Editor's Note: This article was originally published in July 2010. We are bringing it back in archive form for your convenience.
Idly strolling down the main boulevard of Petropavlovsk-Kamchatsky, a tourist is stopped by something he sees in the outermost reaches of his peripheral vision. It has been a long week on the Kamchatka Peninsula. He has gone paragliding and walked around the central port, gawking at snow-capped volcanoes that lean in around the city like leering aunts. He has lost more rubles than he cares to remember at the casino. And now, unbelievably, he sees in the window of a local grocery store a single packet of Doritos. And he hears angels weep.
Under normal circumstances, Doritos don't excite him. But today, this week, the tourist fairly loses his mind. He sprints into the store, pays the equivalent of $20 for the snack-sized bag and feels drool gathering at the corners of his mouth. You see, after seven days of jellied sardines and pelmeni that sit in your stomach like rocks, licking some good ol' American nacho cheese dust direct from an empty bag sounds so good, he can barely stand it.
I'm fairly certain they don't play fantasy football in Petropavlovsk-Kamchatsky. But even in the world's second-largest city inaccessible by car (look it up!), relative values are in very much in play. Sometimes the worst-tasting Dorito is still manna from heaven. And sometimes the best-tasting jellied sardine is the best you can do.
Which brings me to this question: Why does conventional wisdom tell you to load up on running backs and wide receivers early in your fantasy draft? If quarterbacks average the most fantasy points (and in 2009 eight of the top 10 point-getters were quarterbacks), why shouldn't you always take a quarterback with your first-round pick? The answer, of course, involves relative value. A football player's fantasy value isn't based on the absolute number of fantasy points he scores, but rather the extent to which he outscores other players at his position. If you have enough players on your fantasy team who outscore your opponent's players at the same positions, you are going to win games.
Allow me, then, to re-introduce the concept of "Value-Based Drafting." VBD is a way to compare players across all positions, rather than just comparing quarterbacks to quarterbacks, running backs to running backs and so on. I make absolutely no claim to being the inventor or even a particularly early adopter of VBD; heck, Pro Football Reference has been using it for years (and attributes its formulation to Footballguys.com). But I do use VBD regularly. It doesn't result in perfect Draft Day strategies because, well, nothing does. (Reality has a nasty way of impinging on our fantasy worlds, doesn't it?) But VBD does inject rationality, deflate hyperbole and give me a strong baseline from which to build my draft board.
Determining "Baseline" players
Conceptually, VBD is simple. Calculate the difference between each player's absolute fantasy point total and the point total of a baseline player at the same position. This gives you a relative number, which can then be compared across positions. Logically, this allows us to look at past-year performances, determining which positions tended to justify higher draft slots. It also allows us to examine the upcoming season's projections and come up with our draft board.
The issue, of course, is figuring out what we mean by a "baseline" player. Do we compare each player to the average starter at his position? Do we compare each player to the worst starter at his position? Is there some other way to determine the baseline player at each position?
At first blush, the "Worst Starter" theory makes sense to me. As the logic goes, if you forgo this player you're considering drafting with your next pick, and in fact forgo this player's entire position until every other team fills its starting lineup, who'd be the best guy you could get?
For example, if I'm considering Drew Brees (ESPN's top-rated quarterback for 2010) in a 10-team league, I should compare his absolute fantasy point total to the No. 10 quarterback's, because if I let all the quarterbacks go by, that No. 10 guy is who I'd get somewhere down the road in my draft. This would make the No. 10 quarterback our baseline player. It would make the No. 10 tight end and No. 10 defense baseline options, too. If your league has a flex (RB/WR) position -- as does ESPN's standard game -- that complicates the worst starter at the running back and wide receiver positions because some teams will start a running back at their flex spot, and some will start a wide receiver. Let's split the baby, and say our No. 25 RB and our No. 25 WR are our baselines at those positions.
With that, let's look at the 2009 fantasy results through this lens. As I mentioned earlier, in terms of absolute points, eight of the top 10 point-earners were quarterbacks last year:
Top 10 overall fantasy scorers (2009)
The top fantasy point-scorers in 2009.
But if we compare each player to the "worst starter" baseline at his position, here's our list:
Top 10 in "Worst Starter" VBD points (2009)
The top point-scorers compared to the "worst starter" at his position.
Clearly this is a list that presents a much more palatable sense of what really happened in fantasy football in '09. And in truth, there isn't a ton wrong with the "Worst Starter" model.
But the guys at Footballguys.com have evolved this model, I think for the better. The truth is that when I look at the "Worst Starter '09" list and see, for example, Vernon Davis as the season's 12th-most-valuable fantasy player, it smells a little funny. I mean, Davis was great. But so were a lot of other tight ends. And if I'm using this model with an eye toward the future, I definitely smell something funny. I plugged our 2010 fantasy stat projections into a "worst starter" spreadsheet (for a 10-team league, with a flex), and among some very astute suggestions about which players to take where, I was instructed to:
• Select Dallas Clark with the No. 22 overall pick (seems too high).
• Wait until the 30th overall pick to select Ryan Grant (seems too low).
• Take the Dolphins' prolific rushing tandem, Ronnie Brown and Ricky Williams, at the start of the eighth round (seems too low).
• Take the Giants' Steve Smith with the 33rd overall pick (seems too high).
To me, this doesn't pass the sniff test. I think the "Worst Starter" method of determining baseline players isn't as sharp as it could be because it doesn't do enough to weigh the relative risks associated with each position. Quarterbacks get hurt, yes. Tight ends get hurt. But when they don't get hurt, their performance tends to be relatively stable. It's the reason Peyton Manning has been near the top of the charts for a decade. Meanwhile, I think it's arguable that receivers and (especially) running backs are more likely to get hurt, but they're also likely to take sudden downturns related to all sorts of factors, including poor offensive line play, defensive game planning, a teammate's sudden ascendancy, a committee, a vulture back on the goal line and the cruel vagaries of the universe. You know this. You've waited to select quarterbacks and tight ends before so you could stock up on running backs and receivers even though you've already filled your starting rushing and receiving spots. Rushers and wideouts are the (ahem) Doritos of fantasy football. It's very hard to take just one.
