability Model that would be needed in order to develop a proper cover probability model:

1. Probability estimates that factor in the a priori relative team strength implied by the point spread.
2. Estimates of scoring margin distribution, not just a binary win/loss probability

I have not made much progress on either front at this point, largely due to the difficulties in just recreating Brian Burke's original work on Win Probability.  The problem is the sparseness of the data.  There just aren't enough actual game situations that feature, for example, a team down by 4, with 3rd and 2 at their 36 yard line and 2:30 to go in the 3rd quarter.  So, some careful interpolation is needed. Adding in the point spread as an additional independent variable only makes the data that much sparser and requires even more care when interpolating.  Adding margin distributions on top of that gives me LOST-style nosebleeds if I think about it too much.

Turnover Index - Final 2012 Results

Here are the final results for the Turnover Index for the 2012 season.  In my original post on the topic, I excluded week 17 games from the analysis, so there are no more betting opportunities this season.

Week 16 Results

Things ended on a low note, with the week 16 picks going 1-3 against the spread.

Turnover Index through Week 16 (Against the Spread): 18-16-1

That's 52.9% against the spread.  Assuming the standard sportsbook bet (lay $110 for a chance at a $100 profit), you would have realized a whopping 1% return on investment.  That's below the long term performance of 59% against the spread observed from seasons 1998-2011, but at least it's positive.

So what, if anything, went wrong?

Wednesday, December 19, 2012

Turnover Index - Week 16

Here are the Turnover Index numbers for Week 16 of the NFL season.  The purpose of these weekly posts is to find betting opportunities that exploit the market's tendency to overvalue a team's season to date defensive turnovers.  See here and here for more background.

Wednesday, December 12, 2012

Turnover Index - Week 15

Here are the Turnover Index numbers for Week 15 of the NFL season.  The purpose of these weekly posts is to find betting opportunities that exploit the market's tendency to overvalue a team's season to date defensive turnovers.  See here and here for more background.

Wednesday, December 5, 2012

Turnover Index - Week 14

Here are the Turnover Index numbers for Week 14 of the NFL season.  The purpose of these weekly posts is to find betting opportunities that exploit the market's tendency to overvalue a team's season to date defensive turnovers.  See here and here for more background.

Sunday, December 2, 2012

Alabama - Notre Dame Opening Line

In this post, I made a prediction on the point spread and over/under for the two possible BCS title games (Notre Dame vs. Alabama or Georgia).  Even though the game has not been officially announced, Vegas is not wasting anytime (this game is projected to be the most heavily bet game ever).  Opening lines are available and it looks like my prediction favored Alabama too much (by about 3 points).

My rankings as of November 25 predicted Alabama as an 11 point favorite over Notre Dame.  Even prior to the release of the line, that looked to be on the high end of estimates I could find from gambling experts, where things ranged from 7 points to 11 points.  In an earlier post this week, I put my rankings to the test against the Linemakers' rankings.  According to their November 28 rankings, Alabama should have been a six point favorite.

Things still seem to be a bit fluid, but best as I can tell the line is currently 8 points in favor of Alabama, which means my prediction was off by 3 points.  However, my rankings are calibrated to closing lines, so I will probably check in at least once prior to the title game to see which way the line moves.

Saturday, December 1, 2012

Today's Games now available for the NBA

Now that the NBA rankings are up and running, I can now relaunch the Today's Games feature of this site.  Here is a description of what this feature is intended to do (you can also see here for additional background).

Today's Games

The main purpose of this feature is to highlight "unexpected" point spreads and over/unders for games to be played that day.  I can calculate expected point spreads and over/unders using my betting market rankings (see the Methodology page for a description of how the rankings are derived, and see the end of this post for how to use the rankings).

Wednesday, November 28, 2012

The Turnover Index - Week 13

Here are the Turnover Index numbers for Week 13 of the NFL season.  The purpose of these weekly posts is to find betting opportunities that exploit the market's tendency to overvalue a team's season to date defensive turnovers.  See here and here for more background.

Last week's picks just went 1-2 against the spread.  The Rams covered in their game against the Cardinals, but the Vikings and Packers failed to cover in their respective matchups against the Bears and Giants.

Tuesday, November 27, 2012

Betting Market Analytics vs. The Linemakers

While googling line predictions for the BCS Title game for this post, I stumbled across a section of the Sporting News called The Linemakers.  They have a power ratings section that appears to be doing the same thing I'm attempting with my rankings.  Here is the description from the site:

"Power Ratings are numerical values assigned to teams that intend to provide the basis for what the spread should be between two teams at a neutral site. They are updated every week during the season."

That could function as a description of my rankings as well.  To be honest, I don't how they derive the rankings at that site, but I figured I would put their rankings and mine to the test and see which set was best at predicting how the market sets the point spreads for upcoming games.

