Why Standard Fantasy Football Rots and How To Fix It


My mother-in-law plays a dice game called Farkel. It's a game of simple math, but mostly a game of chance—more War than chess, more cleromancy than game theory. If I were stuck with her in a windowless room without my toothbrush but with a woman named Estelle, if we were stuck for all eternity, perhaps we could determine if her, ahem, extreme prudence or my more reckless style were superior. But as-is there's a lot of guessing, a lot of premature revelry and a lot of empty opining about strategy. If I may: the game sucks. It's an excuse for people of a certain age to get together and drink.

Fantasy football is a $70 billion market. I can't believe I just wrote that. That … that number was not planned.

… give me a second …

23 million people play fantasy football, and while I could not find an exact number, a not terribly scientific or thorough accounting of the number of Yahoo public leagues versus the number of Yahoo private leagues, indicates most play with standard rules. And many customized leagues are close enough to standard rules to, for my purposes, be comparable.

Fantasy football played by standard rules is a rotten game, requiring little skill, that, in its crappiness is a bad reason even to get together and drink. Here's why:

Way too much value at the top

Using Pro Football Reference's Fantasy Football tools*, I compiled the last ten years of VBD. If you don't know what VBD is, it's value above baseline. In a standard 12-team league, that means the value of the best quarterback versus the 12th best quarterback, the best running back versus the 24th best running back, etc. Since fantasy football is so designed to reduce all stats into points and put all points into a single bucket, it's easy enough to compare performance across positions. How good was LaDainian Tomlinson in 2006? 271 points better than Warrick Dunn, but also 130 points above top quarterback Peyton Manning. And what does that mean exactly? Well, there's a few ways to think of that.

Here's a graph of the average value of the top 84 players (12(QB + 2RBs + 3WRs + TE)) over that ten year span.
http://mail-attachment.googleusercontent.com/attachment/u/0/?ui=2&ik=73d1a47179&view=att&th=140af97efe58add3&attid=0.1&disp=inline&realattid=f_hkqlxp700&safe=1&zw&saduie=AG9B_P9VkXMuqqNZJTPB07MtA2c7&sadet=1377335665743&sads=IC4cqX_w8MlV78UhL6be3bORQqo&sadssc=1 

You may notice it resembles a power law probability distribution. It's not quite that extreme actually: the top 20% (the top 15 players) only account for 48% of all points above baseline.  Now you might be saying
but I think the implication is clear: Value is concentrated in a few players. Value is especially concentrated in the very best player. The graph becomes very gradual after the first five most valuable players, so that the value difference between the most valuable player and the fifth most valuable player is roughly equal to the value difference between the sixth most valuable player and the 46th most valuable player.

Snaking the draft does not properly compensate

If you were to run a perfectly efficient snake draft, with the best player picked first, the second best player picked second and so on, the resulting teams would project to these point totals:

1: 381.4
2: 337.2
3: 325.8
4: 310.2
5: 296.6
6: 291.6
7: 285.5
8: 284.1
9: 281.2
10: 276.3
11: 277.4
12: 277.8

The most valuable player in any given year accounts for most of that advantage, and value levels out too quickly for a snake-style draft to make up for that advantage.

There is little skill to drafting the best player

You may argue that a draft is not perfectly efficient, and that's true. In fact, that the draft is not run efficiently only deepens the advantage of finding, or should I say, being able to pick that 2006 LaDainian Tomlinson or that 2003 Priest Holmes. 2006 Tomlinson, the most dominant fantasy player in our sample and probably ever, worth 97 points more than 2006 Larry Johnson the second most valuable player in 2006 by VBD, was projected to be the second overall pick. The projected first overall pick: Johnson himself.

True sleeper picks, players that go from mid-round status to become difference makers, are very rare. Most, like 2007 Tom Brady, 2011 Aaron Rodgers and 2004 Peyton Manning, are not so much sleepers as top-rated quarterbacks that have career years. The biggest sleeper of the last ten years was 2010 Arian Foster, who was projected to go at 43, but was worth 192 VBD, 86 more than second place Peyton Hillis. Hillis himself was a big sleeper, as was 2007 Randy Moss and 2004 Larry Johnson, but that's about it. And those picks, and I say this as someone that drafted Foster in 2010, that drafted Moss in 2007 and Peterson last season, are not evidence of some kind of superior gamesmanship or knowledge of the NFL, but a hunch rewarded by luck.

And that's fantasy football in a nutshell. If you pick in the top half of the first round, you have a massive advantage. Barring that, your best hope is something approaching dumb luck. The person that selected Rodgers in 2011 probably did so in large part because their second-round pick landed around 20. Had they the first overall pick, they would have been compelled to select Foster or Peterson or Jamaal Charles and likely had no chance to draft Rodgers. And had they drafted Peterson or Charles, they would have likely lost.

