Previous post:

Next post:

Comparing 2014 Vegas Projections to Estimated Wins

by Chase Stuart on March 28, 2014

in Vegas

Three weeks ago, Vegas released the first set of 2014 win totals for all 32 teams. I immediately wondered how those win totals compare to the estimated wins I created based on 2013 DVOA ratings. I tweeted a request for someone to write such an article, and Warren Sharp (@SharpFootball) was kind enough to oblige. Warren runs SharpFootballAnalysis.com, where he provides handicapping analysis.

One other note before I let Warren take over. If you missed the post on estimating team wins using DVOA, I included this disclaimer:

Even Football Outsiders won’t use these [projections] for more than a starting point — their preseason projections will have the customary tweaks for things like teams getting new quarterbacks, injuries (or the lack thereof) in 2013, rookies, offensive line continuity, etc.

Please keep that note in mind. So when you see “Football Outsiders projects Green Bay to win 7.8 games this year,” that’s just shorthand for “Green Bay’s 2013 offensive, defensive, and special teams ratings, when regressed based on historical data, project a 7.8-win season.” I’m sure with a healthy Aaron Rodgers, Football Outsiders expects more than 7.8 wins in 2013, but the regression formula is ignorant of that fact. And now, I’ll let Warren take over.


On March 7, CG Technology, formerly Cantor Gaming, became the first Las Vegas book to set win totals. For eight teams (25%), the win totals were within one half-game of the estimated DVOA projections: The two sources see eye to eye on Washington, Chicago, Cincinnati, Miami, Detroit, Dallas, Cleveland, and the New York Giants.

For 11 teams (34%), Las Vegas was more enthusiastic than DVOA was, i.e., the books projected higher win totals. The biggest outliers here were Green Bay (10 projected wins by CG vs 7.8 by DVOA) and Houston (8.5 vs 6.5). For Green Bay, we can presume that injuries were the biggest reason for the discrepancy: in addition to Rodgers, Randall Cobb, Clay Matthews, and Casey Hayward all missed significant action. As for the Texans, my guess is that Vegas sees the Colts dropping back two wins from 2013, and the AFC South remains pretty poor.  Houston won 10 games in 2011 and 12 games a year ago, and now faces a pretty easy schedule (the rest of the division, Buffalo, Oakland, the AFC North and NFC East; note that last year, Houston had to face the AFC West, NFC West, New England, and Baltimore.) The Texans were also just 2-9 in games decided by seven or fewer points, a trend that is unlikely to continue.

For the remaining 13 teams (41%), Las Vegas projected fewer wins than DVOA. The two standouts in this category were Arizona (7.0 vs 8.6) and St. Louis (6.5 vs 8.0). As we’ll get to later, CG and Football Outsiders are in considerable disagreement about the fortunes of the NFC West teams.

In some cases, the borderline playoff teams are the most interesting to analyze. There were four teams that DVOA had at sub-.500 that the linemakers have in playoff contention: Green Bay, Houston, Pittsburgh (9.0 vs 7.7) and Baltimore (8.5 vs 7.3). Arizona was the only team in the opposite situation, where Las Vegas projects a losing record despite the DVOA estimates pointing towards a winning record.

The table below shows the numbers for all 32 teams. The Packers had 8.5 wins in 2013, Vegas has set Green Bay’s 2014 wins total at 10.0, and DVOA projects the Packers at 7.8 wins. Therefore, Vegas is 2.2 wins higher on Green Bay than DVOA, Vegas is 1.5 wins higher on Green Bay than the Packers’ 2013 result, and DVOA expects 0.7 fewer wins from Green Bay this year.

Team
2013 Wins
Vegas
DVOA Proj
Vegas - DVOA
Vegas - 2013
DVOA - 2013
Green Bay Packers8.5107.82.21.5-0.7
Houston Texans28.56.526.54.5
San Francisco 49ers12119.41.6-1-2.6
Pittsburgh Steelers897.71.31-0.3
Baltimore Ravens88.57.31.20.5-0.7
Indianapolis Colts11981-2-3
New England Patriots121091-2-3
Denver Broncos1311101-2-3
New Orleans Saints119.58.70.8-1.5-2.3
Atlanta Falcons487.20.843.2
Seattle Seahawks131110.50.5-2-2.5
Washington Redskins376.60.443.6
Cincinnati Bengals1198.70.3-2-2.3
Chicago Bears88.58.20.30.50.2
Miami Dolphins87.57.30.2-0.5-0.7
New York Giants776.80.20-0.2
Detroit Lions787.90.110.9
Dallas Cowboys888.1-0.100.1
Cleveland Browns46.56.8-0.32.52.8
San Diego Chargers988.6-0.6-1-0.4
Buffalo Bills66.57.4-0.90.51.4
Carolina Panthers128.59.5-1-3.5-2.5
Philadelphia Eagles108.59.5-1-1.5-0.5
Jacksonville Jaguars44.55.5-10.51.5
Tennessee Titans76.57.6-1.1-0.50.6
Minnesota Vikings5.56.57.7-1.212.2
Tampa Bay Buccaneers46.57.7-1.22.53.7
New York Jets86.57.8-1.3-1.5-0.2
Oakland Raiders456.3-1.312.3
Kansas City Chiefs1189.3-1.3-3-1.7
St. Louis Rams76.58-1.5-0.51
Arizona Cardinals1078.6-1.6-3-1.4

You may have noticed that the DVOA projected win totals stray away from the extremes, which makes sense when basing projections off of a regression formula. For example, DVOA projects only two teams (Denver and Seattle) to reach double digit wins. In fact, the top two teams in Chase’s Pythagenpat records, also Seattle and Denver, are at only 10.0 and 9.7 wins, respectively. Meanwhile, Vegas linemakers project five teams to reach double digit wins in 2014 (Denver, Seattle, San Francisco, New England and Green Bay).

