BaseballHQ.com

A 1,000-foot view of something normally viewed up close. Photo by Tim Gouw from Pexels.

Average Draft Position (ADP) values have long been used to help fantasy GMs assess the market value of individual players relative to their projected value. In this study, we want to explore how we can use ADPs at a higher level to assess the market value of entire statistical categories, player roles, and other attributes to help guide our draft strategies.

The initial spark for this study is the near-ubiquitous advice from fantasy experts that fantasy GMs must be aggressive in drafting stolen base sources in 2020. This coming on the heels of years of advice and mega-league pressure to more aggressively draft high-end pitchers early. As I read article after article and listen to podcast after podcast offering the same such advice, I wanted to see what the data tell us about whether or not this is good advice, and whether or not fantasy GMs are heeding it. Let’s dive in!

Methodology

Our first step is to determine how much each draft position is worth, not in terms of auction dollar equivalency, but in terms of total contribution to fantasy standings. And in setting out to do so, we immediately find that things get a little tricky. An average draft position is just that — an average — but the player who happens to occupy a given position at a given moment in time is but one player. Simply taking that one player’s stats and treating him as the sole representative of that draft position won’t fly.

So, in this study, we will look not only at each player’s ADP, but also his earliest draft position (EDP) and latest draft position (LDP), as well as the EDPs and LDPs of those nearest him on the ADP list, to establish some bookends of when fantasy GMs are typically considering drafting that player, and then include his stats in a blended average number for each ADP at which he is being considered. For example, as of this writing Yoan Moncada has an ADP of 66, an EDP of 39, and an LDP of 94; for this analysis, we will consider Yoan Moncada’s stats when looking at draft positions 17 places earlier than Moncada’s ADP and 9 places after Moncada’s ADP, so from positions 49 to 75. (You can see the draft range of Moncada and any other player you want by interacting with the chart below.) As you might imagine, the set of players considered at each draft position widens as you go deeper into the draft, as shown here:

Doing this for all draft positions, we then calculate a blended average of how much each draft position is worth in terms of total statistical value, which sets us up to answer all of our other questions.

The second step is to do something very similar to the first step, but instead of looking at each draft position’s overall average statistical contribution, we filter the inputs by any number and type of variable we want. From this, we can determine how much of the value generated from a given draft position comes from a certain statistic or player attribute.

Finally, the third step is to simply compare Step 2 to Step 1 to produce an index value for each statistic or attribute, which tells us how much MORE or LESS of the value generated from a given draft position comes from that statistic or attribute vs. what we’d expect if that draft position generated a perfectly balanced stat line across all hitting and pitching categories. (Note: “balanced" does not mean the same thing as “average.” An average player can have a very imbalanced statistical profile, just as a very above or below average player can have a balanced statistical profile.) An index value well above 100 can be interpreted to mean that fantasy GMs are biased toward that statistic or player attribute at that draft position; an index value well below 100 can be interpreted to mean that fantasy GMs are biased away from that statistic or player attribute at that draft position.

That’s it. Everything from this point on is straight-up analysis using the above process. We’ll cover a ton of ground and hope that, by the end, you’ll have a better picture of how your competition is likely approaching the 2020 draft, and how you can adapt your approach to outmaneuver them pick after pick.

Findings, Part I: Role/Position

First we'll take a look at what kinds of biases fantasy GMs have with respect to player position and role.

Batters vs. Pitchers Volume

Forgetting about value for a moment, it's no secret that fantasy GMs have a bias toward hitters in the first four rounds, and that's exactly what we see here. This bias toward hitting levels off somewhat in rounds 5-8 and then switches heavily in favor of pitching from rounds 9-15 before regaining balance through Round 23. We'll see later how this manifests itself in the accumulation of value.

Pitchers: Volume by Role

Just looking at pitchers, Rounds 1-4 exclusively belongs to starters. Top relievers start to enter the picture in Rounds 5-8, but then Rounds 9-16 are dominated by relievers. From Rounds 18-22, the balance tilts back heavily toward starters as fantasy GMs round out their staffs. Finally, Round 23 is an unpredictable mess of flyers regardless of role.

