Choosing a Quarterback: One Step Closer to Cracking the Code

I admit it. I am more obsessed with deciphering the quarterback position for DFS football than I should be. More bandwidth in the space needs to be allotted to the other positions. However, I honestly believe choosing a quarterback is integral to success, be it in a cash game or tourney and my heavy exposure to Matt Ryan this week suggests I still have some work to do. That said, there is some very good news to report. The projection system I detailed last week kicked some serious ass. I know, it's only one week but this is better than having to start all over.

Let's take a look at how the industry fared projecting the top-ten quarterbacks for Week 10. By means of reminder, the process involved averaging the ranking from several reputable sources and correlating that to the week-ending ranks. No consideration is given to the actual points scored since the projections for each QB is not available. The higher the number, the better the correlation with 1.00 being complete correlation and 0.00 being completely random.

So how did I do? On FanDuel, the correlation was .65 and on Draft Kings it was .66. It's only a week but so far so good.

Since my system was based on actual projected points, we can take the study one step further and correlate points, This hones the process a bit.

Are you ready? Looking at my top-ten projected signal-callers for Week 10, the correlation coefficient is a whopping .88 on FanDuel and .85 on Draft Kings. That's pretty good.

But, as was alluded to up top, I didn't take down Draft King's Millionaire Maker nor the half a mil from FanDuel. A couple weeks ago I was fixated on Ryan Tannehill and this week my bromance for Matty-Ice cost me some cash. I'll contend my mistake wasn't setting lineups with either Tannehill or Ryan. The error was too much exposure and not ample diversification. To be fair, I don't play the same volume as others. I'm above casual but nowhere near a grinder. But if this QB ranking system continues to exhibit some nice results, one of these weeks I'll choose the right lesser-owned (but still highly rated) play or I'll play more and set a multitude of lineups using a bunch of different highly ranked quarterbacks.

For what it's worth, I plan on coming up with a similar ranking system for the other positions. I'm also thinking about how to adapt the process to earlier in the season when player performance averages have too much noise. Perhaps using a trusted projection system in concert with a weighted average of performance to date will suffice.

We'll stick with the QB theme this week. On Friday I'll share the Week 11 rankings along with my thoughts on the more compelling options, Then next week we'll come up with some new angles to research.

See ya Friday!