Let's categorize quarterback performance into tiers
The tiers defined in the table below are approximately based on standard scoring point quantiles from 2018 data.
<table> <thead> <tr> <th> Tier </th> <th> Points </th> </tr> </thead> <tbody> <tr> <td> Bad </td> <td> 0-10 </td> </tr> <tr> <td> Average </td> <td> 10-15 </td> </tr> <tr> <td> Good </td> <td> 15-20 </td> </tr> <tr> <td> Elite </td> <td> 20+ </td> </tr> </tbody> </table>Probability of quarterback performance tier in their next game (NG) given their performance tier in their last game (LG)
The table below can be read as follows: The row labels in the table represents quarterback performance from their last game (LG) while the columns represent projected performance in their next game (NG).
For example: If a QB was "Elite" last game, the probability of them performing "Elite" in their next game is 45%.
<table> <thead> <tr> <th> </th> <th> NG Bad </th> <th> NG Average </th> <th> NG Good </th> <th> NG Elite </th> </tr> </thead> <tbody> <tr> <td> LG Bad </td> <td> 0.29 </td> <td> 0.14 </td> <td> 0.30 </td> <td> 0.26 </td> </tr> <tr> <td> LG Average </td> <td> 0.27 </td> <td> 0.20 </td> <td> 0.20 </td> <td> 0.32 </td> </tr> <tr> <td> LG Good </td> <td> 0.18 </td> <td> 0.24 </td> <td> 0.19 </td> <td> 0.38 </td> </tr> <tr> <td> LG Elite </td> <td> 0.16 </td> <td> 0.16 </td> <td> 0.22 </td> <td> 0.45 </td> </tr> </tbody> </table>Key takeaways
- When a quarterback has an "Elite" performance, there is a 45% chance they will perform "Elite" in their next game
- When a quarterback has a "Bad" performance, there is a 56% chance they will perform "Good" or "Elite" in their next game
Daily fantasy sports strategy
Quarterback drafting strategy
Since quarterbacks tend to rebound after a "Bad" performance, a strategy worth looking into more is drafting quarterbacks that performed "Bad" last week, but have a history of "Good" or "Elite" performances.
Wide receiver drafting strategy
In an earlier post, I talked about how quarterback performance affects wide receiver performance.
A key finding (albeit small sample size) was that WR2 performance given an "Elite" quarterback performance is in on par with WR1 performance given a "Good" quarterback performance.
A good strategy could be to draft low-cost wide receivers that have quarterbacks with a high probability for "Elite" performance in the next game.
If there are enough people interested, I could generate quarterback performance tiers each week and post them here.