Week 15. Fantasy Football Playoffs. Do I start Josh Freeman over Aaron Rodgers? I already know the immediate answer. No. Absolutely, completely, without a doubt, No. You are an idiot for even thinking otherwise. It’s Aaron Rodgers! He’s the best QB in the game! You go with the horse that got you there. Rodgers is a stud and what has Josh Freeman ever done? I heard it all, a resounding ‘Yes’ to Aaron Rodgers … and I started Josh Freeman instead.
Everyone I had talked to hadn’t looked at the numbers for these two quarterbacks this year, and practically no one considered the defenses both these guys went up against. They heard the name Aaron Rodgers, and not only assumed he had been vastly superior to Josh Freeman this year, but also assumed that fantasy scoring systems reflect the true skill level of a player. The reality is that Aaron Rodgers is probably the best skilled quarterback in the NFL, but when placed within the scoring system of fantasy football, Josh Freeman certainly becomes comparable. Side Note: When considered in terms of winning NFL games, Eli Manning tops all of them and is the perfect example of how real world winning doesn’t necessarily translate to fantasy sports winning.
For the season, Rodgers ranked 5th in scoring while Freeman checked in at 12th overall, with the difference between them being 2.62 fantasy points per game (standard ESPN scoring system). In a 12 team league, both quarterbacks would be considered QB1s based on the points they put up, so there is a conversation to be had here. Josh Freeman also had a lower standard deviation in his scoring, which means that he was more consistent in achieving 16.23 points per game than Rodgers was in achieving 18.85. In other words, Freeman statistical performances were less risky.
When you begin to look at the more recent history, the QB comparisons get a lot closer than you’d expect. From Week 5 to Week 14 (9 games), Freeman averaged 18.87 fantasy points per game (fppg) with a 5.56 standard deviation, while Rodgers averaged 19.67 fppg with a 9.00 standard deviation (mind you I’m not cherry picking time frames here; Rodgers scoring numbers went up and it included both players season high point outputs). When looking at this time frame, we are now talking about one quarterback who registers less than a point more per game versus one that is a whole lot less risky. Uh oh, looks like we have a quarterback controversy brewing.
The last thing I looked at were the match-ups, and Freeman won big time in that respect. New Orleans had given up the most fantasy points to opposing QBs in the league, to the tune of 20 fppg, which represented a +1.94 deviation from the league average of 14.84 fppg. They had given up better than average performances in 11 out of 13 weeks to quarterbacks including the likes of Phillip Rivers, Michael Vick, Carson Palmer, and Colin Kaepernick. The Bears, Rodgers’ Week 15 opponent, was the exact opposite, giving up a meer 9.8 fppg (-1.89 standard deviation from league average), best in the league which included a 10 point game to Rodgers back in Week 2. It is also worth nothing that Freeman had his season high 29 against his Week 15 opponent, the New Orleans Saints.
I used these statistics to build a very simple model to give me a sense of projected fantasy points for Week 15 (actually I considered a much more complicated statistical model normalizing all of these stats based on who actually played who, but I’ll leave that out because I don’t have a lot of faith that this math illiterate county can even follow this simple model based on averages and deviations). I included the four other quarterback controversies I was considering for Week 15, which was the reality of all four leagues I was in this year:
Each color band represents the choices I had to make, and in each case, the higher seasonal ranking QB was listed first. All four scenarios represent the exact same situation as Rodgers/Freeman, where we have a well established, well-known stud versus a relatively unknown and unproven commodity. Using the simple model, in all cases the projected points suggested that you should bench your stud in favor of his back-up (the absolute points don’t matter in the comparison, and in the more sophisticated model, Rodgers was more in the 12 – 15 fppg range with Freeman in the 21 – 24 fppg range). Reality seemed to have beared that out as well, since in 75% of the scenarios, QB2 outscored QB1.
So yes, I stand by my decision to start Freeman over Rodgers and with this type of analysis, I will get it right more often than not (hi Bill Belichick). In the particular league where I had the Freeman/Rodgers choice, I lost by less than 13 points, so mostly everyone will point to that example as the reason to always start your studs. However, the reality is that setting your line-up to maximize your points and what actually happens in the game are two separate and unconnected things. The Packers play-calling in game situations doesn’t dictate they should throw to the endzone because Aaron Rodgers is my starting fantasy QB. Above and beyond that, I might not have even won with Aaron Rodgers if I happened to run up against Russell Wilson (especially a downward single deviation that Rodgers has put up in past fantasy playoffs, in particular Week 14 against KC in 2011, Week 14 and 17 in 2010, and Week 16 and 17 in 2009).
In fact, it’s not even the Freeman/Rodgers decision that lost that game for me, it’s how far from the norm Freeman AND the New Orleans Defense deviated from their histories! If the New Orleans played to the league average of 14.5 points given up, which was one standard deviation on the PLUS side for them, I would have won. Had Josh Freeman performed within one DOWNWARD standard deviation of his stats over his last nine games, he would have had 13 points and I would have won. If we could run the Josh Freeman versus New Orleans Saints scenario a million times, 84% of the time Freeman would have put up enough fantasy points for me to win. What we saw on Sunday was the perfect storm, a statistical aberration which resulted in a two and half standard deviation from the norm, a .5% scenario that no one is immune to. Frankly, you don’t set your fantasy team, or make most decisions in life, based on .5% outcomes. You buy insurance for them (or against them).
For the most part, these simple statistical predictions show themselves in the real world, given you have enough scenarios to examine. In my limited fantasy world, I had four leagues where this case existed, and in three of them it proved to be correct. In a broader history, a 36 – 17 regular season record, three first round byes, and two trips to the finals is probably better than most people who play fantasy football can represent. If I had brazenly played Rodgers this week and snuck out the win, it would be statistically obvious to play him against the 20th worst pass defense in the league in Week 16. The point is that sometimes you have to go Saracen when the analysis says so and put your ultra skilled J.D. McCoy on the bench even if it is in the playoffs, (*SPOILER ALERT*) … and sometimes you still just fall short.