Some Pseudo-Random Thoughts
I could go through the exercise of writing a cohesive introduction. Instead, I’ll just list some items and hope that you can see the thread along which I’ve mentally strung these pearls.
Last Week
Last week, I wrote a piece about early-season suckage and mean reversion in NFL DFS. Using our Trends tool, I found that players who do poorly at the beginning of the season experience not only negative Salary Change. They also later in the season experience positive mean reversion and exceed their salary-based expectations — even the ones they failed to live up to earlier in the year. That’s some pretty useful information.
Last Night
Last night, I edited Ian Hartitz’s (strong to quite strong) Seattle Seahawks preview. In it, he (rightly) lists Jermaine Kearse (and not Tyler Lockett) as the team’s current No. 2 wide receiver. As much as I like Lockett, I can’t ignore that he has received only 39 of Seattle’s 75 first-team snaps in the preseason, compared to 63 and 68 for Kearse and Doug Baldwin. Still, I can’t help but think that by the end of the season Lockett will solidly be the No. 2 receiver, because . . .
Last Year
Last year, Lockett saw only 61.45 percent of the team’s snaps (per Pro Football Reference). Kearse saw 71.36 percent. Even with that disparity, Lockett had one more target, two more receptions, and one more touchdown than Kearse. He trailed Kearse by 21 yards receiving — but, while Kearse ‘only’ caught the ball, the versatile third-round rookie contributed in other ways, rushing the ball five times for 20 yards and adding two extra touchdowns to his stats as an All-Pro return man. He finished 2015 with 169.4 DraftKings points. Kearse, 150.5.
People know that Lockett was good last year, especially for a rookie, but he was better than many people know. He led all rookie receivers in touchdowns — and what makes his production especially impressive is that most of it came in the second half of the year.
After Seattle’s Week 9 bye, Lockett became a more regular part of the first-team offense. In Games 9-16, he never once saw fewer than 54.2 percent of the snaps, usually receiving between 65 and 75 percent of the snaps in that time. Here’s what he did with those snaps.
Lockett wasn’t just more productive than Kearse in that span. According to RotoViz’s Game Splits App, in the second half of the season he was also more productive than even . . . dare I say it . . .
You can’t accuse me of beating a dead horse because this horse ain’t e’en near dead yet.
I’m not saying that you should expect more from Lockett than Cooper this year. I’m just saying that over the second half of last year . . .
A Long Time Ago
The first article that Fantasy Douche ever published at RotoViz — back when he and other editors insisted that the ‘R’ was lowercase — explored this basic question: Do we get more by looking at second-half data than we get by simply looking at data from the entire season? In other words, which data set means more for the following season in terms of fantasy points per game?
Naturally, Douche created an app to answer this question. He even built in a filter so that the data could be sorted by age, because . . .
I can anticipate an objection from someone who might say “Yeah, but maybe the second half of the season tells us more about younger players and whether they will break out the following year.”
Touché, La Douché, touché.
And Douche’s implications were right: The larger data set is better. In general, this year’s wide receiver results will be more correlated with all of last year’s data than with merely the data from the second half.
So I probably need to seek immediate medical help due to the engorged four-hour problem that I have for Lockett, but there are still some larger questions regarding rookie wide receivers, second-half performance, and daily fantasy sports.
Let’s explore some of these questions.
An Intermission
F*ck. That was like a 700-word introduction. This might be the first FantasyLabs article ever in which an introduction is longer than the article.
Who am I kidding?
I’ve still got at least another 1,000 words to go.
Do Rookie Wide Receivers Do Better in the Second Half of the Season?
Using The RotoViz Screener (and then my mediocre spreadsheet skills), I found that all rookie wide receivers have combined to have the following splits since 2000:
• Weeks 1-9: 6.40 points per game (points-per-reception scoring)
• Weeks 9-17: 6.32 PPR points/game
Here’s one item to note: The Weeks 1-9 cohort has 439 rookies playing a total of 1,937 games. The Weeks 9-17 cohort has 525 rookies playing 2,487 games.
