Predicting Season Win Totals, part I

In the past when we've run this study we've used a data set going back to 1990, but in light of the fact that the NFL is not the same today as it was in the early 90's (salary cap connundrums anyone?), we've elected to tighten the sample to look only at seasons from 1996-2004. This still gives a healthy nine year data set but sides with the present day over the past when teams were more consistent from year to year.

Now during this period there have been no teams with either a 0-16 or 16-0 record, so consequently we will look at each prior season win total between 1 and 15 (ties are counted as a half, but rounded down for this research).

The data below includes an average change in wins (prior season to current season), and the percentage at which teams surpassed win milestones. We've also highlighted the highest win point at which teams in that category reached a milestone 50% of the time.

[Data includes 1996 to 2004 seasons]

Prior Season Wins	# of cases	Ave Net	4+ Wins	5+ Wins	6+ Wins	7+ Wins	8+ Wins	9+ Wins	10+ Wins	11+ Wins	12+ Wins
1	3	+6.0	100 %	100 %	66 %	66 %	33 %	33 %	0 %	0 %	0 %
2	3	+2.7	33 %	33 %	33 %	33 %	33 %	0 %	0 %	0 %	0 %
3	12	+3.3	83 %	75 %	58 %	50 %	25 %	16 %	16 %	8 %	8 %
4	20	+3.0	100 %	90 %	75 %	45 %	30 %	20 %	10 %	10 %	10 %
5	22	+2.2	100 %	77 %	68 %	50 %	45 %	31 %	22 %	18 %	4 %
6	28	+2.0	85 %	85 %	75 %	64 %	64 %	57 %	39 %	14 %	7 %
7	33	+0.2	87 %	75 %	63 %	48 %	48 %	39 %	27 %	18 %	12 %
8	34	-0.1	94 %	88 %	82 %	76 %	58 %	38 %	17 %	17 %	14 %
9	33	-0.5	90 %	90 %	81 %	72 %	63 %	45 %	39 %	27 %	21 %
10	33	-1.7	100 %	93 %	90 %	72 %	63 %	48 %	30 %	15 %	6 %
11	19	-1.4	100 %	94 %	89 %	89 %	78 %	68 %	57 %	42 %	26 %
12	18	-2.4	100 %	94 %	88 %	88 %	72 %	61 %	61 %	33 %	33 %
13	13	-3.3	100 %	92 %	92 %	92 %	69 %	69 %	61 %	38 %	30 %
14	5	-6.2	100 %	100 %	80 %	60 %	20 %	20 %	20 %	20 %	20 %
15	1	-5.0	100 %	100 %	100 %	100 %	100 %	100 %	100 %	0 %	0 %

- Records only include regular season wins
- Green highlight is highest win number for 50%+ of teams with that prior year win total

ANALYSIS: Obviously the temptation is to take these numbers a little too literally and as indicative of how "any team" in a given category will do the following season (eg assuming that any 6-10 team has a 64% chance of getting eight wins or more). This would be a bad idea in our humble opinion, since the table above is looking at things blindly, without regard for the "season wins" line set on a team -- in other words while those 6-10 teams may as a group get to 8+ wins most of the time, it's highly likely that the ones which were forecast to do poorly (eg having a season over/under line of say 5.5 wins) don't meet the percentage, whereas teams forecast to do well (say 8.5 wins) may be even higher than the class norm.

Having said that, there are some obviously strong historical results to note -- the vast majority of NFL teams get at least five wins regardless of their prior season performance, while few get to 11+ even with a superior record in the previous year.

The general message to take away from the above is that teams coming off strong seasons tend to have a hard time matching expectations and teams that had a rough year often turn it around pretty quickly.

Ah, but let's take the next step anyway, with the caveat that this is just one aspect of a good season win analysis...

Review of NFL Teams by record class

15-1 record in 2004

Pittsburgh (9.5 ov-145, un+115)

There's only been one 15-1 team in the past nine years before these Steelers (and that was the '98 Vikings who were 10-6 the following year after the excruciating loss to the Falcons in the NFC Championship game). Can you say "no sample size to speak of" !

14-2 records in 2004

New England (11 ov-120, un-110)

Well it's not often we see 11 win over/under lines on teams in this era, but the Pats get this honor heading into 2005. History isn't promising since only one of five prior 14-2 teams made it past even 7 wins...but of course the team that did go over the season win line was these very same Patriots. Hmmm.

13-3 records in 2004

Philadelphia (11.5 ov-105, un-125)

Bodog doesn't think too much of the NFC, giving the Eagles a whopping 11.5 line even with Terrell Owens still a holdout. The 'lucky thirteen' teams have actually done pretty well the next year as a group, but heck only 30% have managed to get 12+ wins, so there's an immediate lean to the 'under' here.