Sunday, November 25, 2012

BCS Title Game - Line Predictions

With the BCS Title Game scenarios now down to Georgia-Notre Dame or Alabama-Notre Dame, here are predictions of what the point spread and over/under would be for both potential matchups, using my betting market rankings.

The Ticker - NBA and NFL Versions Now Available

The Ticker is now available for both the 2012 NFL season and the 2012-2013 NBA season.  The Ticker is a feature I first launched for the NBA back in March of this year.  The purpose of this feature is to display how each team's Generic Points Favored (GPF) has changed as the season has progressed.  My GPF ranking can be thought of (almost literally) as the stock price of a team, so the Ticker tracks that movement over time, as a standard stock ticker would.

NFL

I just have one chart for the NFL, displaying total team GPF.  As you can see, the movement is fairly noisy from day to day, but some overall patterns are fairly clear.  You can see how the market's faith in Arizona has slowly bled away over the past two months.  Or how Denver's stock has continued to rise as Peyton Manning has put to rest any doubts about his ability to return from injury.

NBA

There are three charts for the NBA:
  • GPF - Same as the NFL, an area chart showing day to day movement in team Generic Points Favored.  It's pretty early in the season, so there's not a whole lot of movement.
  • GPF+oGPF+dGPF - This chart displays team GPF along with its offensive and defensive components: oGPF and dGPF.
  • GOU - Stands for Generic Over/Under.  This is what you would expect the betting over/under to be when playing a league average team.  This is a function of offense, defense, and pace.  For comparison, the graph also displays the league average over/under (AOU).


NFL Scoring Margin Distributions

This is a follow up from my previous post on building an in-game cover probability model for the NFL.  As a first step in that process, I calculated an expected scoring distribution for any given point spread.  I derived this by building an ordered logistic model off of actual NFL scores, with the Vegas point spread as my independent variable.  In that original post, I shared the expected distribution for a 7 point favorite.  The tool below will provide the distribution for any point spread (from 0 points to 21 points).

Saturday, November 24, 2012

NBA Team Rankings Now Available

NBA team rankings are now available for the 2012 season.  For those of you unfamiliar with my ranking methodology, the goal is to reverse engineer an implied power ranking from the Vegas point spreads.  I'm trying to figure out who the betting market "thinks" are the best teams in the league.  In addition, by combining the point spread with the betting over/under, I can decompose team strength into its offensive and defensive components.  My Methodology page has a simple example of how the rankings work.  You can also refer to my series of posts at Advanced NFL Stats Community, where I first created these for the NFL.

Here is a direct link to the ranking table.  As with my NFL, NCAAF, and MLB rankings, these update every morning with the latest game results and betting information.  My source is Sportsdatabase.com.

Friday, November 23, 2012

In-Game Cover Probability: A Start

This is the first post in what I hope to be a series on the topic of in-game cover probability.  I view the idea as an extension of the Win Probability graphs at Advanced NFL Stats.  For those of you not familiar with the win probability graphs, the general idea is to estimate a team's chances of winning a game given down, distance to go, field position, time remaining, and point differential.  As I understand it, the probabilities are not derived from a Madden-style simulation, but are rather based on actual game outcomes, with some LOWESS smoothing to fill in the gaps where there isn't sufficient data.

What I hope to explore here is whether a similar model can be built to determine a team's probability of covering the spread.  As an example, take the week 7 matchup between Chicago and Detroit.  With about 40 seconds to go in the game, Detroit trailed by 13, but had the ball on the Chicago 12.  The ANS probability model gave Detroit a 1% chance of winning the game at this point, so not a lot of drama, unless you had bet on the spread.  Detroit was a 6.5 point underdog.  A touchdown by Detroit would put the margin at 6 points and Detroit would cover the spread.

Tuesday, November 20, 2012

The Turnover Index - Week 12

Here are the Turnover Index numbers for Week 12 of the NFL season.  The purpose of these weekly posts is to find betting opportunities that exploit the market's tendency to overvalue a team's season to date defensive turnovers.  See here and here for more background.

Last week's picks went 1-1 against the spread.  The Bears-49ers game was a perfect example of what these posts intend to illustrate: The Bears, despite having averaged 3.3 defensive turnovers per game and facing a QB making his first NFL start, could not generate a single turnover.  Regression to the mean in action.

Wednesday, November 14, 2012

The Turnover Index - Week 11

Here are the Turnover Index numbers for Week 11 of the NFL season.  The purpose of these weekly posts is to find betting opportunities by exploiting the market's tendency to overvalue a team's season to date defensive turnovers when evaluating future performance.  See here and here for more background.

