I’ve started this
piece about six different times and I’m not sure how to start it… so we’ll jump
into the process and work from there.
Process
For the last couple
year’s through Twitter I’ve found new sources of fantasy information and
discourse and one idea took hold. The idea of fantasy football, in this case
yearlong, still boils down to a weekly game. The idea being that weekly
variance (high ceiling, low floor), can be just as valuable as consistent but
unspectacular week to week value. I try to incorporate that idea by combining
two numbers. Both sets of numbers start from the same place with the baseline
for each position group being the last starter (based on a 12-team league). For
example, if a player was the 24th best running back each week of the
fantasy season that he played he’d come out as a flat 0 for both sets of
numbers. From that baseline, I take the players score for the week, take the
difference from the baseline and divide that number by the standard deviation
of the position group for the week. I then take and average to find a value.
The first set of numbers doesn’t punish for missed weeks, so if a player played
2 weeks or 13 it won’t affect their average, the second set does punish for
missing games replacing them with 0s. After averaging up both sets of numbers I
bring them together with a weight of 75-25 with the higher weight on the weeks
played average over the missed games punishment average. For sake going forward
I’m going to call these PAR (think golf) and APAR (adjusted PAR). So, once I
have scores for everybody with a pulse and making sure to include everyone in
my final rankings I get at least some idea of how close I looking forward.
Results
I’m going to break
this down in a bit but let’s look at the top 25 position players as adjusted
compared to their peers. We’ve discussed PAR & APAR and AAPAR is the
combining of those two. “Pos Rank” is my position rank in my last rankings, ADP
is the final ADP I recorded and the “Pos ADP” is just the ADP broken down in
each position.
|
Player
|
Pos
|
PAR
|
APAR
|
AAPAR
|
Pos
Rank
|
ADP
|
Pos
ADP
|
|
David Johnson
|
RB
|
1.87
|
1.99
|
1.9
|
7
|
5
|
2
|
|
Le'Veon Bell
|
RB
|
1.79
|
1.28
|
1.66
|
5
|
11
|
5
|
|
Ezekiel Elliott
|
RB
|
1.62
|
1.72
|
1.64
|
3
|
8
|
4
|
|
LeSean McCoy
|
RB
|
1.33
|
1.23
|
1.31
|
13
|
27
|
12
|
|
DeMarco Murray
|
RB
|
1.11
|
1.18
|
1.13
|
23
|
45
|
18
|
|
Melvin Gordon
|
RB
|
1.19
|
0.90
|
1.12
|
16
|
63
|
24
|
|
Jordy Nelson
|
WR
|
1.02
|
1.05
|
1.03
|
10
|
17
|
8
|
|
Antonio Brown
|
WR
|
0.98
|
1.01
|
0.99
|
1
|
1
|
1
|
|
Odell Beckham Jr
|
WR
|
0.96
|
0.99
|
0.97
|
3
|
2
|
2
|
|
Mike Evans
|
WR
|
0.95
|
0.99
|
0.96
|
8
|
23
|
10
|
|
Julio Jones
|
WR
|
0.97
|
0.67
|
0.89
|
2
|
4
|
3
|
|
T.Y. Hilton
|
WR
|
0.85
|
0.89
|
0.86
|
13
|
30
|
15
|
|
Travis Kelce
|
TE
|
0.77
|
0.83
|
0.79
|
5
|
58
|
4
|
|
Devonta Freeman
|
RB
|
0.77
|
0.81
|
0.78
|
8
|
15
|
7
|
|
LeGarrette Blount
|
RB
|
0.76
|
0.80
|
0.77
|
34
|
100
|
36
|
|
Jordan Howard
|
RB
|
0.78
|
0.69
|
0.75
|
93
|
197
|
58
|
|
A.J. Green
|
WR
|
0.89
|
0.13
|
0.7
|
4
|
9
|
5
|
|
Aaron Rodgers
|
QB
|
0.70
|
0.51
|
0.66
|
2
|
28
|
2
|
|
Jay Ajayi
|
RB
|
0.65
|
0.56
|
0.63
|
31
|
102
|
37
|
|
Brandin Cooks
|
WR
|
0.63
|
0.65
|
0.63
|
17
|
29
|
14
|
|
Carlos Hyde
|
RB
|
0.64
|
0.45
|
0.59
|
26
|
38
|
16
|
|
