I’ve been getting a lot of questions about where my numbers are coming from or what a particular column of stats means, so I figured I needed to put together a small walkthrough/FAQ of the method.
The basis of all the data that you will see in my daily fantasy posts are my projections. These projections are created using a similar process to the Simple Projection System detailed at Basketball Reference. The Marcel projections in baseball, created by Tangotiger, is what started all of this.
From these projections, I have an expected number of fantasy points scored per possession. This is very important. Teams play at different paces. I want to know how everyone performs at a neutral pace, first. Then, I will adjust the totals based on that particular game. Using last night (November 19, 2014) as my example, Dirk Nowitzki came into the night with a projection of 0.55 fantasy points per possession and a DraftKings salary of $6,900. I predicted he would play 28.4 minutes in a game that had a pace of 92.2 possessions. With this information, we can predict Dirk Nowitzki’s fantasy points for the game.
.55 * [(28.4 / 48) * 92.2] = 30.0 DraftKings fantasy points.
(30.0 / $6900) * $1000 = 4.35 DraftKings fantasy points per $1000. [Note: the higher the DK/$1k, the better.]
The other piece of information I offer is a particular player’s chances of hitting a level of value, namely 5X and 6X. 5X and 6X are multipliers people use as a guideline for specific point total. In DraftKings, 5X corresponds to 250 total fantasy points, while 6X is 300 fantasy points. If each player in your lineup scored EXACTLY 5X, you would score 250 fantasy points. It’s just like the calculation above for DK/$1K. With a $50,000 salary cap and a goal of 250 fantasy points, it would look like this:
(250 / $50,000) * $1000 = 5.00 DraftKings fantasy points per $1000.
So, in Dirk Nowitzki’s case, here are his 5X and 6X thresholds based on his salary:
5X: ($6900/1000) * 5 = 34.5
6X: ($6900/1000) * 6 = 41.4
This brings me to the 5X% and 6X% numbers that I have in each table. Assuming that a player’s possible outcomes are normally distributed, I will calculate the probabilities that he will hit a certain threshold. I use his fantasy points per possession (in Dirk’s case, it’s 0.55) and his standard deviation of fantasy points per possession (in Dirk’s case, it’s .15) to get a range of potential fantasy point totals for that night. Last night, Dirk had a 30% chance of reaching 5X and a 10% chance of reaching 6X. This helps get an idea of what someone’s upside is for that night. I really like it for the middle tier of salaries, to get an idea of who can perform more like a star.
If anyone has any other questions, please feel free to contact me via comment here, via email or on Twitter. I’m happy to help as much as possible.