36% of Your Draft Will Suck: More Thoughts on ADP

Three weeks ago, I compared 2015 average draft position (ADP) numbers to end-of-season (EOS) rankings. I intended to ascertain which fielding positions are more reliable in terms of fantasy utility relative to their draft position. In other words: I wanted to calculate the probability a player drafted within a particular tier at a particular position would bust.

Granted, the probability that a specific player will bust depends a lot on the player himself: age, recent performance, injury history, and so on. But if you wipe all names from the draft board, looking at only the overall draft slot, the player’s defensive position and his end-of-season rank, we can establish not only which positions are riskiest but also when in the draft they appear to be riskiest.

Like last time, the idea here is to (1) help maximize return on investment (ROI), and (2) minimize risk. (The two go hand in hand, to an extent.) However, I will approach this differently than last time. Instead of deconstructing each position into tiers, I will deconstruct the entire draft into tiers — top 50, top 100, top 200, top 300 — and, within each tier, evaluate the relative to success of each time of player. If you read the first one, this will be similar in terms of results but different in terms of approach.

Also, I will use four years of data (2012-15) in an attempt to improve the reliability of the probabilities to follow. Although I ultimately found composite ADP and EOS data elsewhere, I’d like to thank reader johnnycuff for reaching out and forwarding his data to me.

Bust Rate by Tier

In the table that follows, ADP makes up the left column and EOS rankings populate the adjoining columns. Using the first row as an example, the table should be interpreted accordingly: the top 50 players at the beginning of the season did not finish the season in the top 100 players 38% of the time.

Column “SRB” abbreviates “Same Round or Better + One Round,” and indicates how frequently players actually meet, or beat, their expected value. The calculation is generalized to a player’s draft round, plus another one as an approximation cushion — for example, a player drafted 96th (in the 9th round of a standard mixed league) would have to finish in the top 110, or top 11 rounds, to earn back his draft price.

Bust Rate by Tier, 2012-15

↓ ADP \ EOS →	Top 50	Top 100	Top 200	Top 300	SRB+1
Top 50	56%	38%	19%	11%	37%
Top 100	65%	46%	25%	17%	40%
Top 200	79%	64%	40%	28%	38%
Top 300	84%	71%	50%	36%	39%

SOURCE: FantasyPros, Razzball

SRB = Same Round or Better, + 1 Round

The title of the post gave it away: dating back to 2012, 36% of players drafted in the top 300 have not finished their respective seasons among the top 300. The bust rates have not been localized to the bottom rungs of the ladder: 11% of the top 50, 17% of the top 100 and 28% of the top 200 have qualified as busts outright.

It’s also a coincidence, but: 39% of players — in other words, only four players per round — generated at least enough value to justify their draft prices. And this phenomenon is consistent throughout the draft, hovering between 37% and 40% by tier, historically.

A lot of the (perceived) best players do not make good on their preseason promise. While they’re not total busts — again, only 11% of the top 50 truly bust — they don’t produce the kind of value one would hope to see from such an early draft pick more than half the time. And just from eyeballing the data, a lot of the attrition appears to be injury-related. That’s a hard lottery to win.

Thus, I personally arrive at an impasse. I’m not huge on Andrew McCutchen this year — in fact, I think he’s quite overvalued. Thing is, he’s still a very solid asset. While (I think) he won’t live up to his draft price, the odds of him coming relatively close to it are pretty good in light of an entire career of great health (not to mention incredibly consistent production), considerably raising his floor. The argument could be made, then, that McCutchen’s clean bill of health and high floor is worth paying a premium for.

Also, FYI: that bottom-left corner? That says 84% of players drafted within the top 300 will bust? That doesn’t mean a whole lot — only 50 of those 300 players can fit into the top 50 in the first place, so the natural bust rate must is already set artificially high at 83%. But that 1-percentage point gap is equal to three players — that is, three players drafted outside the top 300 have typically finished the season among the top 50. Another 11 to 12 will wiggle their way into #50 to #100 slots.

Find your late-round lotto pick.

Bust Rate by Position Within Tier

The next table is constructed a little differently. The so-called bust rate of each position is evaluated at each tier. Using the first row as an example, all catchers drafted within the top 200 players do not finish the season among the top 200 roughly 46% of the time.

