Week 15 Reflections
Priors and Profit
Welcome to Throw The Dam(n) Ball!
If this is your first time reflecting, I'd recommend first reading the inaugural piece of this Reflections series to get a sense of what I'm trying to do in this space.
This week we cover: Macro Thoughts on overreacting, Sifting through priors when betting and my Bets+ DFS breakdown. Skip to those parts per your hearts desire!
Macro Thoughts:
I wrote in week 10 about the value and need to sit out as a bettor. In week 10 it was because my views were quite aligned with the market, but this week it was because there was so much uncertainty with COVID situations. But unlike week 10, there was one spot I had conviction: GB O would dominate the BAL D. It was one my highest conviction thesis of the season. So on that assumption, I decided to attack this slate trying to capitalize on that thesis in multiple markets while sitting out the rest.
With the pretty shocking TB & AZ losses, we are once again reminded that variance and randomness is the name of game in the NFL. Reading too much into one game or games is a recipe for long-term failure. This is why we capture season long stats— to see the wide range of outcomes and understand if trends are emerging or if we’re dealing with noise. And rather than “take a wait and see approach” to see if one game is variant, we should take advantage of market volatility(big time off-season project about profiting off volatility) and profit off recency bias. Did we really learn anything new about AZ this week? About TB? Sure it’s a game they both had to win, but do we expect them to play so poorly every week? Should we really be discounting AZ to a tier 2/3 team as market has done? Or can we recognize teams have bad games because winning in the NFL is exceedingly difficult and one game should never alter our priors too much.
I Did not mean to rant for this long, but bottom line: I can’t emphasize enough how our biggest edge as bettors—and even in just understanding market and team value as a casual fan—is understanding noise and sample sizes and, for bettors, to profit off market volatility which unduly prices change and certainty(as we talked about last week).
Team Trends & Quick Hitters
SF:
I don’t care what Kyle Shanahan could do with a “better” QB or how could Jimmy G is in a vacuum as though he were a Madden grade. These are fun academic projects and I enjoy the discussions and conversations like anyone else. But these conversations are not primary when our goal is to profit.
Point is: in his context and scheme in SF, Jimmy G is producing at an elite level and leading his offense to elite results. Jimmy G is: 1st in Yds/attempt, 2nd in RBSDM’s EPA/CPOE composite, and is #1 in PFF grade since his return from injury. As you see above, 4 of his best SF career games have come since his return from injury. And the team results have followed a similar trajectory
Following Jimmy G’s return, SF is best in Drive EPA, Yds/Play and 9th in SR. This has also coincided with unleashing Aiyuk from the doghouse, and Kittle returning to health, giving this SF team their full compliment of weapons. They are #good.
GB: Red Zone Efficiency
Sure Red Zone(RZ) stats are often noisy, but in the first 2 years of the Lafleur era, GB has been elite converting RZ possessions into TD’s.
GB RZ TD% 2019: 68%(2nd)
GB RZ TD% 2020: 77%(1st)
But they’ve struggled in this area so far this season.
GB RZ TD% 2021 Pre Week 13 bye: 55.3%(25th)
GB RZ TD% 2021 Post Bye: 71.43(5th)
I think it is only a matter of time before they regress and the 2019-2020 RZ numbers converge with the 2021 pre bye numbers. And the result will be more points piling up for GB.
PHI: Good in the Chart, Good at Football
Since the philosophical change to embrace the Run & Gun style in week 8, PHI is Top 5 in: Drive EPA, SR, Yds/play, explosive play rate, and #1 in Early down SR— perhaps the most stable of all metrics. And outside of a clunker against the NYG, these results are showing up on the point sheet: (44,24,30,40,7,33,27).
And the defense is top 10 in each of those above categories as well. Don’t look now, but the PHI could spell trouble come playoff time.
DAL: Mirror of KC’s Seasonal Arch:
I wrote last week about how the Cowboys defense and offense have been trending in opposite directions. (The D is really moving up the Y axis of our favorite graph!) And my thought was: this is exactly like the Chiefs seasonal arch! But the difference with DAL and KC is that DAL’s D jump is from 15th best to best and KC’s was from 32nd to 12-16th. But I guess I was in good company with my call…
LAC & KC: Quick hitters:
Considering that both teams are labeled as “fun” in our Drive EPA Graphs, is it a surprise that this was an incredibly entertaining game? In all seriousness, though, I don’t think we learned anything new about LAC. They are highly highly variant, as we’ve often spoken about. Sometimes the 4th downs and big plays go LAC’s way and sometimes they don’t. All this is to say: the floor and consistency game-to-game might be shaky, but they have elite ceiling and the ability to win any game no matter the opponent and they showed that once again this week.
