FSL Shares How They Calculate Player Salaries

by Blinders Last Updated 2010-09-10 18:15:51

Player Salary Cap Values in Weekly Fantasy Football

About the Writer: Blinders is an employee of Fantasy Sports Live and a proven NFL Grinder. Last NFL Season Blinders went 71-10 on FanDuel alone.

How FSL Calculates Player Salaries

At Fantasy Sports Live we take the same approach to setting individual player salary cap costs in fantasy football as we do with all of our daily fantasy sports. There are two main requirements that we seek to meet with individual player cap values. First, the cap values should be transparent and make sense to anyone looking at them. Second, player costs should be relative to average player performance. For example a player who averages Five times as many fantasy points per game as another player should cost Five times as much as the other player regardless of the player’s positions or any other factors. That’s the way things work in the real world and fantasy football should not be any different. If you don’t set things up this way, certain classes of players become too expensive relative to other classes, and as a result become very bad choices (undraftable) for a fantasy team limiting the variety of competitive fantasy teams. At FSL we want all players to be equally good choices on the surface, with no penalty for drafting certain classes of players.

At FSL we have three types of salary caps. A performance based cap called “Ranking Cap”, a pay based cap called “Salary Cap”, and an uncapped cap called “No Cap”. For No Cap there are no cap restrictions on who you can pick. You simply pick who you think will score the highest for each position that week. For Salary Cap, we take the actual 2010 salary for each NFL player and divide that by 16 to get a per game salary. Since actual player salaries are fixed before the season, the Salary Cap values do not change based on performance during the season. For ranking cap we set the player cap values to their average fantasy points scored per game x 10. The 10x multiplier is used to provide a good range of values to choose from when drafting. One nice thing about using this approach is that you can determine what an average score would be for any league that you enter. For example, in a 950 pt ranking cap league if you made random player selections, you would score 95 points on average. When you score above 95 points your fantasy team is better than average. This gives you a good idea of the types of scores it will take to be competitive in a given league. An 850 point cap league would score 85 points on average.

We do not make any adjustments to player cap values outside of the average fantasy points scored. The average fantasy points scored is the obvious cap value for players. It lines up well with the statistical rankings of players that you see in the sports section and on web sites. It is your job as a skilled fantasy grinder to determine the factors that will cause players to score higher or lower than average in a given week. If we adjusted for these factors it removes skill from of the game. For example, if we adjusted for recent results, let’s call it a “heat factor”, and were able to fully take account for the hotness of players, and how that effects their projected fantasy score, this would completely eliminate that advantage from players who consider it. Users who considered player hotness would lose their advantage, because that factor would be fully represented in the player’s cap value. By setting our player cap values proportional to their average fantasy score per game, we allow our users to apply as much skill as possible to overcome the small house edge. Limiting the skill that can be applied lowers the chances for a good player to make money long-term at fantasy sports.

Not all daily fantasy sports sites take this transparent approach to setting individual player cap values. Over at Fanduel, where I am competing against Buffalo66 in an ongoing multisport challenge, they use a formula that is so complicated, that I have no idea how it works. They appear to use some sort of base cap value and add to that based on recent performance, number of games, and other unknown factors. They miss on both of our main points. It is not transparent, and player cap costs are not relative to performance. Let’s look at Fanduel’s Tight End Values for example.

So Gates costs the most at $6600, and they have a ton of poor Tight Ends that all cost $4300. This implies that Gates scores about 50% more fantasy points per game that the worst TEs in the league. But that is not true. He actually scores 5-10x as many fantasy points per game as the $4300 TE group. Gates should cost at least 5x as much as the scrub TEs, but he does not, and that makes all of the $4300 Tight Ends very poor selections and pretty much undraftable. Worse yet is the fact that Davis and Clark actually average slightly more fantasy points per game than Gates at Fanduel. The reason Gates is the most expensive, when he scores 10.2 pt/game and Davis is less expensive when he scores 10.7 pt/game is a mystery that only FanDuel’s management can explain. In contrast, at FSL Gates costs 132 pts because he averages 13.2 fantasy points per game, and TEs who average less than 2 fantasy points per game cost 20 points.

Now let’s look at their Defense cap values at FanDuel. SF is the highest at $6500 and the lowest cost is the same for about 20 of the defenses at $5700. This implies that the spread in average scoring among all fantasy defenses is almost zero. The worst fantasy defense is only about 10% less expansive than the best. In actuality, the worst fantasy defense scores about 90% fewer points, and should be 90% cheaper (not 10%) as a result. So for fantasy defense at Fanduel, you can just forget the cap value, and simply pick the highest projected score, because the salary costs are too close together to matter. So much for picking a bad defense that is expected to do well this week. They probably will still not do better than the best Defenses, so they will be greatly overpriced. In contrast, at FSL SF costs 108 pts because they averaged 10.8 fantasy points per game last year, and Detroit costs 13 pts, because they don’t have much of a defense at all.

In Season Player Cap Value Updates

At Fantasy Sports Live, we update player cap values for the NFL every week. This way they will always be perfectly accurate during a season with a limited number of games. For the first week of the season, they are based on last season’s fantasy averages. For the first four weeks of the season we start moving them from last season’s average to this season’s average. There is a bit of an art to this, but in general we allow each player to move a maximum amount each week towards their 2010 season averages. After 4 weeks we are completely there. For rookies and players getting more minutes, we add them to our draft pool after they have some fantasy experience. That way we don’t have to set high minimums for unknown players. We know who they are and how they should be priced before adding them to the draft pool. As a result we have a smaller pool of available players than some other sites, but all of our players are equally draftable as a result. For the worst fantasy performers we use a minimum cap value that is very low relative to the best players. We try to set this minimum low enough that only a handful of players need it, and those players would probably be poor fantasy choices anyway.

- Blinders