Thursday, January 12, 2012

Divisional Round Playoff Preview


Divisional Round Playoff Preview

New Orleans Saints @ San Francisco Forty-Niners

Saturday January 14, 2012 – 430pm EDT

For the first time ever I am going to publicly handicap this game using all the tools I analyze on a weekly basis to give everyone not only a feel for my style and the numbers I use, but also to give everyone a detailed breakdown of each of these teams that can hopefully be used to assist in your personal handicapping of this game.  I will lay out by # each model, database, or piece of information I utilize, provide some details behind each, and also cover some high level trends and historical items I may use when breaking NFL games down. 

1.       UNITS BREAKDOWN: in this database I breakdown each team’s offense and defense into four buckets: rushing, passing, miscellaneous and overall.  I use 4 stats for rushing, 7 stats for passing, 5 stats for miscellaneous (these are stats that cannot directly be related to either rushing or passing such as first downs, time of possession, etc…), and the combination of all 16 stats for total offense and defense.  In addition to looking at the stats I look at specific SOS categories for each bucket – meaning I break down each team’s opponent’s strength in each of these 3 areas by themselves, and in total.  In this game we have a classic offense vs. defense matchup as the Saints come in with the #1 offense while SF has the #2 defense – and both attained those rankings vs. #31 schedules (SF faced the 2nd easiest schedule vs. offenses while NO faced the 2nd easiest schedule of defenses on the entire season).  Not only is that virtually even, the buckets are also almost identical – the only area with > 3 spot variance is the Saints rushing offense that is ranked #7 vs. the Niners rushing defense that is ranked #1 – both vs. the #32 schedule.  On the other side of the ball SF has an edge as they check in at #16 offensively vs. #9 SOS, while NO’s defense is ranked #24 vs. #18 SOS.  Ranking wise the dispersion is almost equal across all 3 buckets, but the largest SF edge comes in rushing where they are #15/#7 SOS, while NO is #23/#20 SOS.  Edge: San Francisco driven by their rushing attack.

2.       STAT RANKINGS: in this database I take where each team’s rank in the statistical categories I track and compare offense to the opponent’s defense for both teams – I sum up the edges for each team and whoever is plus has an advantage.  Basically, this is just a more granular version of the Units breakdown and it allows me to focus on specific stats and areas that are driving an advantage for one team or another.  This database has a very small SF edge mostly due to a slightly tougher SOS as the two teams are nearly identical in total rankings.  One key number that stands out in this model however is the TO margin – SF led the league at +28 on the season, while NO checked in at -3.  Edge: San Francisco very small driven by SOS and TO Margin.

3.       REGRESSION: in this model I back-tested 10 years of data on every NFL regular season game that was played, using the 16 stats I track game by game, and compared those using Excel to a team’s points scored.  Excel responded with a formula that estimates the approximate weighting each stat has on points scored – I then take an average of each team’s offense vs. the opponents defense and vice versa, apply that forecasted amount for each stat by its corresponding weighting, and sum each team’s points up for a projected margin of victory.  Using this model for this game SF is estimated to win by approximately 0.5 to 1 point.  Edge: San Francisco very small driven by TO margin.

4.       PERFORMANCE RATINGS: anyone that stops by this blog regularly, or reads the Chad Millman column on www.espn.com knows I refer to this model I built on a weekly basis as the projections this feeds me more times than not lead me to picking the right side of a game – which I just realized around Thanksgiving.  As a reminder, this model grades team’s game by game on a scale of 0-160 – there are 160 points per game, so if a team scores 81 they “won” the game according to these ratings.  What’s more, teams that do score 81 or above in this model cover the line 73% of the time – yes, all that needs to be done is project who wins in these ratings and you will win cash 73% of the time.  This model will be a major focus of mine over the summer as baseball season will be a pass for me in order to prepare for football season – starting right here with this model.  I have 3 versions of this model, one uses full season stats, one uses home/away splits, and one uses last three games – for each stat I use I project each team’s performance (i.e. 175 rushing yards), which then gets “graded” based on data over the last ten seasons (i.e. 175 rushing yards = 5 out of a possible 5 points).  Then I sum each team’s offense and defense performance ratings and that figure has been a big piece of my handicapping over the last third of the season – not coincidentally I have not lost one week since realizing the correlation between this model and ATS performance (it would make sense for there to be high correlation with this model and SU performance as obviously the better a team does statistically, the better they will grade out, and thus the more likely they are to win a game SU).  Without divulging the exact ratings this model provides me on this game, I can state NO has the edge in the full season stats and last three games versions; while SF has the edge in the H/A splits version.  In addition to my performance ratings I also break out pts scored vs. pts against for each team’s offense vs. defense in the respective groupings named above, and doing that yields NO -1.8pts using full season data, SF -7pts using home/away splits, and NO -8.8pts using last 3 games.  So at a high level this model is really providing mixed signals as NO is favored in 2 of the 3 in both performance ratings and points scored/against versions.  But since it is not absolute, more analysis is needed looking at the projections for each stat – which will remain for my eyes only.  Edge: New Orleans driven by having a favorable rating grade in 2 of 3 models.

