The DC Times

A New Way to Look at the Cowboys, NFL, and Fantasy Football

By Jonathan Bales

More Evidence ‘Best Player Available’ Is a Poor Draft Strategy

Subscribe to The DC Times
Never miss a post again!

Jonathan Bales

I created a minor ruckus prior to the 2011 Draft when I argued that teams should not necessarily always select the best player available, even at a position of need.  The reason I gave for this was a term called VORP (value over replacement player)–a concept based on position scarcity.  In the past, I defined VORP as follows:

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.

In the article I wrote on BPA, I argued:

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)?

Well, my friends at Code and Football (great site, by the way), posted a short article on why drafting the BPA is simply a way to optimize buyer’s remorse.  In addition to labeling BPA a “hoary old theory” (which I love), C and F writes:

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 remourse.

This idea is obviously different from my own, but it is interesting to note that others also view BPA as having little value, outside of optimal suffering.  The Cowboys clearly drafted for value in 2011.  The question is whether they enhanced their long-term value by including position scarcity into their draft motives, or if they opted for the top-rated player at each individual draft slot and ended up with sub-optimal overall value.

Like this post? Share it with others:
  • Facebook
  • MySpace
  • Reddit
  • RSS
  • Twitter
  • Yahoo! Buzz
  • del.icio.us
  • Digg
  • email
  • Google Bookmarks
  • Netvibes
  • StumbleUpon
  • Technorati

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

22 Responses to More Evidence ‘Best Player Available’ Is a Poor Draft Strategy

  1. Tyrone Jenkins says:

    VORP – let’s discuss that…

    Hypothecial: let’s say Dallas has the 40th overall pick. There would be several potential draftees on the board, obviously, but Dallas has narrowed their selection down to 4 guys. The first is a guy that plays NT and the numerical value Dallas estimates his potential (ceiling) with is 90. Guy #2 is a ILB with a potential of 89. Guy #3 is a OG w/ a potential of 86. Finally, guy # 4 is a RB w/ a potential of 85.

    Now Jonathon, according to your VORP theory, the greater difference in value that Dallas has for the NEXT guy at each position is THE guy that should be the pick. Given the above scenario, if the next best NT’s value is 88, next best ILB is 87, next best OG is 78 and next best RB if 80, then the correct selection at pick # 40 for Dallas would be the OG w/ the potential of 86 since the difference between him and the next best available OG is 9 points – correct?

    If I’m interpreting VORP correctly, then let me mention your argument has merit but I disagree with it. Here’s why?

    VORP assumes that each draft is an independent variable and the potential within future drafts have no bearing on the draft this year – which simply isn’t true. Every year, the draft has different categories of players who tend to be “stronger” than others. This year, it was defensive linemen. Next year, it might be offensive guards. If Dallas took the VORP strategy and selected the OG valued at 86, next year there might be an OG w/ 94 potential available at pick 40. In that case, Dallas lost out on 8 points for the same selection from year to year. Even if this year’s OG selection lives up to his value of 86, he’ll still be LESS value than the OG selected the very next year.

    In my opinion, you select the HIGHEST graded guy at your position of need whenever your selection occurs – ie. they should select the NT. Immediately after that selection, they should cross all other NTs off their board and dont’ draft another one as that is no longer a position of need. Then, they should start looking into drafting other needs.

    In the above scenario, even if the next year’s draft has a NT w/ a higher potential available at 40, Dallas is still filling their team with guys w/ greater potential. VORP strategy might lead to guys w/ lesser overall potential simply because they happened to be greater than the next best alternative for THAT year.

    It is ALWAYS better strategy to gain as many guys as you can who are as close to ELITE as possible even if it means that you get one ELITE guy and 2 average guys. If you can do that, then you’ll more ELITE guys on your team (which can be traded for other solid guys w/ a proven track record if needed). That would be simply a better strategy than gaining 3 solid guys. All that will do is lead a team to having a lots solid talent but is absent of playmaking ELITE talent. Can you imagine an O line w/ 5 Kyle Kosiers, a D line w/ 3 Olshanskys, 4 LBs all comporable to Bradie James and 3 secondary members all about as good as Scandrick?

    Just my .02.