Bold black percentages indicate frequencies that score at least a standard deviation better than average. Bold red percentages score at least a standard deviation worse than average.

Bust Rate by Position Within Tier, 2012-15

Pos	Top 50	Top 100	Top 200	Top 300
C	67%	56%	46%	45%
1B	48%	36%	33%	29%
2B	64%	43%	38%	33%
3B	59%	41%	44%	33%
SS	81%	60%	46%	33%
OF	52%	44%	38%	38%
SP	50%	44%	40%	37%
RP	60%	62%	39%	39%
Average	56%	46%	40%	36%

SOURCE: FantasyPros, RazzBall

Starting with the big picture: most positions experience relatively the same amount of volatility during a season. Second base, third base and shortstop are all slightly better than average, in terms of earning back value. Outfielders and pitchers are all a little worse than average. It’s first basemen who are especially lucrative, and it’s catchers who suffer the most attrition.

And this makes some sense. First basemen are expected to be offensive stalwarts; catchers, not so much. And because EOS rankings depend upon total offensive production, far more first basemen will show up in the top 300 than will catchers.

Except catchers have recently hit as well as or better than middle infielders. Since 2012, catchers have posted a 91 wRC+ compared to a 90 wRC+ for second basemen and an 86 wRC+ for shortstops. It’s not a production issue, then, as much as it is a playing time issue; backstops see frequent rest, making the ones who see 600 plate appearances in a season — Buster Posey, Salvador Perez, Jonathan Lucroy (all when healthy, of course) — all the more valuable because they rack up counting stats in ways part-time catchers don’t.

Climbing up the draft board a bit, we see that higher-tier shortstops begin to be more volatile than other positions. In fact, shortstop is the most volatile position — a fascinating development given the glut of very young, relatively unquantified talent at the position. Perhaps the injury-prone Hanley Ramirez and Troy Tulowitzki are to blame for shortstops’ struggles in the elite tiers of drafts, but the outlook is somewhat bleak for the likes of Tulowitzki, Carlos Correa, Corey Seager, Francisco Lindor and Xander Bogaerts. Granted, these top prospects may be cut from an entirely different cloth than shortstops’ past, but history simply isn’t on their side.

Implications

A lot of the conclusions from my first iteration of this exercise could be repeated here. Every position experiences a good deal of attrition, so any kind of safety or reliability described above should be interpreted in relative terms. First basemen are solid, and catchers, shortstops and relief pitchers are not. Knowing where each position demonstrates weakness can help you structure your draft accordingly.

If I had to give specific advice for certain positions:

Catcher: Draft Posey, or wait.
First base: They’re relatively safe at all ends of the draft, so you don’t necessarily have to plan around them.
Shortstop: I doubt I’ll convince you to lay off of Correa, Lindor or Seager, but you may capture a better ROI in the very late rounds of the draft (think Brad Miller, Marcus Semien, Trea Turner, Oswaldo Arcia).
Closer/reliever: The turnover rate among RPs drafted in the top 100 is really ugly compared to the draft’s overall turnover rate. Given the inherent volatility of the position — it’s built entirely on small sample sizes — you don’t need to load of up elite, or even decent, closers. Anchor your saves column with Aroldis Chapman or Craig Kimbrel, sure. But you don’t need Trevor Rosenthal and Hector Rondon, too.

Reader Participation Segment / A Thought Experiment

Since 2012, the top 10 players have busted at the following rates:

• 63% did not finish in the top 20 (range: 60% to 70%)
• 38% did not finish in the top 100 (range: 20% to 50%)
• 12% did not finish in the top 300 (range: 0% to 20%)

In other words, about six (of the top 10) will be slightly disappointing, four will be
kind of bad, and one will be really bad. Whether it’s by injury or poor performance, who can say.

Now, here’s the current NFBC ADP top 10:

1. Mike Trout
2. Paul Goldschmidt
3. Bryce Harper
4. Clayton Kershaw
5. Josh Donaldson
6. Carlos Correa
7. Nolan Arenado
8. Manny Machado
9. Giancarlo Stanton
10. Kris Bryant

And this is where you come in. Among them, pick the following:

• Six that will finish outside the top 20
• Among those six: four that will finish outside the top 100
• Among those four: one that will finish outside the top 300

Curious to see where everyone lies.