Every week with my columns I try to write about what I think my audience wants to read about. Some weeks, figuring out what that might be is hard and requires a lot of educated guesswork. This week, however, figuring out what that might be has been relatively easy. Everywhere I turn within the fantasy community these days I have been bombarded with questions about tanking. What is tanking? Is this tanking? Is tanking bad? How do you prevent tanking?
Indeed, it seems we are deep within Tanksgiving Season, and all anyone wants to talk about is the turkeys in their leagues. So let’s talk about it.
What Is Tanking?
Part of the problem with discussing tanking is pinning down what, exactly, is tanking. Many will take the position that tanking is like obscenity; they know it when they see it. The problem is that if tanking exists solely within the eye of the beholder, it won’t be long before you get two beholders disagreeing about what it is they see.
Tanking is not merely losing. Bad teams lose all the time. Nor is it making your team worse in the short run. An owner that traded Frank Gore for a future 1st rounder would clearly be less able to compete right now, but I doubt anyone in dynasty would begrudge him or her for making that move.
Tanking is not even making your team “objectively worse”, (with ironic quotes given the impossibility of any sort of objective quality measurement in something as unpredictable as dynasty fantasy football). If a team traded Jordy Nelson for Emmanuel Sanders, it would be worse by pretty much any conceivable measure of value you chose to apply. Still, intent matters in this instance; if the owner was a huge Broncos fan who had in the past demonstrated a willingness to overpay to acquire players from his or her favorite team, I don’t think there’s any realistic way to label that trade “tanking”.
We need a definition of tanking that’s simple enough to cover an endless array of possible situations, but precise enough to avoid returning dozens of false positives. In my leagues I use a definition that has served me quite well over the years. Obviously everyone else is free to use their own definition, but for the sake of clarity, this is the definition I will be using for the rest of this article:
Tanking is taking actions an owner otherwise would not have taken if draft order was not a consideration.
The easiest way to test for tanking is to assume that draft order had already been decided using some existing system. The system itself is not important- we can imagine that draft picks were awarded alphabetically, for all it matters- it is only important that draft order is already set. Let’s give some examples:
Example #1: I am a terrible team eliminated from the playoffs. Jamaal Charles is my most valuable asset. I accept an offer where I give Jamaal Charles and receive two first round draft picks.
In this example, my team is undoubtedly getting worse, and my own draft position will almost certainly improve; however, the question is not whether my draft position benefits from my actions, it is whether I would have taken that action even if my draft position did not benefit. And in this case, yes, I would have still made that move even if it didn’t help my draft position. Jamaal Charles will be 28 next year, and will be well into his decline by the time my team has been rebuilt. It makes sense to get strong value for him while I still can. The improvement in my draft position is a happy byproduct, not a main goal.
Example #2: I am a terrible team eliminated from the playoffs. Frank Gore is my #2 RB, and I do not have anyone behind him who is currently playing more than a handful of snaps per game. I drop Frank Gore to roster Duron Carter.
Again, my team is obviously getting worse and this move now means I’ll essentially be taking a 0 from one of my starting positions. Is this tanking? By applying the test, it is not- Carter has more expected value to my roster than Frank Gore does straight up. Again, the improvement in draft position is a happy bonus and not the main goal. The main goal is to replace an expiring asset with a quality prospect.
Example #3: I am a terrible team eliminated from the playoffs. I trade Mark Ingram to acquire Adrian Peterson. In my personal expectations, I think Ingram and Peterson are very similar assets, though I think Ingram might be the tiniest hair more valuable in 2015 and beyond.
Despite the fact that this superficially seems like a relatively even trade, this is, in fact, tanking. Because I think Ingram is more valuable in 2015+, (regardless of by how much), there has to be some reason other than “player value” motivating me to make that move. In this case, I was motivated by a desire to lessen my production this season and earn a higher draft pick. Ingram might have been a hair ahead of Peterson, but the extra draft considerations are what ultimately tip me over the edge and make me willing to do a deal I would not have done absent draft considerations.
Example #4: I am a dominant and deep team in a league that awards playoff positioning based on potential points. Team A is the worst team in the league by a substantial margin in potential points, but due to a favorable schedule is one win away from securing a playoff berth. I currently own the 2015 first rounder of the 2nd-worst team in the league by potential points. I trade my #5 wide receiver at a discount to Team A, helping him make the playoffs and bumping the 2015 draft pick that I own up from #2 overall to #1 overall.
