You are here
Running the Numbers: BPA Doesn’t Mean Highest-Rated
It’s really an amazing time to be a fan of the NFL. Although most strategic changes have occurred at a snail’s pace, we’re entering a time when many organizations are questioning fundamental concepts regarding how to win. The days of dogmatic decision-making, while still present, are being replaced by more objective, scientific thinking. Teams have embraced analytics in order to make far more accurate and efficient choices. We’ve seen this with a higher pass rate and more aggressive fourth-down decisions—both of which will continue to increase as teams realize they’re advantageous.
The same slow-to-change mindset that led so many NFL coaches to hold onto their “you-have-to-run-the-ball-to-win” and “defense-wins-championships” mentalities is also very much alive in draft strategies. Specifically, teams trumpet the importance of taking the highest-rated player on the board at all costs—an approach many refer to as “best player available.” Like running the ball on first-and-10 or kicking an extra point when down eight, I believe the “best player available” draft strategy—as it refers to selecting the highest-rated player on the board—will eventually die out.
Now, I’m certainly in the minority in my thinking; I know of only a handful of people who agree that teams shouldn’t always draft the highest-rated player on their board. So let me explain why I believe the best draft strategy is a whole lot more complex—and flexible—than “best player available.”
Draft Possibilities Exhibit a Range
So often it appears as though draft strategies are split up into a distinct dichotomy; you’re either drafting the best player available, or you’re drafting for need, they say. It’s so engrained into our minds that it almost seems like a given that drafting for need necessitates forgoing the highest-rated player. But it doesn’t.
The best player available/drafting for need dichotomy fails on two levels. First, it assumes that drafting for need is the opposite of drafting the highest-rated player. Logically, we should know this can’t be true since it’s possible to select the best player available who happens to play the top position of need. When that happens—when a team’s highest-rated prospect plays their primary position of need—drafting is quite easy. Ideally, you’d always prefer to draft the highest-rated player and, if possible, you’d want him to play your top position of need.
But the combination rarely occurs. In most cases, the top-rated player will play a position that’s not the most important need. So what then? Most would say you draft that player anyway, but the merits of such an idea become worse and worse as the position becomes less and less of a need.
For example, if the Cowboys have Geno Smith rated in their top five and he falls to the No. 18 pick, does anyone really think they’ll take him? There’s no chance of it, and there shouldn’t be. That’s because quarterback isn’t a need at all for Dallas, meaning Smith would be the “true” opposite of drafting for need: drafting the top-rated player at the position of lowest need. And it’s easy to see why that strategy, although still a version of “best player available,” is just as bad as drafting the top need regardless of his position.
In reality, draft strategies fall into a range. At the one end, we have drafting solely for need. Such an extreme strategy would be very shortsighted; teams would say “we’re drafting this one particular position, no matter who is on the board.” That’s obviously a problem.
But at the other end of the spectrum is drafting the top-rated player at a position you don’t need at all. In most cases, that’s also a big mistake because the prospect—Smith, for example—might not ever see the field.
Pure Need--------Top Player, Top Need--------Top Player, Lowest Need
In the middle, we have the “Platonic ideal” of drafting—the top player at the No. 1 need position. The closer a potential prospect is to falling in the center of the range, the better he’d be as a pick. When a prospect doesn’t fall into the center of the range, teams should really be balancing their board and their needs. To select a player at a position that’s not a need at all, he would need to be rated really highly on the board—way ahead of other prospects. On the other hand, to draft a pure “need” position, the player should be ranked at least near the top of the board. It’s a delicate balancing act, but superior to blindly selecting the top-rated player.
Uncertainty in Rankings
Another reason that a true “best player available” draft strategy doesn’t make much sense is that teams shouldn’t be acting as though their boards are infallible. There’s all kinds of uncertainty involved with the draft process—it’s quite difficult to project prospects’ production years down the road—and that uncertainty should be factored into draft strategy.
When a team is on the clock, there’s a good chance that their top-rated player is higher on their board than at least the majority of the other teams. If he were extremely high on everyone’s boards, well, he probably would have already been drafted. Teams say that they don’t concern themselves with the views or others—and maybe they don’t—but they should.
If Team A has a particular prospect ranked in the top 10 but all 31 other teams have him with a second-round grade, who is more likely to be correct? If Team A acts as though their views on the prospect are without error, they’ll overpay for a player whose true worth is likely much closer to how the 31 other teams grade him than how Team A rates him.