12-4 records in 2004

Indianapolis (11.5 ov-130, un+100), San Diego (8 ov-130, un+100)

Another huge line set for those offensive juggernaut Colts, but only 33% of 12-win prior year teams have repeated to that level or better, so think 'under' initially. On the other hand the Chargers are deemed to be prime "bounce" candidates (and we'll be looking at the facts on whether the bounce is real in the NFL in a later article) but 61% of 12-4 teams have got to 9-7 or better on the next go round, so begin with an 'over' perspective with San Diego.

11-5 records in 2004

Atlanta (9.5 ov-105, un-125)

Ah, it's Michael Vick scrutiny time as the debate rages on whether he is over-rated or a powerful "lead his team to wins" superstar. A respectable 57% of 11-5 teams have managed to get 10+ the next season so 'over' has a slim edge on this first look.

10-6 records in 2004

New York Jets (9.5 ov-105, un-125), Denver (8.5 ov-145, un+115), Green Bay (8 ov-105, un-125)

Ten win teams have struggled oftentimes, with only 30% making it to the same level or better, and only 15% getting to 11+ (for contrast 5-win prior year teams have won 11+ on 18% of subsequent seasons!). Consequently linemakers are often downgrading these clubs and Bodog has them all below their previous season win total. Of interest is that they have the Packers at a modest 8 win number, so perhaps the 'Favre will finally miss some games' feeling is strong.

9-7 records in 2004

Baltimore (10 ov-110, un-120), Buffalo (8.5 ov+105, un-135), Jacksonville (8.5 ov-125, un-105), Seattle (8.5 ov-140, un+110)

Nine win teams seem to go in both directions -- some rising up to new heights, others tumbling into mediocrity. Only 27% of such teams have earned 11+ wins which makes Baltimore a little suspect at 10 wins to push the bet. It's very close to 50/50 though on how teams have fallen around an 8.5 number.

8-8 records in 2004

Minnesota (9.5 ov-115, un-115), St. Louis (8.5 ov-115, un-115), Cincinnati (8 ov-130, un+100), New Orleans (7.5 ov-145, un+115)

Can Mike Tice get the Vikings to produce for a full season? The historical odds would look slim given only 17% of 8-8 teams have managed 10+ the next year. Indeed only 38% of the 'even-stevens' of the NFL have improved to 9-7 or better.

7-9 records in 2004

Kansas City (9.5 ov+110, un-140), Carolina (9 ov-150, un+120), Houston (7.5 ov-110, un-120)

Both Kansas City and Carolina must have been taking some action over at Bodog since the money-lines are moving. It's an 'over' surge for the Panthers, but an 'under' steam for the Chiefs. Given that only 39% of 7-9 teams have managed to get to 9 or more the next year you'd think the Carolina backers are being a little hasty, yet there is that bounce effect again in play which we'll talk about another day.

6-10 records in 2004

Dallas (8.5 ov-110, un-120), Detroit (8.5 ov-105, un-125), Washington (7.5 ov-130, un+100), Arizona (7.5 ov-125, un-105), New York Giants (6.5 ov-140, un+110)

The most common record in the NFL last season? The 6-10 mark shared by the five team above. If you like big name coaches (whether to back them or go against them) then this is your group, since we have Green, Parcells and Gibbs among the bunch. Interestingly 6-10 teams have often made good turnarounds with 57% reaching 9-7 or better. So 'over' might be the best initial thought here and the NYG line suggests a vision of hard times for Eli Manning in the eyes of the Bodog linemakers.

5-11 records in 2004

Oakland (8 ov-105, un-125), Tampa Bay (7 ov-125, un-105), Chicago (7 ov-105, un-125), Tennessee (6.5 ov-105, un-125)

If there's one wager we'd like to take back from last season it was probably the "over" call on the Bucs (we did go 3-2 overall with season win picks, but perhaps the TMW founder's bias was at work with the unfathomable TB pick). This year it's the Raiders getting some hype, thanks no doubt to the arrival of Randy Moss. Let's face it though, no NFL team is happy with a five win season...something went seriously wrong. History shows 45% of 5-11 sides have come back at 8-8 or better, but that still makes Oakland at 8 to push a questionable starting point.

4-12 records in 2004

Miami (6 ov-125, un-105), Cleveland (4.5 ov-140, un+110)

Exhibit A on a messed up season before things even began took place in Miami with the Ricky Williams fiasco leading into a miserable overall campaign. New coach, Ricky back and who knows whether the bounce will be strong (maybe Shaq can help out!). Four win teams have got to 5-11 or better on 90% of the time though which makes the Cleveland line a little shocking, and understandably there's some significant 'over' bets moving the money line up there.

2-14 records in 2004

San Francisco (4.5 ov-115, un-115)

Then we have the Niners. Unfortunately 2-14 years are so rare in today's NFL that we don't have enough data in this kind of view to make any notable comment.

Again, all of this was just an exercise and should not be construed as a full scale analysis of how to play the season win numbers, for that you'll need to delve deeper...and we will! More articles coming soon on this subject.

Also see:
Season Win Bets Research, part II
NFL Power Rankings
NFL Futures Betting Odds