Hitters: Volume by Defensive Position

Just looking at hitters, we'll review each position (skipping DH/Utility-only since there aren't enough of them) one at at time. Players who qualify for multiple positions are included at each qualifying position.

Catcher: As you'd expect, not much action in the first three rounds, but then signs of life as the top catchers exit the board, and then a mini-run in Rounds 6-8, and then a long lull until the Battle To Not Get Stuck With The Worst Guy begins in earnest in Round 13.

First Base: A lot of wait and see in Rounds 1-4, then everyone decides to grab the best available 1B in Rounds 5-8. Anyone that misses that wave generally waits until Round 13 to grab a third-tier player, with a final burst around Round 19 to fill the corner spots.

Second Base: After completely ignoring the position the first two rounds and then fighting for the top contributors in Rounds 3-4, second basemen are selected at a relatively steady pace for the rest of the draft, with players taken off the board about when you'd expect based on their expected value. But it looks like a lot of fantasy GMs also see value in speculating that second basemen picked up in the endgame can perform just as well as those grabbed in the middle rounds.

Third Base: Third base may be the most top-half heavy position in fantasy in 2020. With the exception of a Round 3 gap between the top and second tiers, third basemen are consistently taken off the board through Round 12 and then largely ignored the rest of the way. If you don’t have your 3B by Round 13, you’re probably not getting the one you want.

Shortstop: Heavier at the top than third base, but levels off completely after Round 3. Fantasy GMs who don't reach in the early rounds are generally able to pick from a decent set of remaining shortstops at any point in the draft without fear of a big run thwarting their plans.

Outfield: There's the obvious early Round 1 crew, but then outfielders seem to be picked up in small waves throughout the draft. This could make it tricky for fantasy GMs to navigate, as you'd rather be at the front end of a wave than the tail end if there's a specific player you have in mind at a specific slot. On the other hand, if you're more flexible on taking the best available, you can always just wait and tack onto the end of the latest run.

Findings, Part II: Reliability

Next, we'll take a peek at whether/how fantasy GMs are thinking (or not thinking) about reliability when making their selections round after round.

Volume by Reliability: Health

It's always nice to see when an analysis supports your intuition. From top to bottom, the above charts show how fantasy GMs are drafting reliable health from A to F. Sure enough, Rounds 1 and 2 are biased toward players with Grade A, B, or C health. Grade A health remains a priority (again, consciously or unconsciously) through Round 5, when higher health risk profiles start to take over. And true Grade F health risks are not really considered until mid-draft at about Round 14, after which point health seems less of a concern or, if anything, a risk worth taking if the skills are present.

Also interesting to note the uptick in focus on Grade A health players in the end game—could be fantasy GMs wanting to round out high risk rosters with safer bets, or, all else failing, guys who will at least take the field.

Volume by Reliability: Playing Time / Experience

If there's a single factor that drives draft position over anything else, playing time security may be it—at almost no point in the first twelve rounds does Grade A playing time / experience (PT/Exp) dip below an index of 100. And rarely does a player with Grade C, D, or F PT/Exp break into the first eight rounds. In fact, there's a nice little wave pattern of C's in Rounds 9-14, followed by D's in Rounds 14-20, followed by F's in Rounds 17-23. It's all about the playing time. This remains a topic for future research later this spring, but my hypothesis is that player "helium" and "lead" effects (i.e., significant movement up or down in average draft position as Opening Day approaches) is very positively correlated with projected playing time increases/decreases. Monitor those spring situations closely to stay ahead of competitors when trying to snag medium-to-high skill players on crowded real-life depth charts.

Volume by Reliability: Consistency

Unlike Health and PT/Exp reliability, our analysis of Consistency paints a less expected picture. The first thing you may notice in the chart is that the Grade F Consistency cohort dominates the top of Round 1! This seems backwards; yet, it makes perfect sense when you factor in recency bias. Last year's top players are those who over-performed even the wildest expectations—a good thing in theory, but mathematically still a contributor to a lower consistency score. Fantasy GMs are making a big bet that those players are going to repeat, or at least settle back to a higher-than-before floor. But there's really no guarantee, and it really makes you think twice before spending that early first round pick on a guy with F Consistency like Yelich or Bellinger, when you could have more of a lock like Lindor or Arenado.