Et Cetera

A head-to-head format, a season that often lasts only 14 weeks, and playoff formats that either admit half the league or compress the season to two games, all only further trivialize what little part skill has in winning at fantasy football. It's not altogether surprising when a dead team wins. That's damning. Even War requires participation. But there's a need to balance the accessibility of fantasy football with its fun as competition. So how to fix it without ruining it? And how to fix it without your suggestion being laughed at?

Use an Auction Draft Format

Advantages: This one's simple enough, and so I won't elaborate too, too much. Auction rids fantasy of its bias toward early round picks. It eliminates the mechanical picking from a list. There's a chance to draft the team you want instead of RB, RB, QB … etc. Managers must put their money where their mouth is when it comes to sleepers. Anecdotally, managers of auction-drafted teams seem more committed to their teams and more willing to play the season through.

Disadvantages: That might be because even with the aid of computers, auction drafts last a dog's age. I couldn't imagine conducting one in a bar among friends—which is incomparably more fun than on Yahoo! against strangers. Auction drafts do not eliminate the hugely disproportionate value of drafting that one guy that flips out. 2006 LaDainian Tomlinson would be worth well in excess of $100 in a draft budgeted for $260. I've yet to see three digits even approached in an auction.

Assign Buckets a la Fantasy Baseball

Advantages: By separating stat categories instead of reducing all to points, it's significantly less likely one player can more or less win you a season. LaDainian Tomlinson scoring 28 rushing touchdowns is then no more valuable than Rickey Henderson stealing 130 bases. It's valuable, very valuable, but it's not enough. You still need someone to score receiving touchdowns, passing touchdowns, passing yards, and whatever else depending on how you segment the stats. This allows for a more or less traditional fantasy draft, and traditional fantasy guides and websites are … well as useful as they ever are.

Disadvantages: For lack of a better way to put it, this sorta nerds up your league. That term's been co-opted. A nerd is now someone that likes technology and sci-fi and fantasy movies—or, as mainstream as an iPhone. But once upon yonder nerd meant weird or strange in some way, and that's the problem with attempting a set of scoring rules that I might be making up right here with this word, this word here. Here. And that means some of the fun is lost. Hemingway said “An intelligent man is sometimes forced to be drunk to spend time with his fools.” To wit, I think common people strive to be different and strange people strive to relate. If fantasy football's one of your ways to connect with others, good luck explaining why Aaron Rodgers' three-yard scramble to kill the clock allowed you to edge out your rush yardage stat and win 7-5.

Also: it might not be fun. I've never tested this.

Reduce the value of touchdowns

Advantages: This is a no muss way of flattening scoring distributions. When Tomlinson scored 431 points in 2006, 186 of those points were rushing or receiving touchdowns. If he scored only four points per touchdown, his VBD would shrink to 218. That's still a ton but it's a step in the right direction. Shrinking it to two would level it further.

Disadvantages: The silly, over-the-top significance of touchdowns in fantasy football is part of the game's appeal. It's also rooted in the game's history, which (fantasy football history) is silly but not over-the-top. Six points per touchdown is intuitive, and allowing for such high value plays creates excitement.

Point per reception

Advantages: This ups the value of wide receivers, putting more high value players in the pool.

Disadvantages: It also ups the value of running backs, making top-end backs that much more valuable—especially since, in the spirit of simplicity, each reception is worth essentially 10 yards. So receivers may be more valuable, but many of the best running backs gain value too. Ray Rice was the most valuable back in fantasy in 2011. His VBD under regular scoring rules was 154. The most valuable wide out was Calvin Johnson at 149. The difference in receptions between Johnson and the baseline receiver, Jeremy Maclin, is 33 receptions so 33 points. The difference between Rice and the baseline back, Benjarvis Green-Ellis, is 67 receptions or 67 points. But that's not entirely fair, because Ellis would be displaced by some schmo that catches a lot of outlet passes. Even by receptions alone, the difference between the top running back, now Darren Sproles, 86, and the 24th running back, Marshawn Lynch, 28, is 58. That approaches the difference (68) between the top receiver, Wes Welker, 122, and the 36th receiver, Julio Jones, 54. In the modern NFL, great running backs tend to be good receivers. Great wide receivers are often high value receivers that trade targets for yards per attempt. Apart from awarding points to something that has no innate value, point per reception leagues devalue great deep threats, do little to reduce the value of top running backs, and sort of bury quarterbacks. As fixes go, it's a botch.

