|
Drinen rambles about something having to do with:
Derrick Mason
Introduction to these player comments
Derrick Mason was terrific in the second half of 2001, and for that reason a lot of people are touting him as a hot prospect for 2002. I like Mason too, but ever since I did this study, I've been very suspicious of the idea that 1st-half / 2nd-half splits are of any use at all. I decided to update that study with more data and a better methodology. The first thing we need is a way to measure whether a particular player did better early in the season or late in the season. The standard way to do this is to compute the difference between his fantasy points per game in his last 8 games vs. his fantasy points per game for the first 8 games. But I don't think that captures exactly what we want to capture. Consider the following two made-up players:
Player A weekly scores: 10 10 10 10 10 10 10 40 10 10 10 10 10 10 10 10 Player B weekly scores: 10 10 10 10 10 10 10 10 40 10 10 10 10 10 10 10 If you just look at the difference between their first 8 games average and their last 8 games average, A looks like a strong first half player while B looks like a strong second half player. In reality, though, they were pretty much the same. Just because B had his big game in week 9 instead of week 8 doesn't really tell us much about how he finished the year. To correct for this, I'm going to take a weighted average of fantasy points, with later weeks weighted more heavily. Here's the formula:
1*(Wk 1 fant. pts.) + 2*(Wk 2 fant pts.) + 3*(Wk 3 fant pts) + ... + 17*(wk 17 fant pts.)
Weighted Average = -----------------------------------------------------------------------------------------
1 + 2 + 3 + ... + 17
Further, a game in which the player did not play will be omitted from the numerator and the denominator. Now, if we subtract a player's weighted average, from his regular old average, then a positive number indicates that he finished the season stronger, and a negative number means he started stronger. Let's call the difference "E/L" (for early/late factor). Just to give you a feel for what kind of numbers this formula produces, here is how everyone did last year. Strong finishers at the top, strong starters at the bottom.
------------- Weekly scores -----------------------
Name E/L 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
--------------------+----------------------------------------------------
M Faulk 2.30 | 18 18 34 21 6 31 21 6 37 22 16 37 40 28
D Mason 2.24 | 12 0 4 5 9 9 6 11 24 13 16 7 14 30
J White 1.94 | 7 1 15 0 4 0 1 2 10 1 1 6 7 27 21 11
S Davis 1.90 | 6 7 9 3 9 11 11 22 9 7 8 17 14 12 18 21
D Jackson 1.83 | 1 0 12 4 3 18 10 16 4 6 7 16 8 2 23 19
P Burress 1.82 | 2 1 3 3 2 15 11 4 1 17 8 2 22 11 22 4
A Thomas 1.75 | 0 0 11 5 24 19 22 9 6 13 23 8 7 23
K McCardell 1.64 | 0 4 7 9 3 11 6 12 8 8 8 15 6 9 19 15
K Stewart 1.62 | 4 3 18 16 1 22 18 14 27 33 4 12 30 35 16 0
D Rhodes 1.62 | 2 3 14 20 4 13 26 6 9 30 21 15 17
A Smith 1.52 | 3 2 27 10 10 7 6 14 22 4 27 13 19 9 21 14
