I know it sounds crazy, but some teams play better in enemy territory than they do at home. For example, when Tony Romo was the quarterback for the Dallas Cowboys, they statistically performed better (scored more points than expected) on the road than they did at their 1.2 billion dollar stadium in Arlington, TX. I think the 2017-18 Los Angeles Rams offense is another rare example of a team doing worse in front of the home crowd.
I believe most sports analytics are ultimately based on the amount of points a team scores and gives away versus the same metrics for whomever they are playing. The final score of the game should be the average of whatever one team's offense typically scores versus the amount of points the opposing team's defense typically gives up, and vice versa. Given below is a graph of the average amount of points the Atlanta Falcons and the L.A. Rams offense and defense have scored and allowed respectively over this NFL season.
At the end of the season, the defenses for both the Falcons and the Rams are allowing about the same amount of points scored for the opposing team (~20 points). The offenses for both teams are different, with the Falcons normally scoring ~23 points and the Rams expected to score ~29. Given these results, you'd expect the Rams to win the game with a final score of ~24.5 to Atlanta's score of ~21.5 . The only problem with this system to predict the final score is that the average amount of points a team scores or gives away is dependent on the quality of teams they face during the year. A better way to measure performance would be to calculate the average amount of points THEY SHOULD SCORE based on the average amount of points their opponents have scored and allowed up to that point in the season. We would expect the offense with the highest scoring average (L.A. Rams at 28.95 points/game) to score more points when playing the defense with the highest allowed scoring average (Houston at 26.87 points/game) and fewer points when playing the defense with the lowest allowed scoring average (Minn at 15.95). During the 2017-18 season, the Rams offense scored 33 and 7 points on Houston and Minn. respectively.
What I've observed over the years that I've been collecting and analyzing NFL data is that teams play different whether they are home or away. As you'd expect, most teams either play better at home or don't have any noticeable differences in performance wherever they play. If you look at the predicted vs. observed points allowed by the defense of both the Falcons (red) and the Rams (blue), you'll notice that the Rams allow fewer points then expected at home (-3) then away (+2), while the Falcons defense is more consistent whether home or away (approx. -1 and -2 respectively). The offense for the Falcons follows this same trend, with the team scoring approximately 2 points more then expected when playing at home, and about -1 less then expected when playing on the road. On the contrary, the Rams offense typically scores what you'd expect when they play on the road but scores about 3 points less then what would be expected when playing at home.
The above graph gives the p-value for the significance in the differences in the predicted vs. observed scores for home and away (ttest) for both the defense and offense of both teams. Although by week 18 none of the curves reach the the typical P<0.05 standard of significance, they are all below a 0.30 probability you'd see this pattern by random chance, so I'd say there is evidence of a home and away effect with both teams and sides. [As a side note, all the measured p-values for performance differences in home vs. away are below 0.50 for both offense and defense for all 32 teams in the NFL, suggesting that there is indeed a location bias].
Another thing I've noticed that affects the outcome of the game, besides whether the team is playing at home or away, is the number of turnovers and sacks a team gets or gives up. Given above is a graph of the average amount of turnovers the defense obtained or the offense gave up for both the Falcons and Rams subdivided by home and away averages. If you look at the average number of turnovers for the Rams defense at home (blue solid line) vs. the Falcons defense away (red dashed line), they are about the same at ~1.25 turnovers/game. We do see a difference when we look at how many turnovers the offenses are giving up, with the Rams averaging two turnovers at home, while Atlanta tends to turn the ball over around 1 time regardless of whether they're playing at home or on the road.
I see a similar trend when looking at the amount of sacks the offense and defense give and take for both teams (above). Again, the average amount of sacks the Atlanta defense gets at away games vs. the amount of sacks the Rams defense gets at home is about the same (~2.3 sacks), but the Rams QB should get sacked about one more additional time at home then the QB of Atlanta at the same game. It's odd that both the number of turnovers and sacks are higher for the Rams offense at home then away, and it is likely that this difference is contributing to their offense scoring less points then you would expect at home.
When I figure in all the above for the final score, I have Atlanta winning with ~19.3 points and the Rams losing with ~18.8 points. Given that Atlanta are 6 point underdogs, I'd say they have a very good chance of beating the spread today but only because for some unknown reason, the Rams offense plays worse in front of their home crowd.
-Michael Edwards, Bioinfo Solutions LLC
*Microsoft Excel for Mac 2011 and JMP Pro ver.13 were used to collect, analyze and display NFL data.