Sunday, August 4, 2013

Points Per 90: an alternate look at adjusting for different numbers of games played

I was doing some research trying to pick my players for gameweek 24 and noticed Lamar Neagle's points per game was a mere 4.4. Something seemed very wrong and that this measurement was extremely under-valuing his production. I looked at some other players: Camilo, Villareal, Rodney Wallace, and all of them seemed to have PPG much lower than what I expected them to be. Most of the players who I found to have a large difference between their PPG and their PP90 were players who have had several appearances as substitutes.

So, I decided to nerd out about it and put all of the points and minutes for all players in the game with 500+ minutes into a spreadsheet. I calculated how many points per minute a player scored up to the all-star game (so only results up through week 23 are included here). I simply multiplied those by 90 to find how many points they would score if the player scored at that rate over an entire game. The tables below show the results by the three outfield positions.

There's a few things I want to talk about as potential uses of this data.

First, we can see that simply judging a player by their total points might not be the best way to judge a player's scoring potential because they might simply have played more minutes than players below them in the total score rankings. Higuain drops 8 ranks when comparing his total score to his PP90 and 4 ranks when comparing his PPG to his PP90 (ignoring players with fewer than 500 minutes played). Okugo drops 15 ranks and 10 ranks for the respective methods of ranking. Michelle drops 11 and 2. There's a lot of players that have higher scores simply because they have played more minutes. This could be because some of other players that show highly in my PP90 are subs (and shouldn't be considered for fantasy unless they lock down a starting role), it could be that a team has played fewer games than other teams (Seattle played 4 fewer games than Colorado), or a player could have missed games due to injury or national team duty.

Second, we can use the PP90 to try to find players flying under (or unrightly in) the radar. Players like Agudelo and Osorio have the potential to be great differentials for us. On the flip side, Okugo and Bieler are not nearly as good as their overall scores make them out to be.

Third, injuries to players ahead of them might lead to the opportunity for some of these players to be sweet differential picks. Hector Jimenez, Soony Saad, Kalif Alhassan, Olmes Garcia, Warren Creavalle, all scream "PICK ME! PICK ME!" should they lock down a starting role. Saad is the closest of that bunch to be able to lay claim to that title and might already have that position in lockdown.