On the KC side, I think this has me one step closer to buying them as the same KC of old, but I’m not there just yet. Another dominant performance will have me back. This team is still not having the same early success and consistency which defined the KC team’s of old.
Bets:
GB O 30+200
I mentioned earlier that GB scoring points was my primary thesis on the slate, and what better way to capture that thesis then minimizing the variance introduced betting the spread or game totals, and instead betting the team over—moving the line to grab some extra juice!
My prior plays a huge role in this bet and I will thoroughly explicate my process from prior to profit(hey, that’s kinda catchy: “Throwthedamball, from prior to profit”). And I do so because priors are kind of tricky. If you hold onto a dead prior for too long you will be buried. At the same time, though, maintaining a strongly held prior can keep you grounded to short term noise. Bottom line: By fulling fleshing out the ideas, we can—I hope— start to get a sense of whether my prior is logical & or bearing out in the data, or if I’m holding onto a sinking ship. Let’s dive in.
Prior: Given that GB has changed little on offense from last year, between players, coaching and scheme, we know they are capable of elite offensive production. They finished last year—as you see above— far and away best in NFL in Drive EPA at 1.94. (this years leader is at .96). That is not to say because they were good last year they will be good this year. But we know they can reach realistically reach this elite offensive ceiling.
And in recent weeks they have begun to play more like the 2020 team. So while it is only a small sample in the 2021 season, it is not such a small if we work under my prior. All of this is to say I am putting more stock in their recent dominant stretch because of last year and the similarities in team, scheme, coaching, and personnel.
Following week 10, GB is 1st in my go-to efficiency categories:
1st in Yds Play by .4 yards
1st in Drive EPA by .14 points per drive
1st in Success Rate by 1%
Rodgers is 1st in PFF Passing grade
So the question remains: if this team is more like the 2020 version, why wasn’t that profile showing up early in the year? Fair question. And I think there are at least 2 reasons: poor RZ and Deep Ball efficiency. While generally unstable but critical areas—RZ efficiency and deep ball efficiency— Rodgers and GB will often be outliers in their success, further suggesting the first part of 2021 and their less dominant profile was fluky more than signal. 2020 and the recent success is the sample to put more stock in. But let’s take a closer look:
First, the RZ efficiency listed above under “Team Trends”— this team was uncharacteristically not converting RZ trips into TD’s. The regression was going to hit and they were going to start scoring TD’s at a higher rate like previous GB teams.
Second, and something we’ve noted in this space: Rodgers through 10 weeks had been atrocious throwing deep: he was 32nd in PFF deep grade(69.2), with a 30% completion rate, throwing for 11.4 YPA. These stats are well well below his career averages. Naturally, in the weeks since then, he is 5th in PFF deep grade(93) throwing for a 45% completion rate, and 16.5 YPA. These are huge differences in efficiency and are much more in line with his career averages. Regression was and is coming.
Back to the bet and thesis: This team is peaking with one the NFL’s best offenses regressing to their mean. And their implied total was not the usual 30-31 for NFL’s best O’s but only 24. The price was very much off in my view.
But the total was also suppressed because this Ravens D has been overachieving in terms of points allowed(9th best), despite a weak analytical profile (To save all of you, I’ve been beating this drum for weeks so I will not go into too much depth). BAL came into the game only allowing 22 points a game, despite allowing 10th worst Drive EPA, 4th most yds/play and worst Explosive Play Rate. They have few points per game allowed, in part, because teams have run 2nd fewest plays against them. But GB is efficient without running a high number of plays(20th most plays run in NFL). Not to mention, this Ravens D was decimated by injury.
Add it up and we had a peaking offense facing an overrated defense. GB would have no issue attacking deep where BAL has been at their worst, piling up points in the process. It was trends we’ve been identifying for weeks playing out in this one spot.
