.
Do you find yourself looking at “Projected Fantasy Points” (we’ll call them PFPs) in your league management service and wondering how in the world they come up with those numbers? Or, have you lost faith in these numbers because you have found they just aren’t that accurate?
This topic is one of the regular conversations we have at PlayThisGuy! Heck, we’re in the fantasy business but how these numbers are determined was largely a mystery to us…until now. Read on to understand how the most common approach for determining PFPs works, why we believe it’s broken and how we are fixing it.
.
Projected Points – Love Them and Hate Them
PFPs are a sort of guilty pleasure – like reading Deadspin.com, TMZ.com or picking up a People Magazine when you’re at the dentist office. Most of us scoff at making decisions based on these projections yet they are very hard to ignore because the most popular league management systems place them smack dab in your face.
You can either love PFPs or hate them but they are here to stay. So, with that in mind let’s dig deeper into how to maximize the usefulness of this statistic.
.
The Incumbant and Why Their Approach is Wrong
We believe most PFPs for sites like Yahoo! and ESPN are provided by a company called AccuScore.com (it’s pretty hard to tell…wonder if that’s on purpose). For those of you unsure about how AccuScore works here’s a summary:
- AccuScore rankings “allow you to view forecasted rankings based on the average performance of each player after 10,000 simulations.” In layman terms they take each player and game and run it through a computer simulation 10,000 times
- The results of the game and players performance are the basis for their PFPs
They provide more statistical speak on their Web site so feel free to read it here.
In our view a better way of predicting fantasy points isn’t based on a game simulations but on actual results from the field. Here’s why…
- A defense catches fire and has a few great weeks (see NYG in weeks 4 and 5 of 2010) – how can the simulation account for a “hot” team?
- A RB with a moderate leg injury may be hindered in the upcoming game. Information on the extent of his injury doesn’t come out until Thursday or Friday. How can the computer include a partial injury in it’s simulation? After all, some injuries have much more of an impact than others.
- A top-peforming QB is ruled out for a game three days prior to kickoff. What impact does that have on the WRs and RBs?
We highly doubt that any computer simulation can account for these situations. Yet, those are EXACTLY the type of situations found in football week in and week out. They are unavoidable.
This specific point is exactly why fantasy football experts add value and their advice is needed. They interpret statistics, probabilities AND the human elements that could impact a team and player and provide a level of context that can’t be replicated by a computer. Their rankings include all elements that impact a player and team.
In our opinion using a purely computer-based model, no matter how intelligent the program is – just doesn’t work for fantasy football.
.
The Better Way of Projecting Fantasy Points
Our idea for creating better PFPs centers on how many points a particular rank has produced over time. For example, if Peyton Manning is the #1 ranked Consensus QB for Week 8 and you want to know how many points he could produce that week you should first understand how many points the #1 ranked QB from previous weeks actually scored. Let me share an example:
Let’s say that during the 2010 season the #1 ranked QB after the games were played scored 27.5, 21.7, 33.8 and 40 points during weeks 1 to 4. If you add these points together and divide by 4 you get an average of 30.8 points per game for the #1 ranked QB in 2010. Now, if the experts believe (and this is why our Consensus Rankings are so important because they average the rankings from more than 30 experts) that Peyton Manning is the #1 ranked Consensus QB for Week 5 you can expect Peyton to generate around 30.8 points if he finishes the week as the #1 QB. When the experts create their rankings they take into consideration all the imporant stats AND the human elements so it’s easier to understand what to expect if Peyton plays up to his ranking.
Of course, Peyton could underperform his ranking for the week and generate less points but looking at projected points this way will also give you a sliding scale of sorts. Since each ranking is tagged with projected points you can easily see how many points he could earn if he under performed by say 5 or 10 rankings. For an example of this in action see our Weekly Conensus Rankings here.
I’m no math whiz and I’m sure some statistician can poke holes in our thought process but here me out…doesn’t it make more sense to base PFPs on what’s happened on the field rather than a computer simulation? To us here at PlayThisGuy! it seems like a more logical and intelligent approach.
.
How We Determine Projected Fantasy Points
As stated above, each of our Consensus Rankings has a PFP based on how the ranking has performed in the past.
Here are some additional details on how we determine PFPs:
- Each ranking has a projected point total (its not based on a player – it’s based on a ranking)
- The projected points are based on the average fantasy points produced by the rank in prior weeks (see above for example)
- As the players Consensus Rankings change during the week (as we receive more inputs from the experts) their projected points can also change
- By Friday and Saturday we’ve analyzed the rankings from more than 30 different sources and the rankings and PFPs are pretty solid
- PlayThisGuy! PFPs are based on a standard scoring model (see below for specifics)
Our method may not be perfect but we hope you agree that our approach just makes more sense and is easier to understand. We’ll be doing some interesting thing around the accuracy of these projected points in the near future so stay tuned.
.
If you have any questions on our methodology or thoughts on how to improve PFPs then drop us a note at contact@playthisguy.com. We’re always looking for great new ideas.
.
PlayThisGuy! Standard Scoring Model
Passing TDs = 4 points
Passing INTs = -2 points
Rushing, Receiving, D/ST TDs = 6 points
Passing Yards = 1 point for every 25 yards
Rushing & Receiving Yards = 1 point for every 10 yards
Fumbles Lost = -2 points
Field Goals = 3 points
Extra Points = 1 point
Sacks = 1 point
Fumble Recoveries & INTs = 2 points
.
Add comment