Interestingly, we also have a wave of Grade D's that follow in Round 2 (along with A's and B's in the back half of Round 1 and Round 2, so the story isn't crystal clear), which again shows how much risk is being drafted early.

There's no clear pattern in the remaining rounds, other than perhaps a return to embracing risk in the last five rounds of the draft, whether it's rookies or high-upside veterans with mixed track records or pitchers transitioning roles.

Findings, Part III: Category Value

Finally, we'll analyze how fantasy GMs are attempting to accumulate value in each 5x5 category throughout the draft.

Batting vs. Pitching Value

In the second chart above, we saw how fantasy GMs are biased in selecting batters vs. pitchers at a given point in the draft compared to what would be expected if they fulfilled their positional requirements evenly as the draft progressed. Here we see the bias in accumulation of statistical value between batting and pitching from pick to pick and round to round. The most notable difference between the two views is that while there is a very clear bias toward the selection of batters over pitchers in Rounds 1-4, there is far less bias in the value accumulated from batters over pitchers in Rounds 1-4. More plainly stated, fantasy GMs are picking far more batters, but the few pitchers being taken are providing just as much (projected) value. It seems, then, that the trend toward more aggressively selecting star pitchers in the early rounds may be a worthwhile strategy, depending on how you feel about the relative risk of batters vs. pitchers.

As the draft progresses beyond Round 4, we start to see more sustained "runs" in value accumulation from batting vs. pitching at different points. These tend to mirror the simple count of batters and pitchers being selected along the way; however, the value indices in this chart are much more exaggerated than the volume indices in the earlier chart, indicating that there may be more "bargains" available for one type of player or the other at different points in the draft.

Specifically, there may be some good opportunities to pick up pitchers in early Round 5, late Round 6 through early Round 8, and definitely in Rounds 9-13 when all of the top relievers are taken off the board. On the batting side, there's a small wave of value getting snapped up around the Round 5-6 turn and the 8-9 turn, and then a looooooooong relative value drought as everyone focuses on accumulating pitching value. But the end game is where the battle to accumulate batting value is won, with a ton of relative value getting drafted in Rounds 17-23. Generally speaking (but more on that in a minute), most of the pitching value is simply gone by the start of Round 16.

Batting Category Value by ADP

Finally, we get to the chart that launched the idea for this study: a look at what biases fantasy GMs have for specific statistical categories at different points in the draft. The indices here represent how much value accumulated at that draft position comes from that particular statistic relative to what you'd expect if all ten 5x5 categories were accumulated evenly. Numbers over 100 suggest that fantasy GMs are going after that statistic at that draft position; numbers under 100 suggest that fantasy GMs are not concerned with that statistic at that draft position.

Right off the top, we see the most talked about bias in action, with an extreme bias toward the accumulation of stolen bases in Rounds 1-4. You might argue that this is simply because there are a handful of players that steal bases in addition to contributing in other ways, and you'd be right. However, everything is relative here. What this is saying is that even accounting for the fact that these guys also hit home runs and have high averages and score lots of runs, fantasy GMs are STILL accumulating more stolen base value than you'd expect in the early stages of the draft relative to all other categories. The other waves of SB value accumulation come in Rounds 9-12 and Rounds 18-20.

After a bias toward high batting average and high run-scoring players in Rounds 1-4, these categories are largely ignored from Rounds 5-13 as owners focus on pitching, steals, and power. From Rounds 14-23, there is a lot of additional accumulation of value from average and runs, mostly at the expense of pitching, but also a bit at the expense of power in the end game.

With the focus on speed in the early rounds, bias toward accumulating home runs and RBI, while still present, is deferred a bit until Rounds 5-8. Like average and runs, power too is ignored in Rounds 9-13 with the intense focus on pitching, but then we see another mini-run of HR/RBI guys targeted in Rounds 18-19; if there's a power hitter you're targeting for the end game, you may want to jump the line and sneak him in at that point. Remember from our first chart above that there is a LOT of variance in ADPs in the later stages—a player with an ADP in Rounds 21-22 range can still be expected to be taken in Rounds 18-19 in a lot of drafts.