In fact, most manipulations produce a botch. It seems, running through my ideas at least, that an auction-style draft is for the best. It might take a long time but a season takes a much longer time, and better spent on something fair and skill-based than testosterone Yahtzi. People may still not realize just how valuable top players are and may underestimate just how high replacement value is in a shallow league, but that's part of the game. There's nothing unfair about exploiting opponent misconceptions and biases. Win, people will learn from your methods, and competition should improve.

Winning, knowing why you have won, being able to repeat and improve upon that method—this is the foundation of a good game. What undermines fantasy football, what makes it frustrating, lowly and kind of ridiculous, is that for all the server capacity wasted on mock drafts and sleeper lists and matchups and don't-draft lists, no one knows how to win with consistency. Organizing and executing an auction draft will take time. You might have to get a fourth Keller Pils to wash down your second cheese steak. But winning because you worked for it, winning because you're smarter, man that's sweet.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

21 Responses to “Why Standard Fantasy Football Rots and How To Fix It”

  1. John Morgan says:

    *For what reason I do not know, PFR calculates their VBD only to the 30th WR.

  2. Anonymous says:

    Instead of pts for reception, how about pts for 1st down? My intuition is it should increase relative value of WR.

  3. Mitch says:

    This is why I play weekly fantasy football. You draft a new team each week and can play for real money. Most games use a salary cap formula and player's salaries are adjusted weekly -- it levels the playing field but still leaves the sophisticated fantasy player a chance to get an edge.

    (the site I play at is DraftDay in case anyone was wondering, there are a few other decent sites).

  4. nottom says:

    I've always thought pts per First Down would be an awesome idea. Receptions that help you team usually result in a first down and would get the same point from PPR but receptions in the backfield that get dropped for a loss wouldn't be rewarded. It would also help RBs who are very good at converting short yardage (like BJGE with the Patriots) would get a well deserved bump

  5. Anonymous says:

    If a league had the 1st round go one way, and then reversed all the subsequent rounds, how close to even would the perfectly efficient draft be?

  6. Tom Bernstein says:

    I assume that if every team in your league drafted from the same cheat sheet, and they only took the optimal player every time... that would make things considerably boring... wouldn't it.

    If any of this were true the first pick would win almost every league every year, and that doesn't happen because leagues are not won on draft day. Overcoming injuries, bye weeks, and Shanahan running backs is how you win leagues.

  7. Dan says:

    Third round reversal roughly balances out the value of different draft positions, on average. It won't be enough in years where a player has a historically amazing fantasy season like Tomlinson in 2006, but those seasons are hard to predict in advance.

  8. Nick Bradley says:

    Most leagues I've played in over the past 5 years have devalued passing TDs - from 6 to 4. This evens everything out...2012 Drew Brees drops from #1 overall to #14.


    The league I'm in now also has completion percentage (a half a point for every completion over 50%).

    Running backs are also penalized a half point for every carry, but compensated with getting a point for every 4 yards. Essentially, backs are penalized for averaging less than 4 yards a carry. league average is 4.25 YPC

    WRs get a half point per catch, 6 yards per point.

  9. Miles Libbey says:

    Another aspect of the luck involved that annoys me is the head to head schedule. As far as I can tell, there is no tactic to take when playing a specific user. That is, my opponent's lineup has zero impact on my lineup. You could score the second most points in the league in a week and lose. in my leagues, the team that has the most total points scored against it over the season rarely makes the playoffs. This could be fixed by calculating the teams record as if they played every other team each week. In a 10 team league, each team would play 9 games a week.

  10. Andrew Y. says:

    Real football can be heavily influenced by luck as well though. I remember a few years ago when the Chargers had a top offense and a top defense (according to this site) but they gave up lots of special teams touchdowns and had a lot of weird plays go against them and they ended up 8-8 and missing the playoffs. There are always teams a few wins above or below their win expectation. Fantasy football reflects this, and personally I think the fantasy version of the sport should be close to the real version of the sport.

  11. Jason Kramer says:

    Increase the size of starting lineups to increase the need for depth.
    I play in a 2QB, 3RB, 4WR, 1TE, 1Flex, 1K, 1D league, and it's awesome.

  12. Toby says:

    In my opinion, 10-team leagues will give everyone a realistic and somewhat similar chance of winning their league. Twelve-team leagues are too large ... I feel like your article could have carried more merit with me if you had recognized this reality upfront.

  13. Daniel Wick says:

    My league uses a yardage model (12 yds per point) as opposed to the traditional fantasy football model. I derived that based roughly on articles here and other sites to essentially equate yards to fantasy points. Some of the highlights of this are that TDs are worth 2, turnovers are worth -4 (or +4 for DEF), and PATs are worth 0 (missed PAT is -1).