5.       HOME/AWAY GRADES: here I take the average of each team’s offense vs. the opponent’s defense rating (same ratings from pt4), and vice versa, and add the variances up to see who has the edge.  Applying to this game the NO offense vs. the SF defense is a dead even wash-out; but the SF offense checks in at 45.1 in their home games while the NO defense’s average rating in road games was 36.8.  What is interesting here is although SF has an advantage breaking the numbers down as such, SF was rated #5 in home game performance this season vs. the #14 SOS, while NO was rated #3 in road game performance vs. the #16 SOS.  When looking at that a little closer, both the SF defense and the NO offense averaged over 50 rating points per game, which is extremely high.  Taking a look at the correlation/relationship between performance ratings and points scored it would appear although SF and NO both averaged scoring the same amount of points at home and on road respectively, SF performed a lot worse than NO – suggesting some of SF’s points scored were “lucky” and “opportunistic” driven by their strong defense often giving them short fields to work with.  Edge: San Francisco driven by both their high offense rating and the NO low defense rating.

6.       POWER RATINGS: in my personal regular season power ratings this game features #1 New Orleans Saints vs. #3 San Francisco Forty-Niners.  In the first aspect of the RS PR which is driven by my performance ratings, I have NO rated at the top alone, with SF on the next level down.  On the season NO was #2 in my performance stats only trailing HOU (#1 off, #20 def, #31 SOS), while SF was #6 in the performance stats (#15 off, #4 def, #30 SOS) – which all equates in my estimation to SF being one notch below NO.  The second aspect of the power rating relates to ATS performance and although it has a smaller weighting than the performance aspect of the formula, it does make an impact – and in that NO has a small advantage there as well partially based on the fact they started the season in a better position in those ratings than SF did. In the playoff version which takes into account regular season performance such as points scored, points against, and strength of opponents (and strength of their opponents opponents), NO has about a 3pt edge which is just about offset by the home field advantage of SF in this game.  Edge: Even

7.       RECENT PERFORMANCE: another big aspect of handicapping is recognizing not only who has performed best over the course of an entire season, but also who comes in playing the best football as January arrives.  I have a sheet that shows me all 32 teams in the NFL, and their offense/defense/total performance rating by game for all 17 weeks of the season.  I also break this down by shading each team’s home and away games to get a better feel for not only how well a team is currently playing, but also how their home/away splits have been recently and for the entire season.  Wins and losses are good indicators of who is or isn’t playing well, but often it does not tell the entire story like analyzing statistics does – and as mentioned earlier these ratings are about as good as it gets in both the ATS and SU world.  When breaking these two teams down we first notice both teams have been playing well of late – remember 80 is the “breakeven” in these ratings – NO scored triple digits 3 of their last 4 games (only one they did not they reached 97), while SF scored 91 and 103 in their last 2, and reached triple digits in 3 of their last 7.  Triple digit ratings are reached by approx. 4 teams every week.  But one noticeable trait seen in these ratings for certain are the home/away splits definitely favors SF.  Edge: SF small based on the home/away splits; entire season favors NO.

8.       ATS HISTORICAL TRENDS: on another sheet in my package I have the last 4 seasons of ATS performance by team broken into a few buckets: four high level buckets as home fav/dog and away fav/dog.  Within those buckets I break lines into 3 categories: 3 or less, 3.5-6.5, 7+.  When utilizing this analysis for this specific matchup we see SF is a robust 15-6 as a home favorite (including 5-0 this year )– but alas they are an underdog this week for the first time all season and in that spot they are 3-3,  going 2-4 SU.  NO on the other hand is 11-9 as a road favorite going 17-3 SU – this season in this spot they are 3-3 ATS including 4-2 SU.  Hard to reach any conclusions using this data as on one hand NO is 17-3 SU over the last 4 seasons as a road favorite – but that is offset some by the fact they only went 4-2 in that spot this year.  Some specific trends we see are NO are 8-0 L8 as a favorite (a streak that started following their SU loss @ STL earlier this season), 8-0 L8 vs. NFC, and 10-3 L13 meetings (dating back to their days as NFC West divisional rivals); SF is 10-2-1 L13 vs. NFC, 13-3-1 L17 overall, and 12-3-1 L16 on grass.  Again, both teams with lots of ATS hot streaks coming into this game.  Edge: Even

COPYRIGHT: THE SPORTSBOSS 2012

No comments:

Post a Comment