What is the Consistency Score?

How many times have you heard the term “boom or bust”? Or worse yet, how many seasons has your fantasy team been one of the highest-scoring teams in the league, yet you miss the fantasy playoffs? I’ve used this to develop the Consistency Score.

Dominating one week, only to lose a 64 to 65 nail-biter the next can be one of the most frustrating things in fantasy and exactly what I want to help alleviate.  Building a more consistent team can make you one of the dominant forces in your league and directly lead to fantasy success. But finding consistency can be elusive at best and many “Consistency” metrics can be misleading at worst.  That is why I took my financial and analytical background and looked for ways to apply it to fantasy.

As I was comparing different funds to invest in for my very first 401(k), I went back to an old statistics lesson that I found helpful for financial analytics. It is called the coefficient of variation, or CV. CV is a measure of relative variability. It is the ratio of the standard deviation to the average.  This is particularly useful in finance when you are trying to compare the returns of different investments, like different mutual funds or bonds.

Fantasy Football

As I was comparing funds it suddenly hit me. I could apply this same formula for fantasy football. The fact that CV is especially useful when you want to compare results from two or more different sets or surveys that have different measures or values makes it perfect for sports analytics. You can use CV to, essentially take unrelated data sets & boil them down into one metric that is then compared. 

Using CV I could take each player’s fantasy production on a game-by-game basis and find out not only which players were consistent but consistently good. Because, let’s face it, if a player scores 4-5 points every week, sure they are consistent. But they don’t do your team any good.  

With CV, the more consistent the results, the closer to zero the CV ends up being.  What happens when you divide the same number by something closer and closer to zero? The result gets larger and larger.

Ex. 10 ÷ 0.5 = 20

Whereas, 10 ÷ 0.3 = 33.33

By using each player’s fantasy scoring to find their CV and then dividing that player’s ceiling by their CV I would have a comparable metric for consistency that was easy to rank and also compare across positions.

This methodology is also helpful because it factors in high outputs as an important part of the equation. Again, I would rather have a player who bounces between 18 and 23 points a week versus a player who scores 11 points each and every week. So let’s take a look at my rudimentary example above, but change the other number in the equation. If I have two players with the same CV, 0.331, but player (A) has a ceiling of 24 and player (B) has a ceiling of 12, how does that change the example?

1. 24 ÷ 0.331 = 72.51
2. 12 ÷ 0.331 = 36.25

So the player with the higher ceiling would have a higher consistency score even though they both have the same CV which is what is essentially measuring the consistency of a player. 

PERFECT!

This was exactly what I was looking for! One metric that could show you not only who was consistent, but consistently good. 

What makes the Consistency Score different?

As I started to keep this metric in order to see how it performed before making it public, I also went out into the fantasy community to see if others were working on consistency or not. 

I found several different consistency ratings/scores/charts/etc. all of which lacked something I wanted out of my metric. Consistency.

Much of the other consistency analysis out there, the main ones I found, were all based on weekly finishes. Who finished as a Top-5, or Top-10 player each week. Who finished outside the Top-40. Etc. The problem with this is simple. What makes a running back a Top-5 RB one week, is not what makes a running back a Top-5 RB the next week. The math would have to change from week to week, making the very metric itself not consistent. I can’t understand how you have a consistency rating for a player, that in and of itself, is not consistent.

Also, those ratings are not comparable across positions.  What makes a Top-10 running back is different than what makes a Top-10 wide receiver in any week. So these ratings are useless in helping make flex roster decisions. 

My Consistency Score is itself, consistent. The formula never changes. We just add more data to it as the season progresses and we have more scores available. Also, since the very idea of CV is a way to make non-comparable data comparable, it is useful in comparing multiple positions to one another. I can look at a running back with a Consistency Score of 6.4 and a receiver with a Consistency Score of 7.2 and feel confident in plugging that receiver into the flex spot on my roster.

Love this? Hate it? Let’s talk about it! Follow Phillip Caldwell on Twitter @DumpsterDiveFF for more of his great content!

Consistency Score Database

Consistency Score Articles