What this means is that in the second half of the season for the last 16 years we’ve seen a lot of rookie wide receivers who didn’t see any action in the first half of the season.
What I’m about to say I can’t back up — because that would require committing hours that I don’t have to doing research that I don’t want to do — but I hypothesize the following:
1) The inexperienced rookie wide receivers who see action only in the second-half account for the 0.8 point/game difference between the two samples.
2) These inexperienced second-half wide receivers actually cover up the real trend: The rookie wide receivers who actually matter tend to do better as the season goes along.
But I can’t prove that using this app.
Maybe I can get there by asking other questions . . .
What About DFS?
This is where we will (hopefully) find some traction. Do rookie wide receivers get better as the season goes along — both in reality and in DFS? That is, do rookie wide receivers become more productive as they gain experience, and in doing so do they provide more value to DFS players as the season progresses?
Let’s go to the Trends tool. Well, first, let’s go to PFR and find all the 2014 rookie wideouts to play in a game. OK, now let’s use the Trends tool to screen for 2014 rookie wide receivers who matter — at least a little bit. Let’s screen for the guys who are projected to score at least one DraftKings point.
Here’s how the 2014 cohort of rookie wideouts does as a whole:
That Plus/Minus surprises me a little. The rookie wideouts we project to contribute to their offenses actually provide DFS value across the season.
Let’s get at these guys based on Games 1-8 and Games 9-16.
First half of the season:
That’s not what I expected. Let’s continue.
Second half:
Well, that’s kind of what I expected.
Let’s explicate this data.
What is Explication?
In 2014, we saw perhaps the greatest all-time collective performance by a class of rookie wide receivers. That year is to wide receivers what 1983 is to quarterbacks and people with the last name “Freedman”: It’s the origin of greatness.
So what we see with this 2014 DFS data very well might not be typical, because this was an outlier year. Still, let’s look at it.
There are three main facts that catch the eye:
1) The rookie wide receivers get better as the season goes along. Maybe that’s something particular to the great rooks of 2014, but it makes sense that first-year players would improve as the season goes along.
2) The platform (DK) starts the season by substantially undervaluing the rookies. That surprises me. I would expect DFS platforms — just like DFS players — to buy the hype on rookie wide receivers and overvalue them at first. Again, maybe these particular rookies overperform and provide DFS value early in the year simply because they are abnormally awesome for rookies.
3) As much as the rookies improve from the first half to the second, the platform’s ability to price them accurately improves even more. These receivers still provide value — for an entire season these rookies provide value and are underpriced likely just because they are rookies — but they don’t provide as much value in the second half. This fact would seem to indicate that, if you want to invest in rookies, especially for tournaments, then you should do so in the first half, when platforms are most inaccurate with their pricing.
What About 2015?
I was really hoping that you wouldn’t ask that, but since you did let’s look at the data.
Per PFR, we find all the 2015 rookie wide receivers to play in a game. And then we screen for the ones projected to score at least one DraftKings point in their games.
And here’s how the 2015 cohort of rookie wideouts does for the season:
LOLzzzzz. YES! I don’t know why that makes me happy . . . but I’m happy.
Here are a couple of points:
1) The platform seems to have bought the rookie hype. All. Season. Long.
2) The 2015 rookies drastically underperform their 2014 predecessors.
3) Even though across the season the platform has priced the rookies down in comparison to the 2014 superstars, the 2015 cohort is still markedly overpriced.
Yes, I felt the need to state the obvious just in case you were reading this while blind or legally brain dead.
Let’s look at Week 1-8 . . . OMG, I am so giddy right now . . .
ROTF.
For eight consecutive weeks, the platform overvalues the 2015 rookie wide receivers. And it’s worse than it seems:
#EpicFail
Look at the average expected points: 7.01. Now look at the Weeks 1-8 average expected points for the 2014 cohort. Yep, 7.01. In case you don’t know what that means; it means that, in the aggregate and through the first eight weeks of the season, DraftKings priced rookie wide receivers in 2015 exactly as it priced them in 2014. Exactly.
It’s almost as if the platform for the first eight weeks of 2015 said, “F*ck it. This price point worked last year for rookies. Let’s just stick with it till we know more and hope that it works out for now.”
