How to Accurately Project Player Fantasy Scores
What Factors Determine a Player’s Next Fantasy Score?
Last year I posted on what factors into a players fantasy score for an upcoming game. There are three things that determine it. The first factor is a player’s baseline fantasy strength. This is the long-term average fantasy score the player would obtain against perfectly neutral match-ups. The next factor is the match-up strength of their next game. This is based on how far the strength of the upcoming match-up varies from what a perfectly neutral match-up would be. Both of these factors can be determined with pretty high accuracy if you can consider everything that goes into them. The last factor tends to throw everything off in a given week, and that is the dreaded variance factor. Variance accounts for the unaccountable. It accounts for when your WR is stopped at the 1 yard line, and the TD gets scored by the RB instead. It accounts for the simple missed tackle that turns a 5 yard reception into an 80 yard TD. These are simply things that you can’t account for in your projections, and is why they end up being off by large margins in many cases. The only good thing about variance is that it completely washes away over the long-term. For every time that you are on the wrong side of it and unlucky, you will be on the right side of it and be lucky. What you want with projections over the long-term is for the sum of their errors to approach zero. When that happens your projections are gold, and you will start making good money at Daily Fantasy Sports.
Projecting Individual Statistics
Let’s move on to the good stuff already. If you want to have highly accurate projections with the sum of the errors approaching zero, you need to go into the most detail as possible, and that requires projecting all statistical categories that are scored by the site you are playing on. This is required so you can see all of the detail of the next match-up. A running back that scores a ton of TDs but not so many yards is better matched against a defense that is tough on yards and weak on TDs. The formula for an individual player’s projection is as follows:
| Fantasy Point Projection | = | Stat Category 1 Player Strength | x | Stat Category 1 Match-up Strength | x | Stat Category 1 Fantasy Scoring | + | Stat Category 2 Player Strength | x | Stat Category 2 Match-up Strength | x | Stat Category 2 Fantasy Scoring | … |
|---|
Variance is left out of the calculation, since it is something that can’t be determined before the game is played. Think of variance as the range of possible values around each statistical category with the more likely outcomes getting more weight (a Gaussian distribution). Some players are low variance in many of their statistical categories, and some are high variance. In general you want to use low variance players when ever possible.
Calculating a Players Statistical Category Strength
A player’s statistical category strength is what they would put up on average, long-term against a perfectly average defense. For sports like baseball, basketball, and hockey once you are about 40 games into the season this starts to converge to their average totals in each statistical category (which is why FSL prices players based on their average fantasy points). This happens because you are starting to get long-term which tends to wash away the variance, as well as wash away any issues with schedule strength. If you want to do projections for football or early season projections in the other sports, using the player’s average stats will not be good enough, as there are substantial errors based on strength of schedule. Variance is also not totally washed away, but there is not much we can do about that. So to get to the players statistical category strength you need to factor out the schedule strength of their previous match-ups. For very early in the season, and for football all season long, when determining the match-up strengths to factor out, you should also consider the strength of schedule that the defenses you were matched up against faced. I will cover this later. For example a defense that has faced a lot of soft rushing yards match-ups, may be currently statistically the toughest against rushing yards, but part of that is based on the match-ups they faced, and not their underlying statistical category strength.
Calculating Defensive Match-Up Strength by Statistical Category
Once you have each player’s statistical category strength at all categories that are scored, it’s time to move on to the next factor. Defenses statistical category strength is how that defense does relative to the league in each category. For example a defense that gives up 110 rushing yards on average against a league average of 100 yards has a factor of 1.1 or 110%. You would expect your player to gain about 10% more rushing yards than their underlying strength when matched up against this defense. For football and early in the season for any sport, this simple calculation is not good enough. You need to also factor out the strength of schedule for that category. To get that you need to calculate team offensive strengths at each statistical category, and average the statistical strength of previous opponents of the defense. For example if you were a 1.1, 1.2, .9, and .8 so far this season you have faced neutral match-ups so far. If you have faced 1.2, 1.1, 1.3 and 1.2 so far you have faced a 1.2 on average. So if your defense is giving up 120 rushing yards on average this season, but has faced a 1.2 average match-up, their underlying average is really 100 yards, and you need to use that number as their defensive strength.
What Factors Are Being Left Out?
You should also factor in home and away strength for both your player and the defense faced to improve accuracy. This complicates things quite a bit, but home players and defenses tend to do better then away. You would need to factor this out of both your player and the defense. It is tough the get an accurate measure for this early in the season, but this is a pretty generic factor, and you could look at last season to help nail this down.
Using this approach you are not factoring in the “hotness” or “streak” of the player or defense. To some extent this is an overused factor. A player or defensive strength has a lot to do with what the recent match-ups were. If you want to use this, it would absolutely have to be adjusted for the recent match-up strengths.
No adjustments are made for injuries. A player may have been injured early in the season, and is now healthy, or vice versa. This will tend to throw off your projections. Because this is so hard to factor out globally, I just assume health in the past, and avoid injured players in the future.
Final Thoughts
To go to this level of projections takes a lot of work, and is something you would never want to do by hand. You need to be highly proficient at Excel or databases to even try it. I personally use a massive spreadsheet to do my projections, and it takes about 15 minutes of work each week to get the projections updated for the following week. There are also other “secret sauce” type items that I am, and you should keep close to the vest. Player projections are site specific based on how the sites score each category. If you would like to see my projections for Fantasy Sports Live, I am posting them weekly, at the FSL blog, and will be posting some here as well.
_________________________________________________________________________________________________
About the Writer: After much research and preparation, Blinders turned his vision and passion for fantasy sports into the first Daily Fantasy sports site with Salary Cap based games in June of 2007: FantasySportsLive. He is a longtime online poker player and blogger, and the only daily fantasy grinder who was willing to take on Buffalo66 in his multi-sport fantasy challenge. Last NFL Season Blinders went 71-10 on FanDuel alone.