M Hasselbeck 1.49 | 1 5 3 0 22 0 16 11 13 10 11 7 10
S McNair 1.49 | 10 4 23 13 7 41 18 16 27 28 26 23 8 26 19
T Barber 1.36 | 8 7 9 8 8 9 12 16 22 13 12 7 15 17
S Mack 1.27 | 0 15 11 2 8 7 6 19 2 21 4 2 13 23 19 6
P Holmes 1.19 | 2 4 40 10 26 10 7 25 7 19 17 39 18 9 20 16
E Smith 1.12 | 4 8 6 4 15 9 9 1 16 12 3 13 13 13
T Holt 1.10 | 8 1 17 13 8 4 13 7 14 13 12 6 5 8 32 9
K Dyson 1.04 | 5 3 0 7 1 2 12 12 11 17 11 5 4 9 15 5
J Stokes 0.95 | 15 0 1 0 1 2 7 19 10 6 1 2 0 16 13
D Staley 0.90 | 9 1 4 11 4 29 13 9 20 16 10 13 2
C Garner 0.88 | 5 10 4 6 8 15 9 8 10 16 8 5 11 12 14 11
C Conway 0.86 | 5 12 10 3 11 5 12 6 17 15 11 3 5 10 11 21
K Warner 0.81 | 12 28 37 29 10 14 9 2 28 11 37 22 37 15 29 20
M Bennett 0.78 | 5 8 7 3 5 4 9 17 23 4 2 11 6
J Graham 0.72 | 9 4 4 4 22 4 9 1 15 3 23 7
R Moss 0.67 | 2 8 9 10 7 7 10 15 36 2 20 21 26 3 2 2
M Alstott 0.65 | 2 3 14 0 1 32 10 6 9 18 5 16 3 16 14 1
J Horn 0.63 | 4 5 4 7 7 24 8 21 14 9 21 19 9 17 1 4
S Alexander 0.59 | 0 0 14 31 28 9 16 45 16 16 16 3 8 25 7 21
C Chambers 0.59 | 5 3 5 5 6 5 23 2 22 9 15 2 12 4 4
E George 0.58 | 6 9 5 19 6 2 11 9 6 7 11 15 3 3 27 6
J Fiedler 0.53 | 23 18 9 10 11 12 14 17 0 31 13 32 0 20 17 18
J Rice 0.47 | 8 0 21 11 4 3 4 8 31 6 15 11 15 8 10 5
T Green 0.47 | 13 6 30 0 3 16 23 3 2 11 11 19 14 11 17 11
E Moulds 0.46 | 3 1 1 16 12 7 3 0 6 31 5 5 5 10 2 8
K Collins 0.45 | 28 0 12 6 10 7 29 24 9 7 7 7 18 19 18 15
C Dillon 0.39 | 19 11 6 6 21 3 38 13 6 6 14 14 23 13 9 20
L Coles 0.31 | 11 2 6 9 19 2 10 10 4 10 12 2 4 6 11 17
T Henry 0.31 | 6 13 3 11 9 9 6 5 4 15 1 21 6
T Banks 0.31 | 2 9 0 11 16 20 17 2 5 20 14 4 9 7 9
J Miller 0.30 | 16 7 -1 21 -1 8 23 3 2 2 16 3 21 12
M Brunell 0.28 | 26 11 -1 15 8 29 15 19 16 14 20 9 14 23 11
D Flutie 0.26 | 0 27 11 6 16 21 23 0 1 1 25 4 22 1 21 16
W Jackson 0.25 | 1 16 9 2 4 11 4 16 2 13 2 10 21 8 2 6
G Hearst 0.20 | 7 2 12 9 11 21 9 16 11 23 20 10 12 7 3 7
D Boston 0.12 | 14 6 7 9 19 12 13 19 20 18 16 13 8 2 24 5
A Toomer 0.09 | 19 5 2 10 6 1 16 6 4 6 7 6 9 12 12 7
I Hilliard 0.05 | 3 2 10 8 10 12 12 10 1 1 4 16 1 9
M Pollard -0.00 | 9 0 12 5 6 6 10 2 18 2 14 2 16 11 3 1
M Harrison -0.01 | 3 32 4 18 21 15 3 35 5 18 14 8 6 18 9 24
K Barlow -0.01 | 1 10 9 4 10 4 7 15 3 4 2 17 4 2 8
R Gannon -0.08 | 23 6 33 15 11 7 30 25 34 27 27 19 12 14 18 15
K Johnson -0.14 | 5 8 8 9 21 10 15 6 5 17 16 9 3 5 13 8
R Dayne -0.16 | 3 10 17 3 15 4 1 8 11 0 2 1 0 10 11 13
P Price -0.18 | 6 14 0 16 0 21 1 11 20 2 1 10 6 7 14 5
L Centers -0.20 | 4 7 4 20 7 1 4 3 7 3 7 1 2 7 9 9
A Freeman -0.22 | 1 9 2 0 19 5 12 12 10 8 10 9 4 2 4 2
A Green -0.23 | 29 14 12 12 13 10 27 11 9 27 16 23 3 16 10 24
C Weinke -0.23 | 17 17 9 8 17 2 0 4 18 10 19 16 20 13 -3
T Brady -0.23 | 2 6 4 26 25 8 28 7 8 35 8 2 7 12 8
H Ward -0.28 | 8 7 6 9 5 11 9 5 17 9 3 13 1 12 5 3
A VanPelt -0.29 | -1 26 0 4 24 27 -5 13 6 14 15 -2
D Patten -0.33 | 4 5 2 1 7 41 10 4 3 9 4 6 4 6 5
J Plummer -0.35 | 14 7 18 15 17 6 14 10 34 10 17 15 16 5 18 3
T Gonzalez -0.35 | 8 4 14 12 4 5 13 5 10 7 2 11 5 2 20 0
J Smith -0.37 | 24 8 8 7 4 17 12 12 6 12 11 17 11 12 5 12