Now, teams obviously don’t know exactly how other teams view each prospect, but it’s still important to understand that each team necessarily thinks the prospects they draft are better than they are. If every team possessed perfect ratings of players, the draft would be a much more efficient practice. It’s still a rather inefficient market, however, meaning teams could benefit from realizing that their boards might not be as accurate as they wish they were. Once that uncertainty it factored into a team’s strategy, the value of “best player available” declines because, well, chances are “your guy” isn’t really the best player available.
Let’s assume that the Cowboys are on the clock with the 18th overall selection and the top player on their board is a safety. We’ll call him Kenny. For the sake of argument, let’s also assume that they really like a bunch of other prospects at the safety position (nearly as much as they like Kenny), many of whom they know they could potentially draft later; we’ll call those players Matt, Shamarko, Phillip, and J.J.
In this hypothetical, it should be easy to see why Kenny isn’t the right pick, even though he’s the top-rated player on the board: he’s not scarce. There are all kinds of “Kenny-like” players available throughout the draft, so it wouldn’t make much sense to draft Kenny in the first round, would it?
Now let’s say the Cowboys’ second-highest player is a guard. We’ll call him Jonathan. Jonathan has a grade just a bit lower than that of Kenny, but the Cowboys don’t have any other guards rated anywhere near Jonathan. The next guard on their board, who we’ll call Larry, has a third-round grade. Jonathan is scarce. Because of that, he’s the right choice for the Cowboys.
Let’s take a look at why this might work out mathematically. The draft is a complex system and there are lots of variables, but for the sake of this example, we’ll use just the safety and guard positions. Here are the Cowboys’ hypothetical ratings for the players they’re considering at No. 18, on a scale to 100:
And here are the ratings for the next-best options at the safety and guard positions:
So what happens if the Cowboys draft Kenny in the first round? Well, they initially gain maximum value. Pretty cool, right? Sure, for now. But in the second round, the ‘Boys presumably wouldn’t want another safety. Even with Matt, Shamarko, Phillip, and J.J. all still on the board, the value of another safety is minimal, or at least nowhere near where it was prior to the selection of Kenny. So the Cowboys reach on Larry, giving them a two-round score of 160. Or, they could bypass the guard position altogether, but they’d still be missing out on the value offered by those second-tier safety prospects.
But what if the team drafted Jonathan in the first round? As the second-rated player on their board, they’d be bypassing the “best player available” draft strategy in favor of scarcity. And it would work. Even if only J.J. would remain in the second round, the Cowboys’ two-round value with a guard in the first and a safety in the second would be 174.
BPA is Shortsighted
So now we see why “best player available” isn’t always the best draft strategy. It’s inflexible, failing to account for team needs, future picks, or position scarcity. When a team says they’re going to draft the highest-rated player no matter what, they’re confining themselves to one specific path. That path maximizes value in the short-term, but it often forgoes it later in the draft. Whereas a draft strategy that incorporates team needs and position scarcity could maximize overall value, “best player available” is too shortsighted of a strategy to do the same. And in many ways, it’s self-defeating; ironically, by drafting the highest-rated player at a position that isn’t scarce, teams could be putting themselves in a position in which they can’t draft the highest-rated player in a later round.
While I advocate the use of a draft strategy that combines a team’s board, needs, and scarcity, it’s really most useful in the early rounds. “Best player available” is a worthwhile strategy in the middle and especially late rounds when teams should be seeking upside at all costs. When the cost of a pick is minimal and the player has a low to moderate chance of making a significant impact, it doesn’t really matter what position he plays; teams can maximize long-term value by selecting the highest-rated player—or at least the prospect with the highest ceiling—in the middle and late rounds of the draft.
Look, there’s no doubt that teams should take the “best player available” in the sense that you always want the guy who will be the best for your team. But that doesn’t mean he needs to be the highest-rated player. Draft strategies fall into a range, with the extremes being drafting purely for need, regardless of the board, and selecting the top-rated player, regardless of the position. Both are problematic because they’re shortsighted, but the “best player available” strategy also fails to account for uncertainty in rankings. In the early rounds, that’s a much bigger issue—one that’s damning to the highest-rated player approach—than teams would lead you to believe. Read