On the low end, neither DVOA nor Pythagenpat project a team to finish with five or fewer wins, with Jacksonville (5.5 in DVOA, 5.9 in Pythagenpat) as the only sub-6.3 win team. Vegas linemakers, meanwhile, project Oakland at 5.0 wins and Jacksonville at 4.5 wins. Keep in mind that these win totals were released prior to the start of free agency, so CG’s win totals do not reflect the Raiders additions of players like Matt Schaub, Justin Tuck, James Jones, LaMarr Woodley, or Austin Howard.

If you compare the projections to the actual 2013 results, there are several teams that both CG and DVOA believe will have much better fortunes in 2014. The biggest improvement will come in Houston. While the 4.5-win increase is DVOA’s largest improvement, Las Vegas believes the Texans will see a 6.5-win increase and finish with a winning record. Four other teams are projected to see +2.5 win improvements by both the linemakers and Football Outsiders: Washington, Atlanta, Cleveland, and Tampa Bay.

On the other end of the spectrum, CG projects a two-game decrease and DVOA a three-game decrease for Denver, New England, and Indianapolis. The linemakers are also far less bullish on Carolina (projected for a 3.5-win decrease relative to 2013), Kansas City and Arizona (both projected for a 3 win decrease).

The NFC West stands out as a sort of battleground between the two sources. Both St. Louis and Arizona are projected much more positively by DVOA, although it’s worth remembering that CG knows that the NFC West teams have to face each other, while DVOA does not. On the other hand, the oddsmakers are higher on both Seattle and San Francisco than DVOA, so perhaps they expect the NFC West heavyweights to absorb those wins and put the Rams and Cardinals in the division basement. Another thing to keep in mind, particularly for St. Louis and Arizona, is that the division is replacing the AFC South with the AFC West.  When the CG lines were first posted, I put together this matrix, which reflects the changes in each team’s schedule for 2014.

According to DVOA, linemakers are overvaluing two rivals in the AFC North, with both Pittsburgh and Baltimore receiving too much respect. That might be a case where the numbers say one thing, and nostalgia says another. Keep in mind that both sources agree on Cincinnati and Cleveland, leaving the wins removed from the non-Ohio teams to be spread elsewhere in the NFL this fall.

It is worth noting that the numbers posted by CG were posted earlier than ever before in Las Vegas. Presumably they wanted the lines posted before the crazy crowds descended on the town for March Madness as a way to attract betting attention (while limiting exposure with low limits). As such, these early season win totals are sure to be batted around, and by the time LVH and other well-known books release their odds this summer, its likely Sharp Football will be revisiting this analysis.

{ 5 comments… read them below or add one }

Sunrise089 March 28, 2014 at 1:38 am

[Insert comment about how Vegas sets lines to encourage even betting on both sides here] ;)

I don’t think that’s generally true, but I always see it accompany these sorts of posts…

Reply

James March 28, 2014 at 11:19 am

I know you said you disagree, but if anyone wants some facts about that…

Vegas doesn’t care at all how much people bet on either side, and if they did care, they’ve been doing a terrible job at balancing bets: “Simmons and Nelson analyzed betting data on 1,008 regular season NFL games on Sportsbook.com from 2009 to 2012. They found the average share of money bet on the favorite was 65 percent.”

So how does Vegas make money if it’s not by balancing betting and pulling in the vig? By setting accurate lines, so favorites only cover 50% of the time: “There Simmons and Nelson found, just as Levitt did, that favorites were about 50 percent likely to beat the spread (413 favorites beat the spread, 415 did not, and 22 were ties).”

So Vegas gets hit hard whenever a lot of favorites win, but they make it all back and more whenever the underdogs cover because the betting is imbalanced. Also, Vegas plays so many games that they are unlikely to suffer a bad streak (law of large numbers and all that), and even if they do they have enough cash reserves to play the long game.

This also makes you think that Vegas should ‘shade’ the line towards the favorite slightly, and it probably should/does a little, but it has to be careful to not allow smart bettors to take advantage of a bad line. So in the end Vegas is best off by creating the most accurate lines they can, and allowing people’s biases to bet on the favorite to profit.

Quotes from here: http://fivethirtyeight.com/features/why-people-bet-on-the-favorite-even-when-the-spread-favors-the-underdog/

Reply

Andrew Healy March 28, 2014 at 4:57 pm

When I saw the numbers, my initial thought was just how good they were. So few opportunities. I’m can’t remember a number like Houston’s that’s 6.5 wins off the previous year. Last year, KC opened at 6.5 and then got bet to 7.5 (juice on the over), I think.

The ones that seemed obvious candidates for opportunities such as Houston, Arizona, and Carolina, all seem to have been adjusted appropriately beyond the numbers in the projections.

Reply

Hobbies April 25, 2014 at 9:45 pm

“Hello to all of you, come sit I will make you something to drink.
Enter this bird sanctuary and colourful birds will greet you.

Hunters use this type of bows mostly for harvesting deer.

Reply

converse all star July 7, 2014 at 12:10 am

My programmer is trying to convince me to move to .net from
PHP. I have always disliked the idea because of the costs.
But he’s tryiong none the less. I’ve been using WordPress on several websites for about
a year and am anxious about switching to another platform.

I have heard excellent things about blogengine.net.
Is there a way I can import all my wordpress content into it?
Any help would be greatly appreciated!

Here is my weblog; converse all star

Reply

Leave a Comment

You can use these HTML tags and attributes: <strong> <em> <pre> <code> <a href="" title="">

{ 2 trackbacks }