Scatterbrained: Tips & Trends From Week 1

Scatterbrained: Tips & Trends From Week 1

Why “Scatterbrained?”

Every Monday and Tuesday each week, as I sift through the ashes of the NFL slate, I find myself drawn in a hundred different directions. … often, I don’t find the end of those trails. I’ve always struggled with wrapping up projects before something shiny and new engages my passion… this is commonplace with folks managing ADHD. It’s not that I don’t want to – I really, really do – there’s just this idea, and that concept, that call to me right now. I cannot help but indulge myself.

How does this apply for fantasy sports as a hobby? It’s more helpful than you may think. I’ll explain.

An unintended benefit of focus difficulty is that I don’t become beholden to any one concept. ZeroRB? Pfft, been there, done that. Streaming? OK, great, whatcha got for me now? Team Totals? Old news. I’m always looking for the next intellectual fix, the new challenge. This year, I’d like you all to come along with me as I wander along this path. I will pull a few Crazy Ivans, and some weeks you may furrow your brow and ask yourself, “What the hell is he thinking? That’s nonsense!”

And to that, I will happily answer, “You’re damned right, it’s nonsense… but we wouldn’t have known that if someone hadn’t plied the idea first.”

Scattered Thinking About Week 1

I’m in the midst of a new data project – I’m essentially reconstructing my football database from the ground up. The original reason for building my own database was my apprehension over paying for one. There are many great datasets out there already such as Armchair Analysis, Stattleship, nfldb, FootballDB and Pro-Football-Reference that provide highly sorted, immaculately tidied data for a reasonable fee (or free).

I like free, but I’m also lazy. Many of the free data sites require manual table scraping, which is monotony I cannot stand. So, what else would I do but learn R (or Python) and write code to do the scraping for me? And that’s the path I’ve gone down. I’ve been working in R for a year this month, and have come a long way with the language… so much so that I’ve been able to leverage it in my professional life with great success. But for this wonderful hobby, it’s allowed me to start (rapid) construction of my own football database.

As I write this article, I’m five days deep into the build. I’m learning much along the way, and always tinkering with a new way to manipulate this data in hopes of gaining a bit of an edge. So far, here are the concepts I’m considering this week…

Eastbound & McCown

Honestly, this is probably the best thing that could happen for this offense. While Josh McCown is old, and Not Good, he is exactly the type of QB that Hue Jackson can prop up in his scheme and excel. If we learned anything from McCown’s time in CHI (and later, TB), it’s that an inspired offensive mind (like Jackson) can pull the best from him. Naturally, it helps having great-to-elite receiving talent around him who can snare inaccurate throws and make McCown look “good.”

We already know what Hue Jackson will do with his playcalling when CLE is behind in the second half of games… pass, pass, and pass.

cin_split2h_2015 cin_split2h_2014

The first three weeks of the McCown Era in CLE will be rough though, as they await the true alpha WR that will open up their offense. In the meantime, I do expect plenty of deep throws to both Corey Coleman and Terrelle Pryor. I also believe we’ll see an uptick from the planned usage we envisioned for both Duke Johnson and Gary Barnidge.

Week 1 Play Splits

If you read my pieces on QB projections for 2016 (some of which are already rotten), you’ll see where I consider strongly the concept of Game Script Splits. What are offenses doing when scores are within a TD, or beyond that? How does playcalling change throughout the game? I find these are highly important when determining streaming options for the coming week(s) if I have a good feeling about game scripts for particular offenses.

Week 1 saw overall play splits turn out like so:

pl.offtotpapa.00ruru.00stst.00
AVE70.237.653.7%25.836.7%6.89.7%
ARI654061.5%1929.2%69.2%
ATL724258.3%2230.6%811.1%
BAL703854.3%2738.6%57.1%
BUF552443.6%2341.8%814.5%
CAR753648.0%3242.7%79.3%
CHI613455.7%2032.8%711.5%
CIN633758.7%1930.2%711.1%
CLE562951.8%2137.5%610.7%
DAL824554.9%3036.6%78.5%
DEN602846.7%2948.3%35.0%
DET703955.7%2434.3%710.0%
GB663553.0%2537.9%69.1%
HOU793645.6%3544.3%810.1%
IND744966.2%1925.7%68.1%
JAC734257.5%2635.6%56.8%
KC775166.2%1924.7%79.1%
LA703752.9%2332.9%1014.3%
MIA623353.2%2032.3%914.5%
MIN703347.1%2840.0%912.9%
NE713549.3%3143.7%57.0%
NO714360.6%2231.0%68.5%
NYG603050.0%2440.0%610.0%
NYJ743648.6%3040.5%810.8%
OAK723852.8%2636.1%811.1%
PHI823947.6%3441.5%911.0%
PIT703752.9%3042.9%34.3%
SD773748.1%3241.6%810.4%
SEA874652.9%3236.8%910.3%
SF843541.7%4250.0%78.3%
TB673349.3%2841.8%69.0%
TEN724359.7%2230.6%79.7%
WAS594372.9%1220.3%46.8%

Legend:

pl.off – team offense
tot – total plays
pa – pass plays
pa.00 – pass play %
ru – rush plays
ru.00 – rush play %
st – spec teams plays
st.00 – spec teams play %

I’m utterly shocked that Seattle ran 87 plays. What’s not surprising is that they were able to squat on the ball for nearly 35 minutes of game clock (only counting timed offensive plays). Granted, these are very early returns, but the splits above give me reasonable confidence that the philosophical shift in their offense late last season is here to stay. Despite being within a TD the entire game, the Seahawks elected to throw prolifically… even with Russell Wilson struggling on one wheel.

