I guess it's that time of year to make my prediction on the Super Bowl, although I wasn't very accurate last year (I might have been too eager to see the Patriots lose). I will say that I am happy this year's game seems to be competitive and that none of the teams are from the New England area (three straight years of the robot Brady and the scowling Belichick would be too much). Most betting sites are favoring Kansas City (KC) +1.5 over San Francisco (SF), but if you look only at season averages, SF should win the game.
Given above in Fig 1 is a box plot of the points allowed (defense) and points scored (offense) for both KC and SF over the 2019-20 NFL season. The KC defense allows ~1.3 more points on average than the SF defense, and the SF offense scores 0.33 points more than the KC offense. When you average points allowed versus points scored for both opposing teams, you come up with a score of 25.17 for SF and 24.36 for KC (Fig 2).
If you've followed any of my NFL posts then you know that most NFL teams play differently when home or away. Both the KC and SF defenses allow the opposing offense to score around 3 less points than you would expect based on their season average, but allow 3 or more points from the opposing offense when playing on the road (the probability that this home and away effect for these two teams isn't real is about a 1 in 20 chance, or p-value = 0.05). Both offenses play better at home (+4 and +4.8 points for SF and KC respectively), but the KC offense tends to score 8.3 more points on the road, while the SF offense performs about as expected (+0.5) based on their season average. If you now adjust for location [Expect Score + (Away Effect x p-value)], you have KC scoring 29.47 points to SF scoring 27.03 points.
If you look at performance for both teams offense and defense over the season for away games (Fig 3), you can see that both the KC and SF defense has seen a decrease in the number of points they've allowed the opposing offense over time, but it seems like this decrease peaked with SF around 10-14 weeks and now both teams' defenses seem to be around the same level of performance. The KC offense, on the other hand, performs almost 8 points better than expected than the SF offense when playing on the road. It is because of this huge difference in offensive performance that I would expect KC to win today.
Besides location, I've found that the number of turnovers and sacks a team gets on defense or gives away on offense has a significant effect on the final score. If we only look at these metrics for away games, then KC has the clear advantage for both. Given in Fig 4 is the average number of turnovers for each NFL team ranked by away games. We can see that the KC defense gets on average 0.5 more turnovers than the SF defense on away games.
If you look at the ability for the defense for both teams to get sacks at away games, then KC clearly has the advantage (Fig 5). When playing on the road, the KC defense on average gets at least one more sack than the SF defense (3.6 compared to a 2.4 sack average). If you look at both the average number of turnovers and sacks over the course of the season for away games (Fig 6), you can see that the KC defense always has a higher average number than SF for both turnovers and sacks. (It is probably not an accident that the increase in sacks for the SF defense occurs around the same time we observe a dip in the allowed score versus the expected (week 10-14). We also see that the KC offense on average is less likely than SF to give up turnovers or sacks, with Garoppolo over 2 times as likely to be sacked than Mahomes while playing on the road.
Based on location and the relative turnover/sack ratios for each team, I calculated that KC would score ~30 points while SF would score ~26 points, giving KC a +4 point advantage. The above analysis assumes an average away game, but I'm sure the Super Bowl is not average, although it is definitely not a comfortable home game where the players can sleep in their own bed the next before their game. I do think that the SF defense plays much better at home and they won't be able to contain a KC offense that is exploding right now, especially when playing on the road. If it's a high scoring game, then the advantage is obviously to KC, but if SF coach Shanahan can find a way to get more turnovers and sacks, they may stand a chance. Either way, it should be an interesting game to watch.
-Michael Edwards, Bioinfo Solutions LLC
*See previous blog post for an explanation on the performance metric
**Microsoft Excel for Mac ver.16.2 and JMP Pro ver.15 were used to collect, analyze and display all NFL data.