  2. David Myers says:

    Tyrone,

    The problem with BPA is that drafting isn’t a science, it’s a hit or miss art. Therefore, the bigger bang for buck comes from better scouting, not drafting theory. You can draft BPA till the cows come home and if your scouts suck, you will fail. If you have terrific scouts, you could throw darts at photos of your best candidates and do well.

    Jonathan,

    Thanks for the H/T. DVOA seems like an interesting approach to the idea of the draft as a marketplace. If I were to restate the notion, I’d say “The draft is a market. Don’t overpay.”

    David.

  3. Tyrone..outstanding argument. I would say you can make your idea even stronger by adding in free agency. Instead of waiting for the higher value in next year’s draft, teams can fill in positions they were unable to upgrade in the draft via free agency. In that way, selecting the BPA at a position of need seems sound.

    My argument will still be the same, however. When you extend the BPA theory to multiple drafts, you are giving it a larger scope than my VORP/DVOA theory and creating better long-term value (possibly…what if three drafts in a row are poor on a particular position of need? Further, most teams don’t have time to wait. Coaches need to win now, so even if BPA was theoretically optimal, it isn’t practical over the long-haul..outside of perhaps NE, PIT, etc where coaches have true job security). Anyway, I still think VORP/DVOA would outshine BPA if also extended to multiple drafts. You might argue that my theory necessitates future knowledge of prospects that isn’t possible, while BPA is a more immediate method, but teams certainly have an idea of who is available in upcoming drafts. I promise the Cowboys, and all 31 other teams, had an eye on 2012 even before the 2011 Draft.

    Another argument I think you’re making is that elite players are exponentially more valuable than average ones. For example, I think you’d argue a 95 and 85 are better than two 90s. I would agree, but I think that can be factored into any draft theory. I don’t think you can discredit VORP because if its inability to secure top-end talent because, if a particular prospect is rated high enough, he will get selected. It all comes down to a particular team’s formulas. Perhaps Team A prefers a 95 and a 85 player to two 92s, while Team B prefers the latter. Even though the second scenario creates higher overall value, the choice comes down to each team’s views on the potential impact of an elite player and the difference he can make over a really good one (which should probably be factored into numerical ratings anyway).

  4. David..no problem. Love what you’re doing there. I think the view of the draft as a marketplace is a good analogy. My main point here is that teams’ views on their upcoming picks can and should influence their current selection, whereas a BPA theory necessarily eliminates one’s ability to perform that task.

  5. Vince_Grey says:

    David hit the nail on the head. Both ways can be right… or wrong, depending on the circumstances at hand.

    Regarding Tyrone’s argument, you simply cannot draft one year with an eye on next year’s draft, which may, or may not, have an abundance of players in any certain position. Also, sh*t happens. That Division II free agent guard you signed as camp fodder might blow you away and be a future pro bowler. Or, that 1st round “can’t miss” screws his back up and never plays again.

    And, no matter what, there’s some luck involved. As usual, I like to look at history for future predictors. Back in 1990, Jimmy went into the draft with a determination to select a pass rusher, but the guys he wanted were all picked off ahead of his slot. He tried trading up, and he tried trading down, but the deals weren’t there at the time. His 4th or 5th option was to “settle” on Emmitt Smith. I’m sure he left that draft thinking it wasn’t all that great because he failed to get his pass rusher.

    Well, of course, Smith became the greatest post season runner ever (IMO, and not too shabby in the regular season either.) and the 49ers were kind enough to basically give us Haley. Three SB’s. Thank you very much.

    When your pick comes up, you just have to decide which is better: that available guard you had rated 2 rounds higher (but don’t really need) or that RB you really need and is right where you thought he’d be, and kind of like, but don’t really love.

    Suddenly, all kinds of questions arise. If you take the guard, is there another RB that will likely be there later that you like almost as much? Can you pick up a decent free agent RB for a year?

    Bottom line, no matter what you do, you have just as much chance of being right as you do of being wrong.

  6. Tyrone Jenkins says:

    Excellent points all.

    David, I think I misled you. I’m not an advocate of BPA. I’m an advocate of BPA at a bona fide position of need. There is a fairly significant difference.

