The Mathlete's Dilemma: NFL Season Projections

The NFL season is rapidly approaching as teams are all gong to have their first preseason games this weekend. As such, its time to start looking at which teams will have bounce-back years, who will fall from the top and who will more or less where they've been. Projecting the NFL though is actually much more difficult than college football. This comes mostly from the fact that at the college level, recruiting and the prestige of the team both play major roles and tend to be pretty well known before you get going. A team with a declining recruitment and high attrition is likely going to face issues down the line.

The NFL though sees quick turn arounds both up and down on a yearly basis. The Jags went from a 3 win nobody to and AFC title hopeful in a matter of season. The Packers went from Super Bowl contender to missing the playoffs thanks to an injury to Aaron Rodgers. This past year though felt extremely odd with a pretty high number of starters missing significant time to injury, and several players regressing in their play from previous seasons.

There are luckily a few tools to look at that can be insightful The first is Pythagorean wins. This is a basicac methodology comparing a teams points surrendered to their points scored. Its one of the better methods we have to draw to identify teams that might break out or will face a fall back.

Team	Pts Scored	Pts Allowed	2017 Wins	Pythagorean Wins	Projection
Arizona	295	361	8	6.121667609	Decline
Atlanta	353	315	10	9.073219929	Decline
Baltimore	395	303	9	10.43407068	Improve
Buffalo	302	359	9	6.383506934	Decline
Carolina	363	327	11	8.985091809	Decline
Chicago	264	320	5	6.207261271	Improve
Cincinnati	290	349	7	6.27203956	Decline
Cleveland	234	410	0	3.348716789	Improve
Dallas	354	332	9	8.607085858	Decline
Denver	289	382	5	5.447465578	Improve
Detroit	410	376	9	8.817798162	Decline
Green Bay	320	384	7	6.297991742	Decline
Houston	338	436	4	5.657083749	Improve
Indianapolis	263	404	4	4.248708158	Improve
Jacksonville	417	268	10	11.84555917	Improve
Kansas City	415	339	10	9.881698326	Decline
LA Chargers	355	272	9	10.44406904	Improve
LA Rams	478	329	11	11.32677694	Improve
Miami	281	393	6	4.977440566	Decline
Minnesota	382	252	13	11.65243707	Decline
New England	458	296	13	11.804685	Decline
New Orleans	448	326	11	10.87875129	Decline
NY Giants	246	388	3	4.056257003	Improve
NY Jets	298	382	5	5.711535996	Improve
Oakland	301	373	6	6.009513373	Improve
Philadelphia	457	295	13	11.81346958	Decline
Pittsburgh	406	308	13	10.52917484	Decline
San Francisco	331	383	9	6.630333593	Decline
Seattle	366	332	6	8.920193743	Improve
Tampa Bay	335	382	5	6.765315773	Improve
Tennessee	334	356	9	7.396422456	Decline
Washington	342	388	7	6.812513678	Decline

The table above lists every team's projection. Teams like Pittsburgh and Philly both overperformed for their 13 wins, so it might be more realistic to have them lose a game or two. Liekwise, the Browns are pretty unlikely to go 0-16 again. Brett Lieblich has his own adjusted Pythagorean formula, where he removes garbage time points to help lessen some of the noise of blowouts. For my own predictions I do a similar thing to keep the program from over evaluating a loss.

The folks over at Footballoutsiders also have some research into defensive improvements year to year that coincides with third down efficicency. Third down efficiency itself is a useful tool, having a roughly .51 r^2 value to wins. Basically defeneses that perform well on 1st and second down but a re poor on third downs tend to improve the next year. From below, we can see teams like Jax, Ten, and Buf all might look forward to a bit of defensive improvement next year. Jacksonvilel will be tough to get much better, but the Titans and Bills both have a good shot to improve. New England's defense might also be one to look at a sleeper for a unit that could quickly be back to being near the top. On the flip side, Cin and Chi might expect a bit more regression this season. And if you're a a Bucs fan, this defense might get really bad this year.

Team	1st and 2nd Down vs League Average	3rd Down vs League Average	1st and 2nd Down vs 3rd Down
ARI	-0.72	-0.42	0.3
ATL	-0.2	-0.3	-0.1
BAL	-0.32	-0.51	-0.19
BUF	-0.52	0.04	0.56
CAR	0.01	-0.06	-0.07
CHI	0.6	-0.55	-1.15
CIN	0.57	-0.8	-1.37
CLE	-0.11	-0.4	-0.29
DAL	0.26	-0.47	-0.73
DEN	-0.64	-0.47	0.17
DET	-0.31	0.11	0.42
GNB	0.15	0.07	-0.08
HOU	0.2	0.26	0.06
IND	1.28	0.04	-1.24
JAX	-1.82	-0.56	1.26
KAN	0.11	0.18	0.07
LAC	0.16	-0.32	-0.48
LAR	0.49	-0.24	-0.73
MIA	0.22	-0.02	-0.24
MIN	-1.6	-0.65	0.95
NOR	0.52	-0.21	-0.73
NWE	-0.16	0.61	0.77
NYG	1.23	0.05	-1.18
NYJ	0.5	-0.17	-0.67
OAK	0.65	0.08	-0.57
PHI	0.21	-0.57	-0.78
PIT	-0.3	-0.08	0.22
SEA	0.13	-0.62	-0.75
SFO	-0.45	-0.1	0.35
TAM	1.69	0.27	-1.42
TEN	-1.41	-0.13	1.28
WAS	-0.62	0.07	-0.69

On the program front, its classification mostly deals with team trends. I use a weighted average of a team's past 4 seasons to get some idea of who should be where. This is that prediction for the 2018 NFL season. I'm not certain that I like it as much, since its got a high variance and this upcoming season seems to have some huge questions for teams at the top of the charts, but its a starting point.

Team

New England Patriots

Minnesota Vikings

Kansas City Chiefs

Pittsburgh Steelers

Atlanta Falcons

Seattle Seahawks

Carolina Panthers

New Orleans Saints

Philadelphia Eagles

Dallas Cowboys

Baltimore Ravens