Dynasty, in Practice: Predicting Stretch Performances

Making predictions about order of finish over the final twelve weeks to evaluate after the season.

It’s Science Week(TM) here at Dynasty, in Theory and Dynasty, in Practice! Earlier, I took a look at reproducing one of my most-referenced studies, with mixed results. But science isn’t just about re-running our tests to check whether the results hold up. It’s about forming hypotheses and testing them out.

Over the years, I have found that, after four games, preseason ADP typically holds as much predictive power for rest-of-season performance as early-season performance does. This is important to know, but it should not be construed as me advocating that we ignore everything we learned and look only at preseason expectations.

Instead, I would hypothesize that a skilled fantasy owner could outperform both preseason ADP and early-season performance. So let’s put that hypothesis to the test. I’m going to take a look at the top players at each position and predict whether their game 5-12 performance will hew more closely to preseason expectations or to early-season performance.

These predictions could be viewed as me taking a stand on which players are good buy-lows and which are good sell-highs. And they are, indeed, that. But mostly they’re my attempt at preserving my current thoughts for posterity. Because, to quote from the very first Dynasty, in Practice:

The problem with revisiting our past decisions, of course, is recalling our past decisions. For most, the process of recalling is one of remembering. The problem, of course, is that our memories are bad. Worse than bad, our memories are biased. They are systemically flawed in predictable ways.

Enough talk. Let’s take some players, make some predictions, and drop them into a time capsule to revisit at a future date. (Once again, I will be using standard scoring for all rankings for future ease of comparison.)

(One last caveat: because of week 4 byes, a few players will not have had a chance to play their fourth game yet. I have included them and marked them with an asterisk; when evaluating my predictions, I will include their fourth game, as well. I’ve also preemptively removed players such as Ben Roethlisberger and Dez Bryant who will certainly not meet the minimum games played requirement for evaluation.)


ADPPlayer1st 4 gamesBetter predictor:
1 Andrew Luck 28 preseason
2 Aaron Rodgers 1 early season
3 Peyton manning 24 preseason
4 Russell Wilson 8 early season
5 Drew Brees 22 preseason
7 Tom Brady 5* early season
9 Matt Ryan 6 preseason
10 Matthew Stafford 26 preseason
11 Ryan Tannehill 13 early season
12 Cam Newton 3 early season
13 Eli Manning 12 preseason
14 Sam Bradford 21 early season
15 Philip Rivers 7 preseason
16 Teddy Bridgewater 27 preseason
17 Joe Flacco 25 preseason
18 Carson Palmer 4 early season
19 Colin Kaepernick 23 preseason
20 Andy Dalton 2 early season
21 Jameis Winston 17 early season
22 Jay Cutler 31 preseason
23 Marcus Mariota 15* early season
24 Derek Carr 14 early season

 Running Back

ADPPlayer1st 4 gamesBetter predictor:
1 Adrian Peterson 3 early season
3 Eddie Lacy 21 preseason
4 Jamaal Charles 2 preseason
5 Marshawn Lynch 49 preseason
6 C.J. Anderson 56 preseason
7 DeMarco Murray 40 preseason
8 Matt Forte 5 preseason
9 Jeremy Hill 13 early season
10 LeSean McCoy 30 early season
11 Justin Forsett 26 early season
12 Lamar Miller 48 preseason
13 Frank Gore 20 preseason
14 Mark Ingram 7 preseason
15 Melvin Gordon 41 preseason
16 Alfred Morris 34 early season
17 Carlos Hyde 12 preseason
18 Latavius Murray 9 early season
19 Joseph Randle 4 preseason
21 Ameer Abdullah 25 preseason
23 Jonathan Stewart 42 preseason
25 Doug Martin 19 preseason
26 T.J. Yeldon 29 preseason
27 LeGarrette Blount 35* preseason
28 Chris Ivory 8 early season
29 C.J. Spiller 28 preseason
30 Joique Bell 63 early season
31 Giovani Bernard 14 early season
32 Rashad Jennings 22 early season
33 Isaiah Crowell 23 early season
35 Alfred Blue 37 early season
36 Darren McFadden 60 early season

 Wide Receiver

ADPPlayer1st 4 gamesBetter predictor:
1 Antonio Brown 3 early season
3 Demaryius Thomas 22 preseason
4 Odell Beckham 17 preseason
5 Julio Jones 1 early season
6 Calvin Johnson 39 preseason
7 Randall Cobb 10 preseason
8 A.J. Green 5 preseason
9 T.Y. Hilton 41 preseason
11 Brandin Cooks 49 early season
12 Mike Evans 77 preseason
13 Emmanuel Sanders 18 early season
14 Jordan Matthews 29 preseason
15 DeAndre Hopkins 4 early season
16 Andre Johnson 111 preseason
18 Amari Cooper 16 early season
19 Golden Tate 54 early season
20 Keenan Allen 6 early season
21 Julian Edelman 19* early season
22 Davante Adams 94 preseason
23 Jeremy Maclin 13 early season
24 Sammy Watkins 65 preseason
25 Brandon Marshall 7 early season
27 Jarvis Landry 30 preseason
29 Nelson Agholor 103 preseason
31 Vincent Jackson 21 early season
32 Mike Wallace 43 preseason
33 Allen Robinson 15 preseason
34 Larry Fitzgerald 2 early season
35 Charles Johnson 127 preseason
37 Steve Smith 11 preseason
38 Roddy White 95 preseason
39 John Brown 37 early season
40 Torrey Smith 45 preseason
41 Eric Decker 26 early season
42 Michael Floyd 98 preseason
43 Marques Colston 56 early season
44 Anquan Boldin 55 preseason
45 Terrance Williams 36 early season
47 Devin Funchess 131 preseason
48 Kendall Wright 28* early season

 Tight End

ADPPlayer1st 4 gamesBetter predictor:
1 Rob Gronkowski 1* early season
2 Jimmy Graham 11 preseason
3 Greg Olsen 5 preseason
4 Travis Kelce 2 early season
5 Martellus Bennett 8 early season
6 Jason Witten 6 early season
8 Jordan Cameron 23 early season
9 Zach ERtz 32 early season
10 Owen Daniels 19 early season
12 Delanie Walker 22* early season
14 Tyler Eifert 3 early season
15 Kyle Rudolph 25 preseason
16 Josh Hill 37 early season
17 Vernon Davis 30 early season
18 Coby Fleener 16 preseason
19 Richard Rodgers 14 early season
20 Heath Miller 17 preseason
22 Larry Donnell 20 early season
23 Charles Clay 4 early season
24 Eric Ebron 10 early season


Well... we aren't really at the conclusions stage of the process yet. We still have twelve more weeks of data to gather. In the meantime, this could serve as a solid list of guys I expect to rebound towards their preseason ADP and guys I expect to keep early-season performance rolling going forward.

After the season has concluded, I'll be able to test my hypothesis that educated fantasy owners can effectively synthesize both preseason expectations and early-season results to create an amalgamation that is a better predictor than either. I will test this by using a combination of both preseason ADP and early-season ranking, (as indicated above). If I'm correct, my correlations will be higher than either other correlation on its own. If I'm wrong, and we really aren't good at predicting which results represent an aberration and which are the new reality... well, the data will show that, too. It's science!

More articles from Adam Harstad

See all

More articles on: Dynasty

See all

More articles on: Strategy

See all