    Jonathan, as far adding free agents to fill in void of the draft, this is the 1st year where that is possible. Normally, free agency occurs BEFORE the draft, and the draft is used to fill in holes left from free agency. That’s why BPA at a position of need is the strategy I advocate. Keeping an eye on next years draft always helps but is, as expected, even more unpredicatable. Jake Locker and Taylor Mays were both consensus top 5 picks had they come out a year early, waiting a year resulted in them dropping. It’ll be interesting to see how high Andrew Luck goes.

    Vince, agree that there are unforseeables in every draft: the 7th rounder who emerges as elite (Jay Ratliff) and the 1st rounder who’s a bust (Bobby Carpenter). But, historically, there have been more elite players from the 1st round than any other round. Generally speaking, you get what you pay for.

    Again, I love the discussion – one of the best I’ve been a part of so far I’d say. Certainly is helping pass the time during this lockout…

  7. moses says:

    The whole idea of the draft is to add to the pool of existing players.
    There are actually several pools to consider:
    1. Current roster & practice squad
    2, Current draft pool
    3. Free agency
    4. Trades
    5. Future draft pool

    I think that this actually frames the options that a team has when considering a player in the draft.

    Can I sign someone that is a high caliber player without having to pay 1st round money or use that draft choice to get another high caliber player and get 2 top players instead of just one?

    Another thing to consider is the scheme that they will play in when they come in. The Patriots pick some off the wall players that have you scratching your head but they fit what they want to do as a team.

  8. john coleman says:

    To VORP or BPA?

    You guys can figure where it fits.

    Simply put, you must draft according to need and value. Let me also say that I like David’s marketplace thoughts and IMO he is right. Why buy Lima beans if you are growing them? So we look at our roster and who fits our plan. Then we develop our board of prospects, who will add value to our roster. Now we have a plan of who we have and who we need. We then develop another board that encompasses all available players who could potentially be of interest. I’m talking whether of need or not.

    We execute the plan by 1st looking for the guy who can help us fill a need while adding longterm value as a roster maker or trade bait. Keep in mind that the tradebait scenario is if you have a SOLID roster already. IMO we are in a fix now because of guys like Spears, Hatcher, Bowen, R. Williams, and Barber who offer no market value, either because of price or production. Solid olinemen, QBs, OLB’s, NT’s, CB’s are always a need and if they can play at all will have some value. RBs(because of the pass happy league) and DE’s(because of the 3-4) are a dime a dozen.

    Looking at our draft-
    #1-T.Smith- Definitely qualifies probably on both the BPA and VORP arguments. He was definitely a need filler. IMO he was also the BPA. In fact he would have been my BPA two or three slots earlier. He is only 20 years old and if he pans out will be a player for 10-14 years. He also will command value to any team if he can play LT. I admit, I was surprised that JJ actually took him at #9, but it is what I would have done. Looking at the FA’s available and the price they command, good shopping. Factoring in the bust factor of others available, good shopping.
    #2-B. Carter-Based on need and longterm value, I have no problem with this pick. This guy could end up being the best ILB in the draft. He also can be an everydown player. Great size and speed, good playmaker. I also feel we had inside knowledge because of coach Davis. Looking at the FA marketplace, good shopping. Good ILB’s are priorities to retain. Looking at our roster he is an instant upgrade when healthy. We had James, Brooking, Lee and who? Brooking should be able to contribute some, until Carter is ready.
    #3-DeMarco Murray-Evidently we have a plan. Either Barber or Choice are gone. I suspect Barber and his lack of explosiveness/huge salary is the guy. This pick also adds a little homerun depth behind Jones. Again he adds roster value if he does as expected. For the record, I am lukewarm with this pick, because of other players available. Rackley, Ellis, Jernigan, Hankerson, and Wilson were all there. I think the team sees a little Reggie Bush in Murray. I look for him to be split out on occasion.

    The team later addressed OG-Arkin(mean streak and has played OT) so missing Rackley could be a nonfactor. Same for Jernigan when we selected Harris. Between Murray and Harris you cover the slot and PR/KR. Later you select Thomas for competition and value, if he makes the team. Same with Chapas and Nagy.

