Could midpoints per game be the future of fantasy football analysis?


The competition in fantasy football is constantly changing, but the industry always uses an obscure note throughout the season for how players end up. You’ve heard the arguments: “So-and-so finished as WR23 last year and I could see him finish in the top 15 this year.” The small step forward was the push for the points-per-game average, commonly abbreviated as PPG. Viewing a player’s “weekly” production is useful for making weekly roster decisions for a season’s leagues as well as for daily fantasy games. However, there is still a better way. This way it is midpoints per game (MPPG will work as an abbreviation). If you can’t remember your college math work, don’t worry, class will begin!

Why finishing all season is not useful

A player’s point total is influenced by too many variables to be useful in predicting how a player will perform in the future. A cumulative number such as 150 PPR points (the typical PPR total for the RB30) gives no idea how this player arrived at that number. There is no allowance for variance and with no games played next to the point total, there is little to create context. Take for example Nick Chubb’s 2020 season. He finished with 207.7 PPR points and was RB11. He was in a place of Kareem Hunt who obtained 218.5 PPR points. The missing piece of context is that Chubb played four games less than Hunt.

Over 14,000 potential candidates for this year and beyond

Why the average points per game is better, but not yet sufficient

Keeping our example above, we can calculate the average points per game (APPG) of Chubb and Hunt. Since Chubb has played fewer games and had a similar point total, his APPG (17.3) is significantly higher than Hunt’s 13.7. This is useful because it provides an estimate of what may be the expected release for Chubb in any given week.

The problem is, this estimate is extremely inaccurate for fantasy football. One problem lies in the mathematics used to calculate an average. The APPG is simply the total number of points over the year (or some other cumulative period) divided by the number of games played. So, by using APPG, you did not break away from the player’s season total.

The other math problem is that an average assumes a normal distribution (think college science bell curve) to be precise. In a normal distribution, the mean and median will be the same since there is equal “weight” on both sides of the mean and the median. However, when the data is not normal or “biased” in one direction, the mean and median will have different values ​​depending on the direction and magnitude of the skew.

To further prove the flaws of using the mean in largely non-normal data sets, statisticians and college math teachers are telling a “joke.” A guy and four of his best friends are at the bar. They all earn roughly the same amount of annual income with an average of $ 50,000 per year. For some reason, Bill Gates, who made over $ 7 billion in 2019, walks through the door. Have a good day ! The guy and his friends have all become billionaires. . . on average. It’s a bad joke, even by daddy-joke standards, but it’s an attempt to explore the impact of introducing a data point that turns a normal curve into a non-normal curve.

Seattle Seahawks wide receiver Tyler Lockett's points-per-game average was 16.8, but his median points-per-game was 10.9 during the 2020 fantasy football season.

Seattle Seahawks wide receiver Tyler Lockett’s points-per-game average was 16.8, but his median points-per-game was 10.9 during the 2020 fantasy football season.

Why midpoints per game are better

As shown in the image above, the median is closer to the “bump” in each case and the mean (mean) remains relatively unchanged. Even in the bad joke above, the median income would still be much more representative of the income of the guy and his friends as opposed to the average income which was the highlight of the joke. When we take a look at these images above, it should be clear that the median is a more precise estimate of where the majority of values ​​fall in a data set. To prove this point, the following image is a histogram of Tyler Lockett’s PPR points for each week of the 2020 season. Lockett’s histogram is similar to the third distribution in the image above. As detailed in the third distribution, the mean is above the median despite the fact that the majority of values ​​are at the lower end.


In the entire 2020 season, Lockett’s APPG was 16.8 despite only six games above that number. That’s a surprisingly low number of 37.5% games above his points-per-game average. Lockett’s MPPG was 10.9, and by definition half of his games (eight) were above that number. Lockett’s 53-point game was more than likely a winner. However, if you knew that 50% of a player’s games would produce 10.9 PPR points or less per game, how interested would you be in having that player on your fantasy roster? For reference, 10.9 PPR points was WR33 at best, WR46 at worst (twice), and would have been WR36 or worse 16 out of 17 weeks in the 2020 season. Lockett’s 16.6 APPG ranks 12th for 2020, but it didn’t produce a weekly score that would make fantasy owners happy outside of three really big weeks. Lockett’s MPPG for the typical fantasy season (weeks 1-16) was 10.2 and was tied for 48th; his APPG was 17th with 14.4.

While Lockett is an extreme example of a player being overvalued by APPG, you can use MPPG to find players who are dumped by APPG. Anytime a player’s median is above their average, they score more points more consistently. Brandon Aiyuk is a prime example. Although Aiyuk only played 12 games, he surpassed his average in seven of those games; just under 60% of the time. His average points per game from Week 1 to Week 16 was 15.4, barely outside the typical WR2 weekly range. Still, his MMPG for that period was 18.6, more than three points higher, and places him squarely in the middle of the WR2 range. Comparing the weeks in which Lockett and Aiyuk both played, Aiyuk scored more points than Lockett six out of 11 times even though his APPG was almost a point and a half short.

The median points per game is the next step in the evolution of understanding football data. It breaks completely with year-end totals without context in a way that APPG just can’t. In addition, it more accurately reflects the weight of a player’s scores, allowing more accurate predictions.

For more on midpoints per game, be sure to follow me on Twitter as I’ll be working the entire offseason to get the word out as much as possible.


About Author

Leave A Reply