Reassessing Best Player Available Draft Strategy: Why Teams Should Often Bypass BPA To Maximize Overall Value
“Reassessing Best Player Available Draft Strategy: Why Teams Should Often Bypass BPA To Maximize Overall Value”
Sounds more like the title of a thesis than a blog post. Nonetheless, I wanted to again delve into what is one of my favorite football topics: draft strategy. Last year, I published an article called “Why Selecting Best Player Available in NFL Draft a Myth.”
The post highlighted a few of my unconventional (and wildly unpopular) thoughts on draft and game theory, the most intriguing of which is that selecting the best player available, even at a position of need, is often a mistake. I will recap that article with a few quotes, but I suggest clicking the link above and rereading it if you are bored at work, or entertained but simply don’t have a lot of work to do, or even if you have a ton of work, or if it is nighttime and you are off of work, or if your name is Betty. Just read it. Here we go. . .
Like many (or even most) of the long-held NFL “truisms,” the concept of selecting the BPA (best player available) is mistaken. As is the case with punting on 4th and 1 or always kicking extra points, selecting the BPA will actually lead to sub-par results.
GMs who say they always take the BPA are simply lying.
The key to this strategy is a concept I’ve discussed in a few of my fantasy football articles–VORP (value over replacement player). In a nutshell, VORP means selecting not the player with the most projected points, but the player with the largest disparity of projected points compared to the next player at the same position who you could secure in a later round. If you read between the lines, you can see game theory is really the backbone of this strategy. To effectively maximize value, it is critical to understand perceived worth.
Game theory is all about understanding opponents’ beliefs and using them to your advantage. If you had perfect knowledge of other teams’ draft boards and knew the top player on your board was rated three rounds lower on everyone else’s, you would wait to select that player. BPA, even at a position of need, promotes the dismissal of potentially useful information.
Although no team has perfect knowledge of a player’s perceived value, the notion that perceptions can and should alter draft theory remains unchanged. BPA as a draft strategy is too shortsighted and could force premature selections, ultimately decreasing overall value.
On a real world example from 2011:
Let’s assume the ‘Boys will select a defensive end and an offensive tackle in the first two rounds, but they’re unsure of the order. Now, let’s provide a numerical value to the possible targets. As a guide, we will use the NFL draft trade value chart and my own 2011 NFL Draft Big Board to assign these values.
As I’ve already proposed, let’s assume Dareus (ranked No. 2 on my Big Board) is available for the Cowboys. At that ranking, he’s worth a whopping 2,600 points. As I’ve argued in the past, however, I think there is a major problem with selecting a defensive end in the first round. By the time the Cowboys’ 40th selection rolls around, there is zero chance that a top-tier offensive tackle will be left on the board. My top five tackles–Tyron Smith, Ben Ijalana, Anthony Castonzo, Gabe Carimi, and Derek Sherrod–will almost certainly be gone by the second-round.
Thus, the top offensive tackle that is left to pair with Dareus, according to my personal Big Board, is Alabama’s James Carpenter. . .all the way down at No. 71 overall. According to the value chart, that selection is worth 235 points, bringing the Dareus/Carpenter duo to 2,835 combined points. Certainly our VORP has been compromised, as Carpenter is terrible value in the second-round. But is Dareus’ BPA status enough to compensate?
To determine this, let’s project the Cowboys’ possible selections if they take an offensive tackle in the first-round. At No. 9, the ‘Boys may very well have their pick of the litter, and according to my board, Tyron Smith (No. 8 overall) is that guy. The eighth overall selection is worth only 1,400 points–a far cry from the 2,600 that we assigned to Dareus.
We can already see the Dareus/Carpenter duo is going to win out. Even if the Cowboys somehow land Cal’s Cameron Jordan in the second-round (which is clearly a pipe dream), his 14th overall ranking–worth 1,100 points–would still bring the Smith/Jordan duo to only 2,500 overall points–335 behind Dareus and Carpenter.
Although the Cowboys selected an inside linebacker in the second round, they still landed Smith in the first. My hunch is that he was their BPA, but he should have been selected even if they had a prospect rated higher than him.
It’s worth noting that, although the optimal tandem turned out to be that which was comprised of the BPA, the process by which we discovered that was still VORP. Thus, teams will often arrive at the right selection, but implement the wrong method of getting there. Selecting the correct player helps you now, but selecting the correct player by utilizing the proper draft strategy will help you in the future.
Mathematics often leads to counterintuitive results, but the teams which disregard their “gut” and utilize the numbers on draft day are generally the most successful. Remember, the “gut feelings” are already implemented into a team’s rankings (whether they admit to it or not). Draft day is not the time to follow hunches.
What I mean here is that film study, interview results, and other non-measurables are already reflected on your board. Ironically, if you disregard your board on draft day (in relation to VORP, not BPA), you will actually be forgoing those gut feelings which were already implemented into your rankings.
VORP is an all-encompassing draft strategy that leads to greater ultimate value than BPA–a more short-sighted draft philosophy which disregards the future in favor of optimal value right now. Would you rather have $100 today (BPA) or $500 tomorrow (VORP)?