    After this ramble, go figure. Point is that IMO drafting is a constantly variable science and factors that only the insiders know influence what is done. I am for taking the BVA(best value available). Players who can make our team initially and offer trade value if they don’t fit longterm. Hence, a Murray, could help us initially KR/PR/3rd down back and be a trade value in two years, if he fails to be a Bushish type player. Just get players that we don’t have to give away. Let Crayton be the example. He helped us, but was a role player. He netted a 7th rd pick. Not bad if you consider what we payed initially. Then consider Barber. Overpriced, under performing and no trade value. Sometimes you miss after the draft. How about Roy?

  9. john coleman says:

    Let me mention the Patriots way as well. Draft the future guys and trade the veteran while he has value. Go out in FA a fill any holes with the BPA for your system and budget. Then Barbers and Williams’s don’t happen. You must always get something in return.

    Hopefully we have learned our lesson.

  10. Good points by everyone.

    @Vince- I agree that the draft is a hit-or-miss process and there certainly a ton of luck involved, but the best teams optimize their chances for success. Even if the disparity between Team A and Team B hitting on a draft pick is just five percent, those numbers become rather significant when extrapolated over, say, three years of picks. Kind of like poker…you or me could beat even a poker pro over the short term, but in the long-run they will always win.

    @Tyrone- You’re right about free agency, and I was referring to this season. I definitely think BPA at a position of need is superior to true BPA, and I don’t think any team really ever takes the true BPA all the time, although the Cowboys came close this draft.

    @Moses- Your point about the external factors that go into each draft pick is a good one. I think an overarching theme here that we all agree upon is that true BPA is too short-sighted of a draft strategy.

    @JC–I like the lima bean comment. Nice analysis of the picks, and I’m just happy that T Smith was probably a true BPA and fit a position of need. It would have been interesting to see what Dallas would have done had someone like AJ Green dropped to them.

  11. craig kocay says:

    @last post
    we would have drafted AJ green

  12. Tyrone Jenkins says:

    I disagree – I think Jason Garrett would’ve convinced JJ and the rest to do the right thing and draft Smith.

  13. craig kocay says:

    ha that was a joke
    or i hope at least they would not have

  14. valmont says:

    I may be wrong, but let’s consider the standard explanation of VORP

    From Mathletics (Wayne Winston) ….

    “The key tool involved in our analysis will be the Value of a Replacement Player Points (VORPP), which was developed by Keith Woolner, formerly of the Baseball Prospectus and now an executive for the Cleveland Indians…

    We know that players create value by doing good things. Woolner appears to be the first person to have realized that players also create value *by keeping bad players out of the lineup*. Woolner asked what a team would do if a player were injured. The team often brings up from the minor leagues a player whose salary is very low (let’s say 0). In theory there is a near inexhaustible supply of such players (called replacement players). To define a replacement player for, say, second basemen, Woolner would put the second basemen in descending order of plate appearances and define the replacement second basemen to be those who rank in the bottom 20% of this list.”

    VORP to me is about correctly recognizing that a pitcher than eats up 200 innings at a 4.50 ERA can be more valuable than a pitcher that can only give you 100 innings with a 2.50 ERA because the former is keeping an inferior player out of lineup. The football analogy would be a Matt Forte or Brandon Marshall.

  15. valmont says:

    the C&F guys are conflating 2 issues that are seperate: BPA and drafting error.

    You agree that drafting error would lead to a ‘optimizing buyer’s remorse’ whether you picked for need or you picked BPA, right?

    If a team misrates players, whether they draft BPA or for need they will still be drafting their most misrated player, right?

    The fix would be neither need or BPA but to throw out your draft board, right?

  16. Tyrone Jenkins says:

    I’m assuming you mean that the pitcher who’s only good for 100 innings (and then gets hurt or is by some other reason not available to pitch) isn’t as good because the replacement player called up from AAA has a value far less than that of the pitcher w/ the 4.5 ERA who would be available twice as long?

    We’ve debated this point before…adjusting that to football, that’d be the equivalent of having the WR on the team who’s good for 50 catches, 600 yards and 5 touchdowns but is healthy all year > the WR who’s good for 85 catches, 1150 yards and 11 TDs but who is only available for 1/2 the games. Your point is that the reason why the former is preferable to the latter is because whenever the latter isn’t playing, his replacement is so much worse that the former’s a better value over the entire season.