Obviously that’s not what happened — not even Gaius Baltar would shirk responsibility in that way (and that reference was for you, Shawn) — but the salary repetition (when coupled with the monstrous differences between the 2014 and 2015 wide receiver classes) highlights a very important fact.
Early in the season, the platform seems to be flying blind on its pricing for rookie wide receivers.
For at least the first eight weeks of the season, rookie wide receivers are the nebula — the place where visibility is minimal . . . where danger lurks and redemption is found — where Cylons attack and Starbuck returns.
Rookie wide receivers are the road to earth.
What Are You Saying?
Here’s the working theory: As Al Zeidenfeld said recently on the “GPP” episode of the Fantasyland podcast, the time to be contrarian is early in the season because people have less information at that time. They have less awareness of that which they don’t know, and that provides DFS players with an enhanced opportunity to steer into the skid of contrarianism and see how far the car can fly once it goes over the cliff.
And, as little as we know about early-season football in general, we know even less about rookies. That’s part of the reason they’re so valuable.
What I’m saying is that DFS sites and DFS players probably have no real idea about how to value or project rookie wide receivers early in the year. That makes them prime targets for tournaments — not right away, because their ownership percentages will likely be too high, but just a little bit later: Once the rookie delirium has died down but before sites are able to price them with anything approaching accuracy and precision.
But I suppose that all of this begs the question . . .
Are You Sure?
Am I sure that sites are better at pricing rookie wide receivers as the season progresses? Fairly.
Here’s what rookie wideouts did in the second half of 2015:
As usual, there are a few items to consider:
1) As we would expect, the cohort does improve from the first half of the season to the second.
2) Yes, the platform — as it does in 2014 — gets better at pricing rookie wide receivers as the year progresses. If a DFS site gets enough data, it will become better with its pricing.
3) As the wideouts improve, their tournament ownership increases. On the one hand, that’s natural. On the other hand, ownership is increasing precisely as the site improves its pricing — and people shouldn’t invest in assets right as the market makers settle on pricing and decrease the possibility of upside.
What this last point means is that, in all likelihood, the time to invest in rookie wide receivers for guaranteed prize pools is not in Games 9-16.
The sweet spot for using rookie wideouts in GPPs is probably Games 3-8 — after the initial spike of hype and before the ownership percentages kicks back up as the rookies noticeably improve.
At least that’s my working (game) theory.
Don’t You Have Only Two Years of Data?
Correct. And next year I’ll have three years of data.
What Are the Takeaways?
So many takeaways . . .
1) If we believe that the 2015 rookie wide receivers are more representative of ‘typical’ rookie wideouts than the 2014 rookies are — and we should believe that — then we should probably also expect the 2016 rookie wideouts to be collectively overpriced earlier in the season. So you can love Sterling Shepard, Tyler Boyd, and Tajae Sharpe as much as you want. Just know that they are more like the 2015 rookie wideouts and less like the 2014 ones.
2) Rookie wide receivers tend to get better as the season goes along. So if there’s a guy you like and he sucks early in the season, that’s OK. He’ll probably get better later in the season. Of course . . .
3) As the season goes along, rookies (just like all other players) are increasingly likely to be salaried at price points commensurate with their probable production.
4) Their ownership in tournaments is likely to increase as they become more productive. If you want to use rookie wide receivers in tournaments, Games 9-16 seem less than ideal spots to do so.
The Conclusion
In retrospect, that introduction looks kind of short.
———
The Labyrinthian: 2016, 86
This is the 86th installment of The Labyrinthian, a series dedicated to exploring random fields of knowledge in order to give you unordinary theoretical, philosophical, strategic, and/or often rambling guidance on daily fantasy sports. Consult the introductory piece to the series for further explanation.
Previous installments of The Labyrinthian can be accessed via my author page. If you have suggestions on material I should know about or even write about in a future Labyrinthian, please contact me via email, [email protected], or Twitter @MattFtheOracle.
Matthew Freedman is the Editor-in-Chief of FantasyLabs.