Defensive Deficiency Metrics

Another performance measure I’ve engrossed myself in deals with defensive efficiency against teams, positions, and situations. I haven’t moved very far into this study yet, but I’d like to share a snippet of where it’s going:

dfpa_app2

dfpa_app

Legend:

TM – offense to observe
wks – number of weeks to use in Adjusted Aggregate Metric
spr – spread
QB – QB Adjusted Aggregate Metric (AAM)
dQB – QB AAM differential, aggregated for that week per the look-back
RB – RB Adjusted Aggregate Metric (AAM)
dRB – RB AAM differential, aggregated for that week per the look-back
WR – WR Adjusted Aggregate Metric (AAM)
dWR – WR AAM differential, aggregated for that week per the look-back
TE – TE Adjusted Aggregate Metric (AAM)
dTE – TE AAM differential, aggregated for that week per the look-back

Above is a brief glimpse at an app I’m writing to identify the defenses we want to target on a weekly basis. The numbers presented above are fantasy points conceded to each of the primary positions, aggregated and average-adjusted by week, against a user-selected time frame (in this case, 10 weeks back in time). Also included are home/road splits, as well as fave/dog splits. This is nothing new, but like nearly everything, I don’t take someone’s word for it without doing it myself first to understand the method and math. I advise you don’t accept data at face value, either.

What’s not shown are the defensive adjusted-aggregate metrics (DAAM?), other than fantasy points against, that I intend to track with respect to projected offensive success. I’ll reveal those as they’re ready here in Scatterbrained, and I think you’ll find them useful.

Random Bits, Lying About

  • Washington elected to punt only once against the Steelers (4th & 1 from the PIT 40), and it set the fantasy world ablaze. It was admittedly baffling, considering Jay Gruden later called a pass play on 4th & 6 from the PIT 38.
  • Teams that won games despite holding the ball less than their opponent:
    • NYG (23.2 min)
    • CIN (25.2 min)
    • KC (26.2 min)
    • DEN (27.8 min)
    • TB (27.8 min)
    • GB (28.7 min)
    • MIN (29.9 min)
  • This makes me giggle maniacally:
    (13:51) (No Huddle, Shotgun) B.Gabbert pass short left to B.Gabbert to SF 15 for -16 yards (T.McDonald). Caught SF19 -4 yrds. YAC Gabbert caught a deflected pass (McDonald) behind the line of scrimmage.
    You almost feel bad for the kid, out there trying to make a play. I always wonder about players that can’t help themselves but catch a ball batted backwards when a play is destined for sadness. The quickest thinkers slap it to the turf and move on. Clearly, Gabbert isn’t MENSA material.
  • Highest Pass Percentages for teams largely ahead in their games.
    • NO (60.6%) – Philosophy
    • CIN (58.7%) – Path of Least Resistance
    • DET (55.7%) – Philosophy
    • GB (53.0%) – Philosophy
    • SEA (52.9%) – New Philosophy
    • PIT (52.9%) – Philosophy
  • Largest Play Split Disparities – expect these to move toward more balance (less QB opportunity).
    • WAS (43 P / 12 R / 4 ST)
    • KC (51 P / 19 R / 7 ST)
    • IND (49 P / 19 R / 6 ST)
    • ARI (40 P / 19 R / 6 ST)
    • TEN (43 P / 22 R / 7 ST)
    • CIN (37 P / 19 R / 7 ST)
    • DAL (45 P / 30 R / 7 ST)

Tidying Up

That will do it for this week’s scattered thoughts. Next week, I’ll unveil more of the DAAM as well as dive a bit deeper into how game scripts are shaping up so far this season. And, naturally, whatever shiny objects that catch my eye along the way.

As always, feel free to find me on Twitter (@FantasyADHD) to challenge the scattered brain. See you next week!

Josh Hornsby

Josh Hornsby leads engineering teams in the oil & gas industry. His background in new product development, combined with nearly 20 years of data-driven fantasy experience, compels him to think outside the box and wreck the echo chamber of current fantasy analysis. Josh loves to challenge popular thinking and typically does so with numbers in hand. You can find him on Twitter @FantasyADHD

Latest posts by Josh Hornsby (see all)



1 thought on “Scatterbrained: Tips & Trends From Week 1”

Leave a Reply

Your email address will not be published. Required fields are marked *