    The league used to be heavily weighted to scoring with yards as a bonus, but after making the switch, almost everyone agreed that the "new" system was better, and more representative of a player's actual output. There was some initial resistance to change -- as there always is -- but moving away from the traditional fantasy scoring breakdown has been a huge success, and I can't recommend it enough.

  14. pacifist viking says:

    Miles: I'm in a league that does just that. We call it Cross Country Scoring, and I've been trying to pitch it to others for almost a decade (we're entering our 13th year). Here's an explanation if anybody is interested:

    Each week, every team competes against every other team. If there are ten teams in the league, the team with the most points that week defeated everybody else, thus going 9-0. The team with the second-most points defeated eight other teams, but lost to one team, thus going 8-1. Yada Yada Yada, the team with the least points was defeated by everybody, thus going 0-9.

    Weekly Scores:
    Most points: 9-0
    2nd: 8-1
    3rd: 7-2
    4th: 6-3
    5th: 5-4
    6th: 4-5
    7th: 3-6
    8th: 2-7
    9th: 1-8
    10th: 0-9

    These are the wins and losses for the single week. For the season standings, wins and losses from every week continue to be added up. Each week, your wins and losses for that week are added to your wins and losses for the season total. Teams scoring the same number of points in a single week have a tie, and a tie will be included in the standings for the week and the season.

    The fantasy season goes all 17 weeks (no playoffs). At the end of the season, the team with the most wins is the league champion (1st tie-breaker: best winning percentage 2nd tie-breaker: total points).

  15. pacifist viking says:
    This comment has been removed by the author.
  16. pacifist viking says:

    I'll add that the benefits of Cross Country Scoring are two-fold:

    1. Fairness: instead of teams with bad scores getting Ws and teams with good scores getting Ls just because of weekly matchups (silly when you can't control your "opponent's" score anyway), every team gets pretty fairly rewarded for its performance. If you have the 2nd best score, 8-1 is a good week. If you were mediocre, instead of luck determining whether you get a W or L, you go something like 4-5 or 5-4.

    2. Fun: every single NFL game is fantasy relevant to you. Instead of, say, 18 players mattering to your fantasy prospects each week, 180 do. This makes Monday night really, really exciting.

  17. Nate says:

    It seems like fantasy football is more about entertainment value than anything else. Starting with that notion, maybe things should be structured so that players can make more meaningful decisions as the season goes on rather than just keeping track of the numbers. Similarly, some kind of rubber band mechanic might help keep more people in contention longer.

  18. Luther says:

    John,

    First of all, let me say that I agree completely with the substance of your post. The enormous variance in week-to-week (and year-to-year) performance of NFL players (and teams) makes it all but impossible to eek out any discernible edge against your opponents. However, I'm curious to know if you think that playing fantasy football in volume would in any way change your analysis.

    Here's what I mean: suppose it was the custom with amateur poker fans to get together once a year with their friends and play a single winner-take-all, no limit Hold'em tournament. Many of the analytically-minded poker writers and pundits would rightfully point out the almost overwhelming randomness of such an event. No player at the table will ever see enough hands for his or her true skill to emerge. The results or each individual tournament will be highly random.

    Eventually, someone would point out that much of the inherent variance of the tournament could be mitigated by playing lots and lots of tournaments over the course of a year rather than just the traditional single entry. I know that this is true for poker; do you think that it might also be true for fantasy football?

  19. Anonymous says:

    Not to put too fine a point on it, for 23 million people to make a 70 billion dollar market, they average $3043 each. That seems a big high.

  20. Anonymous says:

    Point per first down reception seems like a perfect compromise that I really wish would catch on. Seems like more and more people are flocking to auctions every year so hopefully snake drafts will soon be a thing of the past.

  21. John Morgan says:

    Luther, there is certainly some skill to playing fantasy football. You need to know Percy Harvin is hurt, for instance, and not draft him. But much of that skill is removed by the fantasy provider: through pre-draft rankings and up-to-the-minute news. Maybe there's some skill yet beyond that ... but how rare is that skill and how hard is it to attain? I would guess it's common and pretty easy, respectively.

    Unlike poker, to understand if someone is good at fantasy football, you would need both volume of teams and volume of years. Otherwise, a player with a lot of good hunches one season, that is, a fantasy owner that picked, say, Arian Foster in 2010, would seem good but good in a way that maybe couldn't be repeated.

    I don't want to say fantasy football isn't fun at all, or that it's a total crapshoot, only that it's too damn random and would be more fun if it were more reliant on skill.

Leave a Reply

Note: Only a member of this blog may post a comment.