Tuesday, November 12, 2013

Week 10 - Retrospective

This week I added tier accuracy, as defined as the percentage of higher tiers that actually came in scoring higher than a lower tier.  At a glance this seems like a nice metric that I'll keep track of going forward.

This week we have: QB 90%, RB 86.7%, WR 71.4%, TE 60%, K 33.3%, DST 50%.  This seems consistent with past weeks -- QB/RB/WR great, TE mediocre, K/DST not so great.  I believe WR has been higher in past weeks. 


Disclaimers:
- all points are a little higher than yahoo standard due to my data source (but it evens out)
- please let me know if there are any remaining name mismatches (i.e. A.J. Green vs AJ Green)
- if anyone has a better source than fftoday that uses standard scoring, please let me know!



QB
Tier 1 median: 27.6
Tier 2 median: 23.9
Tier 3 median: 14
Tier 4 median: 16.7
Tier 5 median: 8.9

Tier accuracy: 90.0%

It was a great week for QB's.  Only tiers 3 and 4 got mixed up, and not by much.


RB
Tier 1 median: 11.4
Tier 2 median: 12.8
Tier 3 median: 6.2
Tier 4 median: 7.3
Tier 5 median: 5.1
Tier 6 median: 2.2

Tier accuracy: 86.7%

RB consistently behaves well and has been (to my naked eye of doing this since week 6) the most predictable position.


WR
Tier 1 median: 20.7
Tier 2 median: 8
Tier 3 median: 7.8
Tier 4 median: 4.7
Tier 5 median: 4.6
Tier 6 median: 4.8
Tier 7 median: 4.5
Tier 8 median: 10.8

Tier accuracy: 71.4%

The top 3 tiers of WR each week continue to overwhelm the rest of the pack.  This has been a trend since I started doing retrospectives in week 6.

It was a bad week for a handful of tier 1/2/3 WR's, but on average the tiers still held true. Within a high tier, we expect most (not all) players to score highly.

It's worth noting that without the two tier 8 monstrosities, WR tier accuracy would have been absurd.


TE
Tier 1 median: 5.9
Tier 2 median: 1.9
Tier 3 median: 1.7
Tier 4 median: 1.5
Tier 5 median: 6.1

Tier accuracy: 60.0%

This was an extremely low scoring week for TE's across the board, though tier 1 did noticeably better. TE has been much more noisy than I anticipated all season long.  I've been attributing this to the fact that a higher percentage of TE points come from touchdowns (compared to WR and RB).


K
Tier 1 median: 7
Tier 2 median: 8.5
Tier 3 median: 9
Tier 4 median: 7.5

Tier accuracy: 33.3%

The lowest accuracy of the week, no surprises. Kickers have been the most noisy every week.


DST
Tier 1 median: 6.5
Tier 2 median: 4
Tier 3 median: 3
Tier 4 median: 7

Tier accuracy: 50.0%

The second lowest accuracy of the week.  Also not surprising. This is quite strong evidence for supporting streaming K and DST.



No comments:

Post a Comment