Play fantasy football? Try our 2010 Fantasy Football Subscription.
By Jonathan Bales
Fantasy football is increasingly becoming a game of mathematics and game theory. Yes, football knowledge is important, but I would be willing to bet I could develop a computer program which could overtake 95 percent of fantasy owners over the long haul.
Thus far in my fantasy football articles, I’ve spoken about how to predict a running backs’ yards-per-carry, how to use tiers to garner maximum value, and why 2010 is the year to draft a quarterback in the first round.
Today, I will talk about the act of “handcuffing” your top running backs, i.e. drafting their backup as insurance. Handcuffing isn’t a particularly new strategy, but it is one that hasn’t been questioned much. The primary reason owners handcuff their top running backs is to limit the downside of an injury. Frank Gore owners who draft Glen Coffee, for example, can just plug him in if Gore goes down.
I have never been a fan of handcuffing. While it does limit the downside of losing a player to injury, it also limits a team’s upside. Being an intelligent fantasy football owner means invoking a combination of both: limiting downside while maximizing upside.
In game theory, this term is called “minimax.” In short, it is an attempt to maximize the minimum gain. Its fantasy football equivalent is VORP:
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.
Ultimately, the question we must ask (regarding handcuffing your running backs) is whether or not the minimization of downside outweighs the loss of possible upside. Let’s take a look at a real-life example, displayed on the chart below.
Suppose you are a Frank Gore owner and are looking to handcuff him with Glen Coffee. The players listed at the top of the chart represent a running back/wide receiver combination one could expect if Coffee is selected, and one if he is not.
Notice that Jeremy Maclin is a better option than Johnny Knox (3.0 more points/game) because owners who handcuff must select that backup player earlier than normal to ensure they do not lose him. Thus, the initial “cost” of handcuffing Gore is the loss in wide receiver points.
The VORP category lists the amount of “extra” points the combination would gain (or lose) in certain situations. For example, the Coffee/Knox combination would only be suitable to the Maclin/Scott combination if Gore gets injured. But what are the chances of that? 5 percent? 10 percent?
Even if we assume Gore has a 20 percent chance of getting injured in any particular game, the statistics still show handcuffing is a poor strategy. The reason? Whether or not Gore gets injured, Team B is securing more value from the wide receiver spot.
Perhaps I should use more conservative numbers. Instead of assuming wide receiver B (Maclin) would score three points more per game than wide receiver A (Knox), let’s assume it is just one point. Further, let’s raise the likelihood of Gore getting injured to 15 percent.
In that situation, Team A would hold a total VORP of +0.65 (-1.0*0.75+10.0*0.1-1.0*0.1). Team B would hold a total VORP of +1.0 (1.0*0.75+1.0*0.15+1.0*0.1). Thus, even with incredibly generous assumptions, handcuffing (at least in the middle rounds) appears to be a poor strategy.
The reason handcuffing is such a suboptimal strategy is the low injury rate. Even NFL running backs aren’t that likely to get injured. Further, I have been doing my simulations as if the injuries occurred immediately. If Gore got injured mid-season, the VORP of Team A would be cut nearly in half.
So, is there ever a time when handcuffing is a sound strategy? Yes. In my previous example, the VORP of the wide receivers was realized at all times because I assumed Maclin and Knox were approximately No. 3 wide receivers (and thus starters in most leagues). On the other hand, Coffee’s VORP was only realized during a Gore injury, as that is likely the only time he would be started. Scott’s VORP was non-existent because he is not a handcuff in that scenario and would probably never be started.
If the selection of the handcuff comes at a time after all positional starters are filled and no significant VORP can be acquired, however, then it may be a prudent strategy. However, at this period of your draft, there is no way to secure any of the top “handcuffs”–Darren Sproles or Tim Hightower, for example.
Instead, the Steven Jackson owners of the world should be targeting Chris Ogbonnaya. He can be drafted in one of the last rounds and the potential upside of another player is probably not as “great” as the limited downside which comes with Ogbonnaya’s selection.
As a general rule of thumb, don’t look at handcuffing your players until at least Round 14 or so. Up until then, game theory dictates you bypass the ability to limit your downside by securing high-upside players.
Besides, you don’t win fantasy football championships by being conservative.