clock menu more-arrow no yes mobile

Filed under:

NFL All Star - Lesson 05 - Handcuffing

Daily fantasy football is a game of identifying value, but how we define ‘value’ can change. In tournaments, value might be upside minus cost, whereas that’s different in cash games. One of the things that undoubtedly affects overall lineup value is correlated play. When you start…

Daily fantasy football is a game of identifying value, but how we define ‘value’ can change. In tournaments, value might be upside minus cost, whereas that’s different in cash games.

One of the things that undoubtedly affects overall lineup value is correlated play. When you start a wide receiver in one game and a wide receiver in a different game, their production isn’t correlated; what one scores has nothing to do with what the other scores. When you start two players on the same team or in the same game, however, their production is now linked. A wide receiver’s production is obviously connected to his quarterback, who also affects his running back, and all of their production is linked to the play of their defense, and so on.

Harnessing and exploiting these correlations is one of the best ways to find value in daily fantasy football. When you pair two or more teammates together in your lineup to acquire the benefits of correlated play, it’s called ‘handcuffing’ or ‘stacking.’ There are different ways to handcuff different players to improve your team’s odds of winning.

With handcuffing, you’re basically trying to get something for nothing. When you start a quarterback with one of his wide receivers, you’re improving your lineup’s upside without necessarily needing to spend any additional money on those positions. Here’s a look at the correlation between quarterback and wide receiver points.

This is a sizeable effect. When a quarterback has scored 30 points on DraftKings, the average points scored by his wide receivers has been just over 70 points. This is why handcuffing a quarterback with one or more of his receivers has been such a popular and fruitful DraftKings tournament strategy.


The easiest way to determine how useful certain types of handcuffs are is to look at the historic correlation between their play. Luckily, rotoViz has already calculated this for us in this image. Every correlation is shown from -1.0 to 1.0. A negative correlation means as one thing goes up, the other goes down. There’s a negative correlation between the points for a quarterback and the opposing defense, for example.

A positive correlation means that as one thing increases, so does another. There’s an obvious positive correlation between quarterback and wide receiver scoring. The higher the number, the stronger the positive correlation.

QB/WR1: 0.34

This number isn’t a surprise. There’s a fairly strong correlation between a quarterback’s fantasy scoring and that for his top wide receiver.

QB/WR2: 0.39

Believe it or not, this is the strongest correlation of any teammates in football. That doesn’t mean that WR2s score more than WR1s, but just that their production is more closely linked (barely) to the quarterback. This actually makes intuitive sense since, when an offense really goes off through the air, they’re usually getting more than just their top receiver involved. Meanwhile, a top receiver can sometimes produce decent numbers even if the offense struggles as a whole.

QB/WR3: 0.35

The correlation between quarterback scoring and wide receiver scoring is about even for each of the top three receivers on an offense. Again, this is related to game script; when a No. 3 wide receiver does well, there’s a good chance that the offense was rolling. Of course, WR1s and WR2s are still the best to pair with passers just because their bulk production is so much greater.

QB/TE: 0.16

Surprisingly, tight end production hasn’t been as closely linked to quarterback production as that for receivers. This is especially surprising to me since tight ends are usually dependent on touchdown passes from their quarterbacks to put up elite production. It could be that we’re seeing the overall numbers skewed by a handful of teams that effectively don’t use their tight ends.

QB/RB: 0.09

While this correlation is weak, it’s still positive, i.e. as quarterback production increases, so does that for running backs.

RB/WR1: -0.07

However, running back production is negatively correlated with wide receiver production (just barely). The same goes for WR2s and WR3s.

RB/Opp RB: -0.31

Finally, note that there’s a strong negative correlation between production for opposing running backs.


1. Definitely handcuff a QB/WR in GPPs.

When it comes to handcuffing, nothing is more powerful than QB/WR.

2. Consider using a QB/WR/WR stack in tournaments.

Due to the strong correlations for all receivers and their quarterback, a QB/WR/WR stack makes a lot of sense, especially on high-upside offenses. I think there has been some concern that wide receivers on the same team can cannibalize on another’s fantasy scoring, but it appears like there’s enough upside to go around.

3. QB/RB/WR stacks haven’t been effective overall at increasing upside.

There might be times when it’s okay to add a running back into your lineup when you handcuff a quarterback and wide receiver—primarily when that running back can catch a lot of passes—but overall, wide receivers and running backs tend to rely on different game scripts and end up stealing points from one another.

4. Don’t start two running backs in the same game.

One of the strongest negative correlations in fantasy football is between running backs in opposing teams. That should be obvious since, unless the game is tied in the fourth quarter, one team is likely going to be forced to throw and probably hurt their running back’s production. There’s almost no time that it makes sense to start opposing running backs.

I am a promoter at DraftKings and am also an avid fan and user (my username is bales) and may sometimes play on my personal account in the games that I offer advice on. Although I have expressed my personal view on the games and strategies above, they do not necessarily reflect the view(s) of DraftKings and I may also deploy different players and strategies than what I recommend above.