2012 Fantasy Football: Developing A New Algorithm For Fantasy Rankings
I recalled the unit in my Operations Management class on Forecasting Demand and decided I would try a 4-period moving average forecast technique.
Why 4? Teams do not have static schedules; they have waves of difficult games and cupcakes. A four game moving average is sure to keep our statistics useful and relevant, as well as capturing before and after the bye-week for many teams.
In order to adjust to strength of schedule, I also included in the spreadsheet, a 4 period moving average of the Points allowed by position. Since the algorithm is still in test phase, I will just be using it for QBs this week, since quarterback is generally the most consistent of all the fantasy positions.
As I was calculating the “forecasts” or “projections” for this week, I noticed how much things were trending toward the middle (17 points to 14 points range). Take for example Aaron Rodgers: Rodgers has a 4-game moving average of 24 points. This week, he faces a Cardinals defense that has allowed the 3rd fewest points to opposing quarterbacks over the last 4 weeks. (10).The formula so far was just add the two numbers and divide by 2 (24 and 10); I decided that this was too much of a penalty for Rodgers, as he dropped to 4th in the rankings.
With the formula giving to much weight to defenses, I went back into the rankings and picked out a small handful of quarterbacks from each extreme, and altered the weight given to each number. Elite quarterbacks like Rodgers and Brees, have their personal 4 game average count for 70% of the total average whereas quarterbacks like Russell Wilson and Blaine Gabbert, who are more prone at this point in their careers, to be “less in control of the game” than the defense is, got a 60% weight to their personal numbers. (Instead of 50/50)
A final note on another formula adjustment, is since a lot of QBs trended toward the 2 TD average range, I used 4 points as a standard deviation (the points for a passing touchdown), and increased or decreased their projection by 50% (2 points).
(ECR expert consensus ranking in parenthesis)
1(2) Drew Brees
2(1) Aaron Rodgers
3(5) Matthew Stafford
4(8) Michael Vick
5(3) Peyton Manning
6(4) Robert Griffin III
7(11) Josh Freeman
8(14) Matt Schaub
9(7) Cam Newton
10(12) Andrew Luck
11(9) Ben Roethlisberger
12(6) Matt Ryan
13(21) Brandon Weeden
14(15) Jay Cutler
15(13) Tony Romo
16(16) Carson Palmer
17(22) Russell Wilson
18(10) Eli Manning
19(24) Ryan Tannehill
20(26) Christian Ponder
21(20) Ryan Fitzpatrick
22(19) Joe Flacco
23(17) Andy Dalton
24(18) Philip Rivers
25(23) Matt Hasselbeck
26(28) Blaine Gabbert
27(27) Matt Cassel
28(25) John Skelton
29(29) Matt Moore
30(30) Nick Foles
Explanations of specific rankings:
Question: Do you really expect the Bengals to contain Peyton Manning? You have Michael Vick ahead of him, and Manning is 2 spots lower than anyone else.
Answer: The Bengals allowed 6 passing touchdowns in the first 4 weeks. In the past 3 games, they’ve allowed 3. Furthermore, in the first 4 games the Bengals did not face more than 20 rush attempts in any game. In the past 3, they’ve faced 31, 30 and 24 rushing attempts, to go with 4 TD’s
Question: Surely you’re not suggesting starting Mike Vick over RG3 and Peyton Manning?
Answer: Good question, I’m not, but I do think Mike Vick’s range of possible outcomes is slightly higher than RG3’s and Manning’s.
Question: Schaub at 8? Ahead of Matt Ryan and many other great starting QB’s?
Answer: Mark Sanchez threw for 3 touchdowns vs. the Bills, Matt Cassel threw for 301 yards and 2 TD’s, I think I’ve explained myself perfectly.
Written by Josh Kay exclusively for TheFantasyFix.com. Send all your PPR questions to Josh on Twitter @JoshKay_Fantasy.