    I disagree somewhat. I think your point is valid only when it’s applied to certain positions and on certain teams. If Peyton Manning was less durable, then your point would be made in that the backup QB for the Colts is THAT much worse than him that it’d be better to have a more durable QB w/ less ability.

    But, I think that situation rarely happens – especially in football. Romo was injured for 1/2 the year and his replacement did just about as good than he did. Dez Bryant went down but another WR stepped up and caught passes – not as frequently as Dez did but the replacement did have some value. The assumption that replacement players are worthless isn’t valid – they do have SOME value. Some more than others. And, it’s unrealistic to know exactly how much time high value players will be available. For your baseball scenario, it’s impossible for the coach or GM to know that your 100 pitcher will only produce 100 innings that year (insted of 200) before the year starts.

    IMO, it is FAR better to have a 95 rated player and an 85 rated player (even if the 95 rated player only plays 1/2 the time) than 2 players both rated at 90. If the 95 player ends up averaging only 50% (or less) over 3 years, then I cut him and go get a more durable player rated as high as possible at his position. But I don’t “settle” for one w/ less ability because he’s able to play more….

  17. valmont says:

    Tyronne,

    I’m not arguing for VORP.

    And I could also be wrong.

    But my point is to explain what the VORP insight really is. The insight of VORP is that there is value in keeping a bad player off the field. In baseball that tends to mean giving ‘extra credit’ to more durable players.

    I tend to agree that this is less applicable to football.

  18. Tyrone Jenkins says:

    c – nice explanation

  19. Vince_Grey says:

    JB – You’re right, but I say it’s that extra “spurt” of luck that very often separates the champs from the contenders. For all their drafting prowess and Belichick’s coaching, if they don’t luck into a HoF QB in the 6th round (Brady) they’re not SB anything. But, IMO, the Pats drafting greatness is not so much their ability to pick good players as it is their ability to consistently acquire extra premium round picks.

    That’s also the key to why Gil Brant and the Cowboys seemed to suddenly lose their drafting mojo after the `77 season. They stopped acquiring all those extra 1st, 2nd, and 3rd round picks.

  20. David Myers says:

    Valmont,

    I’m hardly conflating draft error and BPA. I’m saying one is so large that the other doesn’t buy you very much. I’m saying the proportional net value you gain by improving draft error would more than offset any theoretical gain you would have by worrying overly much about the net value differences between say, drafting for need or BPA.

    I calculate that error level here:

    http://wp.me/p1m41i-aS

    Rick Reilly also recently did a “redraft” calculation, and his errors, by his method, far exceed mine.

    David.

  21. valmont says:

    I could be wrong so let’s examine the facts.

    Can we agree that you will have drafting error whether you draft for need or BPA? For example, if you need an OT you will draft the highest rated OT on your board, right? So if you draft BPA you draft the most overrated player on your board, if you draft an OT you draft the most overrated OT on your board, right?

    Your argument is then is that by drafting BPA you’re maximizing drafting error. I guess the logic is that by drafting for need two things happen. One, need is random so the drafting error related to your need position may happen to be smaller. Two, you’re getting benefit from filling need.

    Here’s the real key assumption you’ve made. You have assumed that the NFL draft is an efficient market. If you believe that then there are a whole bunch of implications. Why have a scouting department at all? Why have a board? Just take an average of the consensus boards and follow that.

    Let’s also examine your analogy with Burke. Let’s think about Berkshire Hathaway making investment decisions. On one hand, Warren Buffett will not participate in auctions because of the ‘winner’s curse’. Again, I could be wrong, but aren’t you then extrapolating that logic and saying that Buffett shouldn’t make any investments because it will only maximize his ‘investment error’? He’ll simply be buying the companies he’s most overvalued?

    I think you can see that how that rests on the assumption that the market is indeed efficient. If you allow that there can be market inefficiencies, then it makes perfect sense to both buy undervalued companies and draft BPA.

  22. valmont says:

    David,

    I should add the following. I agree that your logic is correct if your assumption that the NFL draft is efficient is true.

    I just disagree that the NFL draft is efficient.

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>