ADP, Pt. III: In Which Rounds Do Sleepers Sleep?

I introduce to you today the third part of an ongoing and rather aimless series regarding my research on average draft position, or ADP. Yesterday, I reprised an earlier piece of work in which I look at how often players “busted,” on average, in certain points of drafts. A bust included both low production by (1) poor performance, (2) lack of playing time due to injury, or (3) both.

Before I proceed, I’d like to make a quick acknowledgment. I’ve been asked on multiple occasions how often busts are attributable to injury rather than poor performance. Unfortunately, I would have to comb the data and handle each red flag on a case-by-case basis. While it’s not totally unruly, I don’t really have time to tackle it right now. I was also asked how age might play a part in ADP. I don’t have an answer for that yet, but I do have future research planned that examines the upward bias regarding organizational (or universal) top prospects. Again, it’s something for a future post, but not today.

Lastly: scroll to the bottom for quick results on yesterday’s poll and for another pseudo-interactive segment.

On to the heart of the matter. The title of the post says it all, but it could alternatively be stated as…

When is it OK to start reaching?

As should be expected, players selected earlier in drafts tend to be more reliable than those selected later in drafts. However, there reaches a point where any kind of edge held by higher-ranked players over lower-ranked players starts to wane or disappear.

I would construct a table that depicts, round by round, how players drafted at certain points in a draft outperform their expectations, but a couple of quick sentences might better describe what historical ADP data tells us.

First: one-eighth of players who rank anywhere from 11th- to 90th-best in terms of end-of-season rankings will be drafted anywhere from the 201st to 299th picks. In other words, of the players you expect to provide 2nd- to 9th-round value, one in every eight will come from the 21st through 30th rounds. That’s pretty wild. Like, you know it’s true, intuitively, but you know you would be shamed for ever reaching that early in a draft.

Additionally, players drafted outside the top 300 gain traction starting in the 5th round. On average, about seven players drafted 300th or later will wind up ranking 41st through 80th, equating to roughly two per draft round. Things really pick up in the 9th round, with more than 80 players drafted 300th or later occupying slots 81 through 150 in the end-of-season ranking — good for roughly three per draft round.

It’s inexact. It’s a ballpark figure. But the crux of the matter is players become less reliable as draft progresses. And that makes sense, because if it didn’t, the draft would be deeply flawed, or we, as rational human beings, would be deeply flawed. Which we are. So many that’s not out of the question.

But really, history shows us that unreliability permeates the draft as early as the second round. And when I talk about unreliability, I mean beyond what we already perceive unreliability to be in terms of baseball’s unconquered statistical frontiers and general ebb and flow in player performance. Reliability happens early and often.

I hereby un-bury the lede and declare: it’s OK to reach. In case you were self-conscious about it.

Specifically, beginning in the 9th round.

I averaged the ADPs for each end-of-season ranking as aggregated by round. In simpler terms, I looked at, say, the top 10 players at the end of the season from 2012 through 2015 and I averaged their ADPs as a group to create what could be considered round-by-round ADPs, the results of which are depicted in the table below.

As should be expected, the average ADP for the 10 best players is lower than any other tier. This trend continues for a bit — the next 10 are better than the 10 after them, and so on — until the draft reaches the 9th round. Then all hell breaks loose.

That was a bit hyperbolic, but the point stands: the trend disintegrates around the 81st through 90th picks, as the average ADP fluctuates unpredictably relative to the steady progression of the first eight rounds.

Thus, while it’s perfectly reasonable to start reaching as early as the second round, it should be socially acceptable to reach, and reach far, as early as the 9th round. Given how often players bust — again, whether by injury or bad performance, who’s to say — and how often unheralded players reign supreme, there’s no problem reaching for someone you like. But if you reach, and you blow it — and, at the same time, you inevitably pass up on a solid pick ranked concurrently with that point of the draft — then you risk looking like a dingus. The human ego is too fragile to shoulder such shame.

So that’s it. That’s the takeaway. Don’t be afraid to reach early. But don’t reach too far. There are a whole lot of other elements to that advice — namely, that players far down the draft board will likely be there in a couple of rounds anyway. But if it’s the 10th round, and you like someone in the 13th round, sure, why not?

Results from Yesterday’s Poll

Yesterday, I presented some sobering statistics regarding first-round draft picks and asked readers to predict who they thought would be 2016’s busts, by varying degrees of bustitude. As of very late last night, the polls have informally closed and the results are in. Out of 21 respondents…

Six to fall outside top 20, in order:
Kris Bryant (21), Carlos Correa (19), Manny Machado (18), Nolan Arenado/Giancarlo Stanton (16), Clayton Kershaw (13)

Among them, four to fall outside top 100, in order:
Bryant (19), Correa (18), Arenado/Machado (13)

Among them, one to fall outside top 300:
Machado (5)

While it’s a small sample size, so to speak, it’s an interesting exercise in ceilings and floors. For example, Correa garnered the second most votes to finish outside the top 100 but received less than half as many votes as Machado to fall outside the top 300, suggesting a lower ceiling but a higher floor. Or, instead, that he’s less volatile than Machado — which, I admit, is kind of funny to me. Meanwhile, Bryant, unanimously voted to fall outside the top 20 (given the parameters of the poll), did not receive the most votes to fall outside the top 300. Perception is so interesting.

At the other end of the spectrum…

Fewest votes to fall outside top 20, in order:
Paul Goldschmidt (2), Mike Trout (3), Bryce Harper (6)

Fewest votes to fall outside top 100, in order:
Trout (0), Goldschmidt (1), Harper/Kershaw/Josh Donaldson (3)

Fewest votes to fall outside top 300:
Trout/Goldschmidt/Donaldson (0)

Oh, yeah, my response to the poll, with almost zero explanation

Except for this disclaimer: these are listed in no specific order.

20+: Harper, Correa, Arenado, Machado, Stanton, Bryant
100+: Correa, Arenado, Stanton, Bryant
300+: Stanton

Because I can’t help but bank on the annual Stantonian freak injury, and can’t bear to think of Kershaw suffering an elbow-related one himself.

And now, as promised, another poll

This is less structured than yesterday’s question. Just do me this:

• Per NFBC ADP, pick a player currently being drafted between picks 200 through 299 to rank among the top 100 players at the end of the season.
• Pick a player with an ADP of 300 or higher to rank among the year-end top-150.

I know, not quite as exciting as yesterday. Still, the link provided in the first bullet point offers an interesting array of talent. I won’t name names to prevent coloring anyone’s perspectives, but man. Seems like 2016 is deeper than the typical draft.

Epilogue

I realized I never really answered the question stated in the title of the post. I kind of attacked it in a roundabout way. Frankly, sleepers exist throughout the late rounds of the draft. There’s no formula that tells you when sleepers are most likely to emerge in a draft — or, at least, I don’t have a formula for you. That’s why I think when to start reaching might be the better tidbit worth knowing.