After I posted that article, I found more support in my critique of BPA draft strategy from Code and Football in their article on why drafting the BPA is simply a way to optimize buyer’s remorse:
Consider this scenario: you have three players in the middle rounds you are considering, whose “true career value” is about the same. We’ll assume drafting is an efficient market for now, so our estimation of the value of these picks is a normally distributed estimate whose mean is based off their true career value. Which one of these men do we draft? We draft the player whose value we have overestimated the most. Consequently, we draft the player most likely to underachieve our expectations.
Since in most drafts there are very few times a true BPA falls into the lap of teams (i.e. players where one is wildly superior to all other candidates), it would seem that BPA is a way of optimizing how heartbroken a team will be over the draft choices it actually picks. Though this approach would appear to gather the best athletes, in a draft with a large error, and multiple situations where you’re picking from nearly equivalent athletes, perhaps all BPA will get you is maximally suffering from buyer’s remorse.
This idea has differences from my own, but it points out the fact that the value of taking the BPA is often minimized because of a team’s overestimation of a prospect’s ability. The argument that BPA can help a team secure the most elite prospects, then, seems less compelling. With 32 teams all acquiring nearly the same information, the chance that a single team will obtain a player whose value is so much greater than the other options that it overrides the value of VORP is slim at best.
In the two main articles I posted on draft strategy (here and here), we had over 50 really insightful comments. Many of these raised interesting critiques of VORP, game theory, and other draft strategies, and I wanted to address a few of them now. . .
“VORP doesn’t address the ‘real world’ value of specific positions.”
It does. Much as the measure of “intangible” things like heart and determination are actually reflected in advanced football statistics, the value of a specific position over another (quarterback to linebacker, for example) is reflected in a team’s draft board.
“VORP as a long-term strategy will lead to less overal talent than BPA.”
VORP leads to the greatest overall value because it has a far greater focus than BPA, or BPA at a position of need (BPAAAPON, if you will). VORP is in the business of temporarily bypassing short-term value to secure greater value in the future.
The most valid critique of the draft theory is that it requires too much knowledge, i.e. you can never have enough knowledge of other teams’ intentions for the draft theory to work in practice. It is a pipe dream, some might say.
While this might be possible, I think that argument would be a better one when applying VORP to the later rounds of a draft. For the most part, we all know which players will go in or around the first couple of rounds, so predicting the abundance of talent at a specific position is made easier early in the draft. That task only becomes very difficult after the first few rounds when boards do not match up as comparably.
In effect, VORP becomes a less valuable draft strategy as the draft rolls along and opponents’ beliefs become less predictable. In the later rounds, when the goal is to maximize upside (as opposed to the goal of minimizing downside early in the draft), selecting the BPA has more merit.
“The draft is a crap shoot, so VORP is no more valuable than any other strategy. Just select the best players.”
The draft is a crap shoot in which a large majority of success stems from luck, but at the same time that doesn’t negate the value of specific draft strategies. The idea is comparable to blackjack, where the outcome of any single hand is determined almost solely by “luck.” A great blackjack player might win perhaps one out of 100 hands more than an average player.
In a sport like football where the competition is so stiff, though, very small advantages equate to big success. It is the difference between a DeMarcus Ware and a Bobby Carpenter, for example. One draft selection can lead to a monumental difference in production.
“VORP is a baseball term and should not be applied to NFL draft strategy.”
I simply use VORP (value over replacement player) as a label. Change it to ‘position scarcity’ if you would like. The idea is the same: short-term bypassing of the BPA can lead to overall greater value, particularly in the draft’s early rounds.
“Too many things can happen in the future which make using VORP now usless.”
The same thing can be said for any draft strategy. We can only work with the information we have at hand. What if Tony Romo tears his ACL in camp? There are a lot of things in the future that could alter the efficiency of past decisions, but the best way to maximize the opportunity for future success, and the only way, is to use all present information.
“VORP assumes that each draft is an independent variable and the potential of future drafts has no bearing on the draft this year, which simply isn’t true”
Again, all present information can and should be factored into rankings. I’d say the ability for any organization to effectively evaluate the positional talent for a draft which is over a year away, however, is small enough that it should have little to no bearing on current decisions. In addition to not knowing draft slots, teams don’t have a grasp on which prospects will enter the draft, which ones might get hurt, how they will perform in their final collegiate seasons, and so on. Bypassing a talented player at a position of need which VORP draft strategy suggests to select because the subsequent draft might be full of talent at that position seems like a poor philosophy regardless of the draft technique used.
“The draft is an inefficient market, devaluing the use of VORP.”
VORP’s value would certainly be decreased if the draft was an inefficient market, but for the most part it is very efficient. Yes, teams hit and miss all the time, but that doesn’t make the market remarkably less efficient. When we take a large sample size of drafts into account and analyze the success of players based on draft slot, we see a very specific downward curve. Whether we compare draft spot to years as a starter, Pro Bowl selections, or career approximate value, higher picks tend to perform superior to lower ones: http://www.advancednflstats.com/2009/04/career-success-by-draft-order.html
In short, although there are random deviations (which are to be expected), teams are generally efficient in the draft market.