Temper Your Expectations: Identifying Down Weeks From Top Players
Throughout the season, we'll hear it multiple times from our favorite fantasy analysts: “You're still going to start him, but temper your expectations.” This sentiment refers to a player that consistently finishes the week near the top of his position in fantasy points scored, but is facing particularly tough conditions.
A classic example of this scenario is when a quarterback like Drew Brees is playing on a cold, wet night in Seattle. Brees is still projected ahead of most signal callers, but it's apparent that fantasy owners shouldn't expect the production that they're used to from their elite quarterback.
There are two main issues with the “temper your expectations” analysis; the expected drop in production is not quantified compared to the player's normal output, and owners are advised to start their elite player despite tough conditions, but are rarely guided on how they can, if at all, make up for the probable dip in fantasy points.
In this mini-series, I'm going to illustrate how fantasy owners can utilize projections as a tool to recognize expected down weeks from top players, and how we can use specific roster construction techniques, namely targeting boom or bust plays, to offset the likely drop in scoring.
When to “Temper Your Expectations”
John Paulsen recently wrote a brilliant article highlighting the most consistent quarterbacks using fantasy points and standard deviation. Using a similar methodology and 4for4's weekly standard projections from 2013, we see a similar consistency from the top players in their weekly projections, whereas John was looking at actual points scored. By considering deviations in a player's week to week projections we can quantify actual “temper your expectation” scenarios using a metric called Points Below Average Projection (PBAP).
As we'll see as we go through each position, the variance in weekly projections over an entire season for elite players is considerably low. In order to identify legitimate expected down weeks, we will use standard deviation to set PBAP thresholds for each position. Again, a lower than usual projection for a top player will usually still be worthy of a start, but a high PBAP will serve as a good indicator that we may want to tweak our lineup elsewhere.
Rather than arbitrarily picking a group of top players that we will usually start no matter what, I defined “must-start”, or elite, players as those that make up the top half of starters at their position, in either end of season points or fantasy points per game. Assuming a standard league of 12 teams that starts 1 QB, 2 RB, and 2 WR, the elite group is made up of the top 6 quarterbacks, top 12 running backs, and top 12 wide receivers.
Examining the deviations in each groups projections as a whole, rather than each individual player, gives us a PBAP guideline that we can use when looking for lower than usual scoring weeks from each position.
Note that tight ends were not considered in this study. Outside of Jimmy Graham and maybe Rob Gronkowski, an expected drop in production from a tight end will almost never elicit considerable changes in our roster building strategy. Tight end is an inherently volatile position, and most tight ends can simply be sat in favor of a better option if we expect an unusually bad week.
Over the entire course of 2013, 4for4 projected a total of 107 games for elite level quarterbacks and those weekly projections had a standard deviation of just 1.46 fantasy points, meaning that the weekly projections for any top tier quarterback almost always falls within a 3 point range.
As stated before, we should take notice if a projection falls 1 standard deviation below a top player's average projection, but a deviation of just 1.46 fantasy points below normal will be somewhat frequent. For top end fantasy quarterbacks, the actual “temper your expectation” occurrence is when their projection is 1.5 standard deviations, giving us a 2.19 PBAP threshold.
Last season, 4for4 projected elite quarterbacks 2.19 PBAP 11.2% of the time, which is about 2 games per season for each top tier quarterback.
The standard deviation for elite running back projections in 2013 was 1.86 fantasy points, based on 193 weekly projections. Like signal callers, the top running backs' weekly projections will almost all be in a 3 to 4 point range, but since they score less total points than quarterbacks, running backs' relative variance is actually larger (think back to John's explanation of CV).
Since running backs have more week to week variance than quarterbacks, our BPAB threshold, in terms of standard deviations from the mean, is lower than it is for signal callers. At about 1.25 standard deviations, or 2.79 PBAP, the projections reflect cause for concern for your must-start running back that week.
Recognizing when a must-start receiver might have a bad week by comparing weekly projections to their average projection seems unreasonable at first glance. Because wide receivers are inherently volatile on a week to week basis, their weekly projections are usually fairly conservative and fall into an even tighter range over a full season than those of quarterbacks or running backs. There is still enough variance in their projections, however, that we can identify when we might mitigate our hopes for a top wide out.
Elite wide receivers weekly projections have a standard deviation of 1.15 fantasy points. This low standard deviation highlights their conservative week to week projections. Like running backs, if a top tier receiver's projection is more than 1.25 standard deviations below their mean, or 1.44 PBAP, there is significant cause for concern.
There are going to be times throughout the fantasy football season when one of our stud players is in a particularly poor situation, but we will still start him because his lower than usual projection is still higher than any of our other options. By looking at average projections and deviations in those projects, we can actually quantify “temper your expectation” situations through the use of PBAP.
Projections that are just a couple points below average might not seem like much of a difference, but when we consider standard deviation, the small difference is truly indicative of player that we expect to perform well below their usual output. The numbers laid out here are good guidelines to recognize when a stud player may falter.
Since all the heavy lifting is already done as far as calculating standard deviations within projections, you can simply use the PBAP thresholds laid out in this article to diagnose when we really should have lower expectations for our must-start players.
Early in the season we can extrapolate our mean weekly projection for a player by taking season long projections and dividing them by 16. After we have enough data points (about 4 weeks) we should be able to compare a player's current weekly projection to their actual average projection on the season. (Hint: export 4for4 CSVs every week.)
In the next iteration of the “Temper Your Expectations” series, I'll discuss how we can use nuances in 4for4's weekly projections to target boom/bust players that may make up for lowered expectations from elite players.