Pitching Category Value by ADP

On the pitching side, we witness the same scarcity phenomenon, but with saves instead of steals. Even though they contribute in three categories (saves, ERA, and WHIP), no one ever takes a closer in the early rounds. But value in the saves category starts to come off the board in Rounds 5-8 and then the jockeying for projected saves sources simply dominates the draft from Rounds 9-16 before petering out in the late game—again, at least in terms of projected saves sources.

ERA and WHIP mirror each other almost exactly, which makes sense. First, there's the value earned from the top tier pitchers, which fantasy GMs simply cannot underestimate—if you want value, take at least one top pitcher in Rounds 1-2. Ignore the brief visual gap in ERA/WHIP value in Rounds 3-4, which seems more a function of the extreme values in Rounds 1-2 and those in Rounds 5-8 and especially Rounds 9-16. It's worth noting that WHIP value seems to last a bit longer than ERA value in that mid-game pitching run, so probably best to prioritize ERA by the end of Round 13 and then maybe sneak in a few homer-prone low-WHIP guys who may still be available after that.

Similarly, wins and strikeouts also rise and fall somewhat together, which again makes sense as these are largely the domain of starting pitchers. Still, there are some major differences worth calling out. Strikeouts are definitely a bigger part of the value accumulation in Rounds 1-4 than are wins. Fantasy GMs also seem to favor projected strikeout sources over projected wins sources in the mid-game. But then the strikeout well seems to run dry at around Round 13, whereas wins continue to be accumulated at a steady clip through the rest of the draft as GMs round out their pitching staffs with the league's remaining starting pitchers, especially in Rounds 20-23, at which point a bunch of additional strikeout value comes along for the ride.

Player Over/Under Ranking and Uniqueness by ADP

As fun as it's been to fly in the ADP clouds for this long, we'll end by bringing it back down to the earthly materials we're really trading in at the end of the day, which is individual players. Typically, ADP analyses look at which players are going above value in drafts/auctions (i.e., cost more than they're worth) and which are going below value (i.e., bargains). The above charts shows this—those above the imaginary diagonal line that cuts through the crowd are being over-drafted vs. their value, while those below said line are being under-drafted—but with a new twist. We use our draft position profiling from the previous steps to also show a uniqueness factor for each player with respect to his current draft position. (That last point is important: a player's uniqueness can change entirely if his relative draft position were to change.)

Our uniqueness score is again presented as an index, where 100 represents a typical variance from the package of statistics that you could expect from a player being drafted where the player in question is being drafted. A score well under 100 (more blue) means that the player is really similar to what you'd typically get in that slot, where a score well over 100 (more red) mean that the player is really different than what you'd typically get in that slot.

This is a new concept, so to reiterate, the colors do not mean that a player is necessarily projected to be better or worse than his draft position; they mean that the player has a really similar or really different statistical makeup than what you might typically find at his draft position. For example, Jonathan Villar (ADP 33) is represented by a medium blue dot: the numbers say he's heavily overpriced, but his relative mix of statistics is pretty similar to what you'd find around this point in the draft. Alternately, Cavan Biggio (ADP 136) is represented by a pinkish red dot: his ADP is right around where you'd expect based on his projected value, but his mix of statistics is pretty different than the other players taken around that point in the draft.

You can hover around the chart to see which players are being over/under drafted and which are unique/uniform for their draft position. Use these data to plan your optimal draft and develop contingency strategies for when things don't go as expected. If you don't get "your guy", who should you target instead? How long can you wait until you grab him? How will that shift your team makeup and, given that insight, who should you draft in the meantime?

These are the kinds of questions the "experts" ask themselves all off-season. Ultimately, winning comes down to three things: knowing the player pool, knowing the market, and running into some good luck. Hopefully the above analysis helps the first two of these directly, and gives you more opportunities to benefit from the third.

Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.