Last week's picks went 2-0 against the spread.  Cincinnati covered the spread (and then some) in their matchup against the Giants.  The Giants managed to force only one turnover, despite having averaged about 3 per game for the season.  The Bears failed to cover the spread against the Texans, having lost Jay Cutler to a concussion midway through the game.  Chicago forced two turnovers in the game, which is above the league average of 1.6, but below their season-to-date average of 3.5.

Wednesday, November 7, 2012

The Turnover Index - Week 10

Here are the Turnover Index numbers for Week 10 of the NFL season.  The purpose of these weekly posts is to find betting opportunities by exploiting the market's tendency to overvalue a team's season to date defensive turnovers when evaluating future performance.  See here and here for more background.

The Bears - 28 Turnovers in 8 Games

Before getting to the Week 10 numbers, I wanted to take a look at the Bears and their impressive turnover performance thus far this season.  In their first 8 games, the Bears have generated (received?) 28 defensive turnovers.  Going back to 1998, only the 2005 Bengals managed more turnovers (29) in the same amount of time.  See the table below for the outcomes for all teams that managed at least 24 defensive turnovers in their first 8 games:

Monday, November 5, 2012

Ranking Table - Mouseovers

In my ongoing effort to crowd 10 pounds of data into a 5 pound bag, I've added some mouseovers to my NFL and NCAA Football rankings that provide additional layers of drill-down for the numbers I'm reporting.

NFL Tables

  • Last Week - If you mouse over the "LstWk" sparkline for each team, you'll see the actual daily GPF (Generic Points Favored) numbers for the team going back 7 days.
  • Projected Wins - Mouse over the bar graph for the numerical win total probabilities.  For example, as of November 5, the Falcons have a 3% chance of going 16-0 in the regular season, according to my simulations.
  • Strength of Schedule - The SOS columns on the far right summarize the average GPF of each team's opponents, split by opponents already faced and opponents remaining on the schedule.  Mouse over the number for the week by week detail.  As an example, the Broncos have the easiest remaining schedule as of November 5, with an opponent average GPF of -2.3 points.  Mousing over that number shows that only one of Denver's remaining 8 games comes against a team with an above average GPF.

Wednesday, October 31, 2012

The Turnover Index - Week 9

Here are the Turnover Index numbers for Week 9 of the NFL season.  The purpose of these weekly posts is to find betting opportunities by exploiting the market's tendency to overvalue a team's season to date defensive turnovers when evaluating future performance.  See here and here for more background.

Last week's picks went 1-1 against the spread.  Carolina covered the 7.5 point spread in their game against the Bears.  The Cowboys, as a 2.5 point underdog, were unable to cover the spread, done in by Dez Bryant's pinky (and instinct for self preservation).

Sunday, October 28, 2012

College Football Rankings - Now Updated Daily

Over the past few weeks, I had taken several stabs at creating a betting market power ranking for College Football.  Now that I've got the methodology ironed out and the code automated, I'm launching a daily version of the rankings.  These will update every morning with the latest lines and over/unders.  The format is very similar to the format I've established for my NFL rankings.

Here is an overview of the table (you can also hover over the column headings in the table for a brief description):

Wednesday, October 24, 2012

The Turnover Index - Week 8

Here are the Turnover Index numbers for Week 8 of the NFL season.  The purpose of these weekly posts is to find betting opportunities by exploiting the market's tendency to overvalue a team's season to date defensive turnovers when evaluating future performance.  See here and here for more background.

The Turnover Index went 2-0 against the spread last week, although the Bears-Lions game was about as close as you can get (the point spread was 6.5 points, which the Lions just covered by closing the gap to 6 points with 36 seconds left in the game). The other pick, the Colts, beat the Browns by 4, having been favored by 2.5.

Saturday, October 20, 2012

NFL Rankings - Strength of Schedule

At the risk of turning my daily NFL rankings into a cluttered, confusing mess, I have added a couple new columns.

Strength of Schedule

The rankings table now shows a set of columns labelled "SOS", which stands for Strength of Schedule.  This is a very simple calculation where I take each team and calculate the average GPF (Generic Points Favored) of their opponents, split by games already played (the "Past" column) and by remaining games (the "Future" column).  The numbers shown are in terms of GPF, with the corresponding rank beside it.

College Football Rankings - Week 8 Update

As I mentioned in the comments of my week 8 college football rankings post, the rankings looked a bit out of whack.  I've tinkered with things a bit more and I now think that I've got an approach that I can stick with week to week.  Here is a rundown of the methodology (or you can skip to the bottom for the rankings):

Wednesday, October 17, 2012

College Football Rankings - Week 8

Here are the betting market rankings for College Football for week 8.  See my first post for background and my Methodology page for a simple example of how the rankings work.

For this version, I have added the offensive and defensive components of each team's ranking (oGPF and dGPF).  I can derive these by looking at the over/under in combination with the point spread.