J Garcia -0.37 | 20 18 13 25 44 18 24 33 24 16 21 5 10 13 27 34
B Favre -0.38 | 22 25 24 7 31 15 8 19 13 21 38 12 18 23 6 25
V Testaverde -0.38 | 19 5 8 18 18 6 -1 8 3 8 8 5 19 18 3 9
J Thrash -0.38 | 1 29 2 4 13 0 22 18 1 3 6 7 1 2 21
T Owens -0.40 | 9 5 10 23 36 6 24 23 15 16 13 5 4 12 9 23
K Johnson -0.41 | 7 9 8 14 15 6 6 7 8 5 8 16 11 4 2
E Grbac -0.42 | 17 10 18 16 5 0 10 10 13 19 21 7 8 6
M Pittman -0.59 | 14 9 7 19 4 7 6 8 10 15 8 2 11 11 1
W Dunn -0.65 | 7 15 20 4 3 17 15 16 4 5 5 13 6
J Morton -0.66 | 11 16 4 12 17 7 2 6 15 6 1 5 7 2 9 15
T Brown -0.70 | 19 2 4 11 14 4 22 14 12 23 20 3 4 2 7 5
I Bruce -0.73 | 8 20 3 7 4 4 23 9 13 7 4 5 26 0 7 0
C Chandler -0.74 | 6 18 29 -5 21 17 7 19 5 4 6 21 23 0
T Brown -0.82 | 16 8 3 4 11 18 17 16 4 9 15 5 8 7 6 4
B Johnson -0.85 | 11 7 8 26 16 20 21 13 16 4 9 13 2 27 9 1
M Booker -0.87 | 4 15 19 4 13 6 14 7 34 0 7 1 9 7 11 3
T Couch -0.97 | 4 17 8 16 19 18 17 13 -4 7 9 1 5 8 28 0
Q Ismail -1.02 | 8 6 10 15 7 14 14 5 18 6 4 13 11 3 3 2
B Schroeder -1.05 | 17 10 12 17 3 16 11 3 16 3 14 1 2 16
C Carter -1.07 | 10 1 12 8 17 12 5 14 10 4 3 4 3 8 1 6
D McNabb -1.25 | 26 30 26 22 12 8 21 36 10 7 23 21 14 9 30
L Smith -1.38 | 24 11 6 22 13 9 3 5 6 9 13 11 1 5 0 21
J Kitna -1.39 | 13 18 9 9 15 6 16 9 10 -2 8 2 -4 -2 26 10
C Martin -1.51 | 16 17 21 27 19 6 19 6 30 7 10 15 10 16 17 5
M Smith -1.53 | 29 7 11 9 8 14 7 6 2 8 6 8 11 2
A Brooks -1.56 | 27 16 7 21 14 28 5 27 30 18 30 19 23 17 4 -6
L Tomlinson -1.62 | 24 11 30 17 14 6 8 11 9 14 14 17 6 6 15 10
P Manning -1.68 | 17 43 12 15 19 20 24 22 14 8 22 -1 29 15 10 16
R Smith -1.72 | 19 28 8 17 10 8 21 21 7 8 8 6 16 11 9
R Williams -1.82 | 17 10 23 23 11 9 12 17 22 15 17 14 7 4 8 2
B Griese -2.37 | 32 27 9 9 15 9 16 14 22 13 9 9 8 18 1
OK, enough preliminaries, let's get down to the question at hand, which is: do players who finish strong in year N tend to improve more in year N+1 than players who fizzle in year N do? I looked at all QBs, RBs, WRs, and TEs from 1995-2000 who met the following conditions:
How many Better in Yr N+1 Worse in Yr N+1 Average change
-------------------------------------------------------------------------------------------------
Finished stronger in year N 234 39 pct 61 pct -1.0 FP/G
Started stronger in year N 181 38 pct 62 pct - .6 FP/G
A split like this is nowhere near being significant, as there is a very high probability of obtaining it by chance. In other words, we cannot conclude that there is any general tendency for strong finishers to improve any more or less than weak finishers. I broke it down by position, and still found nothing significant. I broke it down by age, figuring maybe young players who finish strong are more likely to improve, but again I found nothing. You're welcome to investigate the complete data set to see if you can find something for yourself, but until somebody does, I'm going to assume that looking at early- versus late-season performance is as likely to steer you wrong as right. I still like Mason. I just don't like him anymore than I would if his 2001 season had happened in reverse.
|