Latavius Murray
|
RB
|
0.61
|
0.43
|
0.56
|
14
|
34
|
14
|
|
Greg Olsen
|
TE
|
0.53
|
0.57
|
0.54
|
3
|
44
|
3
|
|
Drew Brees
|
QB
|
0.56
|
0.39
|
0.52
|
4
|
47
|
5
|
|
Dez Bryant
|
WR
|
0.61
|
0.26
|
0.52
|
9
|
13
|
6
|
Some interesting
things pop up here, first we notice that the top running backs were pretty
beastly when compared to their peers with five players returning greater than a
standard deviation (around 2/3rds of a sample should fall in that range) in
both categories. The receivers only had one such representative which suggests
the mid-range receivers where much stouter than mid-range running backs. To
break this down a bit more the top 25 performers comprised of 12 RBs, 9 WRs, 2
TEs, and 2 QBs. The top rep for the latter two were Travis Kelce (13th)
and Aaron Rodgers (18th). All-in-all in generally lines up with how
we think of positional value, WR/RB and then TE/QB, with minor quibbles in the
middle.
Moving towards more
focused positional breakdowns we’ll start with how well my final ranks
correlated with how it all sorted itself out. On face value I was pretty damn
good on QB and TE this season and was pretty good on WR and RB. In general, a
positive correlation is coefficient + or - .5 is generally good. QB was my most
accurate set with a correlation of -.735 (negative because the closer to 0 the
rank the higher the points so it goes from top left to bottom right on a
graph), TE was second best with -.729, in the past WR has been my strong suit
and this year it wasn’t bad with a correlation of -.687 and RB was my roughest
of the skill positions at -.678. Overall none of those are bad and I’d say it
was damn good on the whole. You may be wondering about DEF and K and to be
honest those are both usually around -.200 correlation and you’ll never hear me
say I can properly figure out either groups fantasy output.
Next is a look at the
#1s (Top 12s) for each position starting with the QBs…
|
Player
|
Pos.
|
PAR
|
APAR
|
AAPAR
|
Final
Rank
|
Pos
ADP
|
|
Aaron Rodgers
|
QB
|
0.7
|
0.512
|
0.66
|
2
|
2
|
|
Drew Brees
|
QB
|
0.56
|
0.387
|
0.52
|
4
|
5
|
|
Andrew Luck
|
QB
|
0.58
|
0.29
|
0.51
|
5
|
4
|
|
Matt Ryan
|
QB
|
0.48
|
0.344
|
0.44
|
23
|
19
|
|
Kirk Cousins
|
QB
|
0.26
|
0.181
|
0.24
|
17
|
13
|
|
Tom Brady
|
QB
|
0.36
|
-0.338
|
0.19
|
8
|
7
|
|
Ben Roethlisberger
|
QB
|
0.1
|
-0.053
|
0.06
|
6
|
6
|
|
Blake Bortles
|
QB
|
0.07
|
0.047
|
0.06
|
7
|
10
|
|
Dak Prescott
|
QB
|
0.07
|
0.051
|
0.06
|
26
|
23
|
|
Philip Rivers
|
QB
|
0.04
|
0.028
|
0.04
|
9
|
11
|
|
Jameis Winston
|
QB
|
0.04
|
0.033
|
0.04
|
13
|
15
|
|
Cam Newton
|
QB
|
0.03
|
-0.096
|
0
|
1
|
1
|
So, from the start
let’s point out that Cam was technically a QB starter but he wasn’t close to
what I or the rest of the fantasy community expected, I’ll take my lumps on
being four spots worse than ADP on Ryan and Cousins, the former had a monster
year under Shanahan and the latter I focused too much on what I think of as a
player and less on his on-field opportunity. Dak was obviously the positive surprise
in this category, Blake Bortles was weirdly useful, and other than that it was
a pretty predictable year on the whole. It should be noted that Russell Wilson
was my 3rd (ADP of 3 as well) ranked QB heading into the year. QB
was tidy up top, but what about their backfield mates?