This is a unique scenario that actually happened to me this year. This is the kind of situation that would not get picked up under most definitions of “tanking”. With that said, it fits many of the key criteria- I am deliberately making my team worse for no reason other than to improve my draft position. The fact that this specific scenario fails my “tanking test” is, in my mind, one of the biggest strengths of my particular definition of tanking. It can pick up on much subtler and more complex behavior than a simple “deliberately losing” definition.
Bad Reasons To Oppose Tanking
A lot of the complaints about tanking come from a position of self interest. You see bad teams complaining that another team is tanking and going to be rewarded with a higher pick as a result. You see good teams complaining that their competition is being handed free wins by tanking franchises and they might miss the playoffs as a result. The upshot is “these things hurt me, so therefore they are bad”.
The problem is that self-interest, as a rule, makes for terrible policy. Self-interest is the reason why allowing leagues to veto trades rarely ends well. Many owners will see a perfectly fair trade that makes one of their rivals better and cast a veto out of pure self-interest. Thus, by everyone acting in their own interest, you wind up with a league where trading is impossible, which is in no one’s interest.
The problem is that fantasy football is by nature a zero-sum game. Every gain by one team is, by definition, a loss by another. If one team gets a win, another team takes a loss. The zero-sum nature extends beyond simple wins and losses down to odds, too.
Consider this: in a league where six teams make the playoffs every year, if you add up the true playoff odds of every team in the league, you will always wind up with 600%. If you have twelve perfectly even teams, then each team will have a 50% chance at making the playoffs, and 12 times 50% is 600%. If you have six teams that have clinched playoff berths and six that have been eliminated, then the odds for those that have clinched are 100% and the odds for those eliminated are 0%. 6 times 100% plus 6 times 0% gives us 600%.
That 600% is a fixed pool, and each individual team can take anywhere from 0% to 100% from it. Anything I do that improves my own individual odds must, by definition, reduce the odds of the rest of the league. If I take my odds from 50% to 75%, then the rest of the league only has 525% left to divide among the remaining eleven teams. If that reduction is spread evenly, then when I increase my odds by 25%, everyone else sees their odds decrease by 2.3%.
It’s not just playoff odds that behave this way. Odds of winning a championship and odds of earning the #1 overall pick are both fixed, and across the entire league they will always sum to exactly 100%. Yes, someone tanking will hurt the odds of everyone else getting the #1 overall pick, but tanking isn’t bad because it hurts those odds. Everything everyone does is going to have some impact on those odds, and we aren’t about to outlaw everything.
Looking at things from the league’s perspective, the league doesn’t care whether the #1 overall pick goes to Dave or it goes to Sarah, and so complaints from Dave about how Sarah is hurting his chances do not seem particularly compelling, just like complaints from owners when their rivals are making good trades should ideally fall upon deaf ears. Everything that is bad for Dave should be equally good for Sarah, and their contrasting outcomes should offset and sum to zero. Hence the phrase “zero-sum game”.
Good Reasons to Oppose Tanking
So everyone’s odds behave in a zero-sum manner and the league doesn’t particularly care the name of the team that gets each pick. Why, then, is tanking bad from the league’s perspective?
While the league as a whole may not care which particular team gets which particular pick, the league still has a preference about which TYPE of team gets which pick. Leagues are explicitly designed to funnel the best picks to the worst teams. Leagues are specifically geared towards sending the best teams to the playoffs.
Tanking causes a league to deviate from its design. If not-the-worst teams are able to tank their way to a top pick, that circumvents the express design of the league. It also distorts the top half of the league, too- if teams are making the playoffs, not because they’re the best, but because they were handed free wins at the end of the year, then that’s also violating the design of the league.
To put it another way: if I donate a game to Toys for Tots, I should not care whether it goes to a 6-year-old named Jennifer or an 8-year-old named Alex, but I should absolutely be pretty upset if it winds up in the hands of a 38-year-old named Todd.
Now, no league is going to be a perfect meritocracy, where the worst team always gets the best pick and the best teams always make the playoffs. To some extent, it’s impossible to define such ephemeral terms as “worst” and “best” in fantasy football. Even if we could define it, the identity of the best and worst teams would change relatively frequently, possibly even weekly.