Instead, I like the Footballguys idea of figuring how many guys at each position are typically drafted within the first 100 players, and using those numbers as my baseline. For a 10-team league, among quarterbacks this baseline player is the 16th-ranked quarterback, because on average 16 quarterbacks are drafted in the first 100 picks. At other positions, it's the 38th-ranked running back, the 33rd-ranked wide receiver, the ninth-ranked tight end and the fourth-ranked defense. Why 100 players? I agree, it's relatively arbitrary. But I think it captures the notion that you don't have to fill all your starting positions (QB, RB, RB, WR, WR, RB/WR, TE) with your first seven picks. Rather, it contends that there's often value in not doing so, and instead loading up with a few extra players at riskier positions (i.e. RB and WR). With this "Top 100" model, we assume that, in a 10-team league, each team will have filled out its seven primary starting spots after Round 10, rather than Round 7.
To close the circle, let's look at the top 10 performers from '09, using the "Top 100" baseline method:
Top 10 in "Top 100" VBD points (2009)
The top point-scorers compared to the baseline of the top 100 players.
Compared to the "Worst Starter" baseline, all we did was flip Gore and Andre Johnson. But further down the list, Vernon Davis goes from the No. 12 player in fantasy to No. 32. Doesn't that sound better?
Top 100 method looks forward
Sure enough, using the "Top 100" baseline method on those same 2010 projections clears up some of the more puzzling recommendations for this summer's fantasy drafts (including the four bulleted above). Here's what our top 50 VBD looks like using this method for the upcoming season:
We don't get our first tight end until No. 51 overall, and we don't get our first defense until No. 72, and those sound about right. (And remember, these calculations are based on our group ranks; personally, for example, I prefer Adrian Peterson over Chris Johnson.)
Are you swimming in numbers right now, ready to exile me to Kamchatka for going on like this? If you remember nothing else about VBD, remember that it helps explain why you don't just automatically go after the highest absolute point-getters. By way of example, look at the following players from last year, chosen from different levels in the '09 fantasy-scoring hierarchy, to see the vast difference in actual value they supplied:
Why not to be dogmatic about VBD
Any system that promises to give you an absolute draft order is obviously writing checks its rear end can't cash; it's impossible to account for the hundreds of caprices that the forthcoming NFL season will dish our way. And let's realize this about VBD: It's still built on projections, and projections are, by their very nature, almost entirely wrong. (I mean, if I guess that Ray Rice will gain 1,248 rushing yards and he gets 1,249, technically I didn't nail it.) Biases and what will turn out to be flawed assumptions are baked into any draft board, in any sport. We do what we can.
Beyond that, there are more reasons to be flexible when using the VBD concept:
How many players at each position have already been taken? Your VBD-based cheat sheet might insist that you should take a quarterback with your next pick. But if nearly everyone else in your league has already grabbed a starter at quarterback, you might be able to afford to wait on your preferred signal-caller, get another running back or wide receiver, and still get the same quarterback with your next pick.
How many picks until my next selection? The extent to which you'll feel safe waiting depends on where you're drafting, and how close together your VBD ratings at a given position are at any given moment. For instance, if you're picking in the middle of a snake draft, you'll have to wait as many as eight picks (in a 10-team draft) between each of your selections. That certainly creates less incentive to let the highest-rated VBD player on your board slip by, unless there are several similar VBD players remaining at a single position, about whom you're basically indifferent.
Understanding ADP. According to our projections and resulting VBD analysis, Brandon Jacobs ranks 51st overall in VBD. But that doesn't necessarily mean you should take Jacobs at the very beginning of the sixth round. As of this writing, according to MyFantasyLeague.com, Jacobs is going 72nd overall in mock drafts. If you can convince yourself that your league mates will adhere to Jacobs' Average Draft Position (ADP), you can afford to wait a round (and maybe two) to grab him. (Of course, if you're running your league for free on ESPN.com -- and what self-respecting fantasy football fan wouldn't? -- your league's default draft-room ranks will be our default ranks, and you might not be able to find as many such gaps between ADP and VBD.) Naturally, waiting to draft a player is a game of chicken, which leads us to
Knowing your league. You know your league better than I do. Are there a bunch of sheep owners ready to grab big names regardless of their 2010 prospects? Do 25 running backs tend to come off the board every year in your league's first 30 picks? Does everyone grab a starting quarterback in the first two rounds? Use what you know about your league to tweak your VBD list.
Conclusions
In the time it's taken me to write this article, our Petropavlovsk-Kamchatsky tourist has boarded a plane, flown to Casa Grande, Ariz., and taken a tour of the Doritos manufacturing facility there. (And perhaps begged a few free samples.) But my pain should be your gain.
I'm always floored when I answer chat or Facebook questions that start out, "Should I go WR-WR in my first two rounds?" Gah! There's obviously no correct answer to that question. It depends on the players that get selected before your pick. It depends on your evaluation of talent leading into this year's NFL season. And it depends on value. Hopefully the concepts underlying VBD can help you make informed inter-positional decisions about which players at which positions represent significant value, relative to the lesser lights at their position. Win the value battle, and you'll begin your fantasy season with a leg up.
Christopher Harris is a fantasy analyst for ESPN.com. He is a six-time Fantasy Sports Writing Association award winner. You can ask him questions at www.facebook.com/writerboy.