|
Player
|
Pos
|
PAR
|
APAR
|
AAPAR
|
Final
Rank
|
Pos
ADP
|
|
David Johnson
|
RB
|
1.87
|
1.99
|
1.9
|
7
|
2
|
|
Le'Veon Bell
|
RB
|
1.79
|
1.28
|
1.66
|
5
|
5
|
|
Ezekiel Elliott
|
RB
|
1.62
|
1.72
|
1.64
|
3
|
4
|
|
LeSean McCoy
|
RB
|
1.33
|
1.23
|
1.31
|
13
|
12
|
|
DeMarco Murray
|
RB
|
1.11
|
1.18
|
1.13
|
23
|
18
|
|
Melvin Gordon
|
RB
|
1.19
|
0.9
|
1.12
|
16
|
24
|
|
Devonta Freeman
|
RB
|
0.77
|
0.81
|
0.78
|
8
|
7
|
|
LeGarrette Blount
|
RB
|
0.76
|
0.8
|
0.77
|
34
|
36
|
|
Jordan Howard
|
RB
|
0.78
|
0.69
|
0.75
|
93
|
58
|
|
Jay Ajayi
|
RB
|
0.65
|
0.56
|
0.63
|
31
|
37
|
|
Carlos Hyde
|
RB
|
0.64
|
0.45
|
0.59
|
26
|
16
|
|
Latavius Murray
|
RB
|
0.61
|
0.43
|
0.56
|
14
|
14
|
Alright
the RB 1s and the first thing we notice is that I need to be shamed because of
that man at the top. I was too skeptical of David Johnson this season and I readily
admitted such a possibility in the run up to the season. That said, skepticism
went much too far. Next up Demarco & Hyde were much better than the fringe
RB 2s that I thought they would be leading into the season and they fit the
same mold as players I wasn’t sure would handle the lead this season well.
Howard was a miss because I thought he would be chasing Langford all year, and
lastly Ajayi lost his job before the season and Blount well I have no idea if
that’s repeatable with Lewis back next season but he greatly surpassed what I
thought of him. I will take the congrats on Melvin Gordon and his performance before
a crushing injury late in the season and having him 8 spots higher than
consensus. A good quarter of the RB 1s were relative surprises, but the biggest
surprise was the dismal performance of Gurley and the Rams rushing attack. I
say this even though I knew damn well the O-line and QB were going to be shit.
I thought he’d be able to keep doing what he did last season and he will be one
of the trickier ranks next season. Now let’s head to the end of the line and the
big bodies at TE.