Moreover, most leagues have added elements of luck and randomness, (with head-to-head scheduling being by far the biggest), largely because they add excitement and fun. Most leagues are willing to ease up on their meritocratic ideals to fit more fun into the mix.
The problem is that, while tanking likewise spikes the meritocratic spokes, it doesn’t add any fun to offset. There’s no excitement about playing a team whose starters are all injured or on bye. There’s no fun in having playoff seeding decided by whoever got to face the bad teams late in the season after they’ve already given up instead of early in the season when they were still trying.
As a result, while from an individual self-interest standpoint the gains and losses of tanking all offset and zero out, from the perspective of the league as a whole tanking is a net negative. Leagues where no one ever tanks are better for all involved than leagues where tanking is rampant, even if all else is equal. Therefore, it is sensible and defensible for leagues to legislate against tanking.
Turning a Footrace into an Arms Race
A McDonald’s and a Burger King are located on a busy street. Their signs are right next to each other, such that for the southbound traffic the McDonald’s sign is blocked and all they can see is the Burger King, while for the northbound traffic the Burger King’s sign is blocked and all they can see is the McDonald’s.
Both restaurants exist in a state of equilibrium for a while until the McDonald’s owner, sensing the opportunity to gain an advantage, pays $1000 to commission a new taller sign. Now McDonald’s remains the only sign visible to northbound traffic, but southbound traffic is able to see both.
The Burger King owner, accustomed to getting 100% of the southbound diners and 0% of the northbound diners, sees his clientele cut in half as he is now splitting the southbound diners with the McDonald’s. In response, he spends $2000 to create an even taller sign still. Now he once again obstructs the McDonald’s sign from the vantage of southbound traffic, but both signs are visible to northbound traffic.
The McDonald’s owner responds by raising his sign even more, and the Burger King owner raises his sign even more. The two restaurants trade salvos until finally they reach an upper limit; whether due to prohibitive cost, or the limits of engineering, or municipal ordinance, they have reached the highest they can construct their signs.
In the end, the two restaurants settle back into the same equilibrium, with the Burger King (and its 100-foot tall sign) visible to southbound traffic and the McDonald’s (and its 100-foot tall sign) visible to those heading north. At the end of the day, they wind up exactly where they were in the first place, but with both parties out however much they spent to build and rebuild and re-rebuild their roadside signage.
This is the perfect example of an arms race- a situation where two parties engage in a competition that has no absolute goals and where success is measured solely by position relative to the other competitors.
The eventual outcome of an arms race is always the same- all parties wind up in the same relative position that they were before they started, but out whatever resources they had to spend on the arms race. Basically, arms races make everyone worse off.
Tanking can very easily devolve into an arms race. Losing in fantasy football is trivially easy, if someone is so inclined; there are always players who are not getting touches in any given week, and an unscrupulous owner could easily put up 0 points week after week after week until the season ended.
If two owners with the same record both decide to start tanking, then at the end of the process both owners will… still have the same record. The scoring differential between them will still be the same. Neither will have improved his relative standing in the slightest. They will, however, have wreaked havoc on the league’s competitive balance in the process.
And if owners fear getting into an arms race, they might decide to begin the tanking process even earlier. Remember, once all parties commit to tanking, relative gains cease to be made. In order to get ahead, a team must start tanking before anyone else would countenance the thought. This creates another kind of race, as teams race to be the first to give up on the season. Carried all the way out to its logical conclusion, in a league where anything went you might see bad teams scoring 0 points for the entire year as they all recognize in advance they have no shot at the title and begin gearing up for draft position as early as they can.
When faced with the possibility of an arms race, the only winning move is not to play. If the McDonald’s and Burger King owners had merely agreed from the outset not to raise their signs, they both would have been much better off. Likewise, in order to prevent an arms race that leaves no one better off and everyone worse off, it is in a league’s best interest to legislate against tanking.
In Which I Preemptively Apologize For the Most Obvious Pun Ever
So that’s tanking in a nutshell— what it isn’t, what it is, why it isn’t bad, why it is bad, why leagues should do something about it. For a discussion of what, exactly, leagues can do to combat tanking, tune in to this week’s Dynasty, in Practice column; practical advice isn’t really what we do here. If you want to continue the discussion or felt I otherwise overlooked something, please feel free to give me a shout on Twitter. Otherwise, tanks for reading!