Why this alt number? 30 points allows us to be on the right side of a key number in 31, and is right around the range of where Vegas prices elite offenses on a given week. At +200? sign me up!
3 team Teaser: GB -1, TEN +8.5, SF -2. +160
3 Team Teaser:
TEN+8.5 @ PIT
List of teams PIT has beaten by more than 8 this year….
That’s right, zero. And I don’t think this is random— they lack the talent on either side of the ball to build and extend leads and there was no real angle to say that they would start building and extending leads vs TEN.
ATL @ SF -2:
Looking at the Drive EPA graph tells the story of ATL— they are bad. Football Outsiders has ATL as the worst team in the NFL and most power rankings similarly have them as a basement team. On the SF side, there’s no need to repeat what I said earlier, SF is #good and moving in the right direction. On the right side of 3, this was a perfect teaser leg in my view.
GB -1 @ BAL
My main thesis, of course, was that GB would put up points, but that would mean they would likely win the game as well. Adding GB to the teaser leg on the right side of 3, even if not my main thesis, was an obvious second order effect of the thesis and added some upside for 2 other legs I thought were particularly intriguing.
DFS:
General note:
I try not to pay attention to ownership projections and only consult a few articles/sites. I hope to avoid group-think and overthinking my rosters based on perceived ownership projections. As a single entry player, my goal is to put out a lineup I think will score the most points every Sunday. It is not to be over/underweight players.
I write this because my lineup is a perfect mix of serious chalk and very very low owned plays. I have no sense of how the ownerhsip will turn out week-to-week but who cares!?
The GB DFS angle is capturing upside for the main thesis if the team went for 42+. With dynamic scoring, DFS is the proper market to capitalize should this team erupt. I was going to align myself for profit on the chance this thesis (30-35 points)be one standard deviation off and this team go for 42-45.
I wasn’t going to start guessing between Jones and Dillion and between MVS and Lazard or even a random game from St. Brown. But in hindsight, I probably should have guessed because there are no paths to 45 points without other players having huge scores. A mistake. Anyway, I chose Deguara as just a cheap option at a TE— it seemed that all TE’s were a guess this week. I’ll take the savings and hope for a TD or 2.
Brown was priced down as this BAL offense has been inefficient in recent weeks, but he was going to see enormous target volume and always has a downfield role. At a shaky position and price point, I was going to take an upside shot with Hollywood.
Lamb: Lamb has been Dak’s favorite WR, seeing 8+ targets a week. With Adoree Jackson out, NYG had nobody with the speed to cover Lamb. In a spot where I thought the NYG had no chance at offensive success, I thought DAL would run a high number of plays. With a great matchup and higher play count, I thought this was a better spot than usual for a Lamb blow-up game. The dude is just # good and he was a path to ceiling on a slate with little ceiling.
5.4K is too cheap for Robinson who was entering a workhorse role—on a run first team in a good matchup as a favorite. And 4.3K was too cheap for a Parker who was likely to see 7-9 targets with Waddle out. He is a go-to-target deep and in the red zone giving him clear angles for ceiling against a helpless Jets team.
RB was so weak this week that I was more willing to stomach the low floor of Michael Carter and spend up a bit more on defense. Carter had questions both because of gameflow and workload, but if the Jets went down early, I thought Carter had immense receiving ceiling(9 and 14 targets earlier this year in heavy pass games). There was uncertainty for most RB’s this week and I thought the Carter+Niners combo rather than a 5.5K back and CAR D provided me with a better floor and ceiling on the chance Carter was featured again or found himself in pass heavy game script.
Weekly Results Recap:
Moved up to 62/918 in the Prediction game, very solid for the bragging rights. But I should also note @deepvaluebettor has done some great working showing how the skills in the game translates to real life (sport)folio management and , ultimately, the PNL. I’d highly recommend checking it out on twitter.
Bets are up to 25U on the year after another successful week
DFS: narrowly missed out on our 2x goal which has become a theme this year. But DFS is about those tails— just trying to position myself to bink one soon!
That’s all!
Don’t want to miss a reflection or any other article? Subscribe right below and get these straight into your inbox!
And you can follow and support my work by sharing this substack and giving me a follow on Twitter @throwthedamball.
As always, my DM’s are open and I really love all kinds of suggestions, comments, criticism! See ya next week!