|
Player
|
Pos
|
PAR
|
APAR
|
AAPAR
|
Final
Rank
|
Pos
ADP
|
|
Travis Kelce
|
TE
|
0.77
|
0.83
|
0.79
|
5
|
4
|
|
Greg Olsen
|
TE
|
0.53
|
0.57
|
0.54
|
3
|
3
|
|
Delanie Walker
|
TE
|
0.47
|
0.37
|
0.44
|
2
|
5
|
|
Rob Gronkowski
|
TE
|
0.65
|
-0.37
|
0.4
|
1
|
1
|
|
Jordan Reed
|
TE
|
0.51
|
0
|
0.38
|
4
|
2
|
|
Jimmy Graham
|
TE
|
0.33
|
0.36
|
0.34
|
19
|
13
|
|
Kyle Rudolph
|
TE
|
0.24
|
0.25
|
0.24
|
9
|
21
|
|
Tyler Eifert
|
TE
|
0.44
|
-0.5
|
0.2
|
12
|
8
|
|
Cameron Brate
|
TE
|
0.13
|
0.14
|
0.13
|
23
|
29
|
|
Martellus Bennett
|
TE
|
0.03
|
0.03
|
0.03
|
10
|
12
|
|
Zach Miller
|
TE
|
0.16
|
-0.39
|
0.02
|
16
|
16
|
|
Eric Ebron
|
TE
|
0.06
|
-0.25
|
-0.02
|
11
|
18
|
Let’s
start off with the fact that I hit on 9/12 TEs in the top 12 and went 5/5 on
the top five. Of course, so did ADP. On the other four I hit ADP had me beat on
Eifert 12-8, but I hit on Bennett 10-12, Rudolph 9-21 and Ebron 11-18. I’m
pretty happy here Brate was gaining steam heading into the season, but ASJ’s
talent was a quagmire to deal with, Graham was coming off an achilles injury
and there wasn’t an injury expert who knew how that’d work out, and Zach Miller
was a slightly better than expected. The three I missed on were Fleener (7),
Allen (8), and Barnidge (6). Overall, I’m happy here and there wasn’t much embarrassment
on my part. And last but certainly not least from a fantasy perspective… The
WRs.
|
Player
|
Pos
|
PAR
|
APAR
|
AAPAR
|
Final
Rank
|
Pos
ADP
|
|
Jordy Nelson
|
WR
|
1.02
|
1.05
|
1.03
|
10
|
8
|
|
Antonio Brown
|
WR
|
0.98
|
1.01
|
0.99
|
1
|
1
|
|
Odell Beckham Jr
|
WR
|
0.96
|
0.99
|
0.97
|
3
|
2
|
|
Mike Evans
|
WR
|
0.95
|
0.99
|
0.96
|
8
|
10
|
|
Julio Jones
|
WR
|
0.97
|
0.67
|
0.89
|
2
|
3
|
|
T.Y. Hilton
|
WR
|
0.85
|
0.89
|
0.86
|
13
|
15
|
|
A.J. Green
|
WR
|
0.89
|
0.13
|
0.7
|
4
|
5
|
|
Brandin Cooks
|
WR
|
0.63
|
0.65
|
0.63
|
17
|
14
|
|
Dez Bryant
|
WR
|
0.61
|
0.26
|
0.52
|
9
|
6
|
|
Doug Baldwin
|
WR
|
0.48
|
0.5
|
0.49
|
18
|
22
|
|
Davante Adams
|
WR
|
0.48
|
0.51
|
0.49
|
75
|
80
|
|
Michael Thomas
|
WR
|
0.48
|
0.37
|
0.46
|
40
|
50
|
First
thing’s first, the consensus gets a slight edge on the Nelson Evans switch but
that’s a virtual tie and not much of an embarrassment for either side. Same
could be said for the Hilton-Cooks flop. I hit Dez right on the nose, and was
higher on the three that snuck in the bottom of the #1s. I and the consensus
missed bad on Adams, who hit about where I expected Cobb to bounce back to, I
was 10 spots higher on Thomas who came out like gangbusters as a rookie and I
was a bit higher on Baldwin having him as a solid #2 WR instead of a fringe #2.
Overall not bad at the top here either, the biggest misses being the shared
experience of Robinson/Hopkins disasters and the hellscape that was the Kennan
Allen injury for me personally. I hope to regain what was a thoroughly
enjoyable run at WR.
In summation,
it was a good year for my preseason predictions, there were some slip ups and
foible everywhere but on par, get it, I was happy with the end results. Next
year luck and continued improvement permitting this column will be even more upbeat.
Thanks for reading in the preseason, we’ll see if I add regular season ranks
weekly ranks next season but that’ll depend on how I approach my weekly college
football articles. So on that note, thanks again for reading anything of mine
you have and now it’s onto draft season and prospect discussion.