Monday, October 15, 2012

The Turnover Index - Week 7

San Diego and Denver have yet to play, but they are both on bye next week, so I can go ahead and update the Turnover Index with data through week 6.  If you don't know what this is, see here and here.

Things evened up this week on the season-to-date results, with the Turnover Index predicting Oakland to cover the nine point spread against Atlanta (which they did, in a wild, losing effort).

NFL Rankings - Projected Wins

"IT'S A SIMULATION!"
There is a new feature added to my daily NFL team rankings. Building off of a suggestion from commenter Ben Moore on my MLB rankings, there is now a column labeled "Projected Wins". With each day's rankings, I also simulate the remainder of the regular season (10,000 times) and display the following:

  • Average Projected Wins - This is the average projected regular season wins, based on my 10,000 simulation runs.  This only simulates future games (wins in-hand are taken as a given).
  • Projected Win Ranking - How that team's projected wins compares to the rest of the league
  • Projected Win Distribution - On the far right is a bar graph displaying the team's projected win distribution, once again summarized from my 10,000 run simulation. The scale is identical for each team, with the x-axis varying from 0 wins to 16 wins for each team.  This can be viewed as an application of Edward Tufte's concept of small multiples.
I simulate the season as follows:

Saturday, October 13, 2012

College Football Rankings - Week 7 Update

My first post on college football rankings was calculated prior to the release of the betting lines for this week's games.  See below for an updated version that reflects the latest lines (LSU dropped and South Carolina shot up to reflect the fact that the Tigers are only a 2.5 point favorite at home over the Gamecocks.

The Turnover Index - Week 6

The Turnover Index got off to a rough start this season.  The first and only pick was for Jacksonville to cover the spread as a 6 point underdog against Chicago.  Just missed it.

Turnover Index Through Week 5 (ATS): 0-1-0

Here is the outlook for the week 6 games (see here and here for background):


"home.dto" and "away.dto" are the season-to-date defensive turnovers for the home and away teams. As before, we're looking for any games where there is at least a 10 turnover differential between the teams. The only game that qualifies this week is Oakland at Atlanta. Oakland has 11 fewer turnovers, so the expectation is that Oakland is more likely than not to cover the 9 point spread.

Sunday, October 7, 2012

College Football Rankings - Week 7

The 1892 Alabama Football Team
When men were men and hats were hats
Here is my first attempt at creating a betting market power ranking for College Football.  The goal here is to try to reverse engineer an implied power ranking from the point spreads set by Vegas and the offshore sportsbooks.  In a nutshell, I'm trying to figure out what the betting public "thinks".  See my Methodology page for a simple example of how the rankings work (also see here for a detailed NFL example).

Here are some details on the methodology:
  • Home field advantage is assumed to be worth 3.5 points (I derived this from 5 seasons worth of point spreads)
  • Recent games are weighted more than games played early in the season.  The weight used is as follows: weight = 1/(2.8 + days), where days is how many days ago the game in question was.
  • The key output is GPF, which stands for Generic Points Favored.  It's what you would expect the team to be favored by against a Division 1A average opponent at a neutral site.

Wednesday, October 3, 2012

The Turnover Index - Week 5

One of my first posts on this blog was on the topic of NFL Turnover Differential and the Point Spread. In that post, I found that teams with fewer season-to-date defensive turnovers than their opponent tended to cover the spread more often than not.  The presumed reason for this is that:

  1. Turnovers are more random than the betting public realizes, and so the point spread "overcorrects" for prior game turnover performance.
  2. To the extent that turnovers are not random, they are more a function of the offense.  Interceptions are more "thrown" than caught and fumbles are more "lost" than forced.

Saturday, September 29, 2012

NFL Rankings - Now Updated Daily

There is no shortage of NFL power rankings from which to choose these days, of varying quality and usefulness.  One unique feature of the rankings I publish here is that they can actually be updated daily.  Standard stat-based NFL rankings need to wait for new data to be updated, with most of that data coming on Sunday.  Since my rankings are based on the point spreads, which update continuously (much in the same way that stock prices do), I can run these daily in an effort to take the pulse of the market.

The purpose of this post is to announce that the daily rankings for the NFL are now up and running.  A few notes on the ranking table:

Sunday, September 23, 2012

NFL Team Rankings - Week 3

As promised yesterday, here are the NFL Team Rankings as of Week 3 (with lines updated as of Sunday morning).  For those of you new to this site, these rankings are an attempt to reverse engineer an implied power ranking from the Vegas point spreads.  If you're gambling on the NFL, think of this as competitive intelligence, in that your competition is Vegas, the sportsbooks, and the combined wisdom of the betting public.  The goal of this site is to provide greater clarity on what may be driving how point spreads (and over/unders) for each game are being set.