Stick to Stacking? Exploring the Classic Boom/Bust Strategy (Fantasy Football)
If you have been playing fantasy football for a while, you might have heard the term ‘stack’ thrown around. A ‘stack’, simply put, refers to multiple players from the same NFL team in your fantasy lineup.
The idea behind ‘stacking’ is simple: imagine that you have Chargers’ duo Justin Herbert and Keenan Allen rostered on the same team (QB/WR is the most common type of stack). If Herbert throws a touchdown to Allen, you score double the points: once for the passing yards and TD and once for the receiving yards and TD. However, if Herbert can’t get anything going and doesn’t sling any touchdowns, Allen won’t, by definition, catch any!
Because of this dynamic, stacking is considered a boom-bust technique. By stacking, you increase the probability that your roster will ‘go off’ because of the double-scoring when your stack ‘makes music’ together. However, you also increase the probability of busting: if one player in your stack has a poor outing, it’s more likely the other did as well.
More formally, players in stacks are correlated: their performances tend to move together. This creates a sort of ‘all your eggs in one basket’ situation for your lineup. The question, though, is how much? What kinds of correlative strength can we expect from the different types of stacks, and how does this change the profile of your roster’s range of outcomes in redraft leagues?
Stacking Data for 20+ Years
To try to answer these questions, I considered QB, RB, and WR data from nflfastr going back to 1999. I removed RBs and WRs that scored zero points on the week and QBs that scored less than seven points (all in Half-PPR) since these are usually backups or other fantasy irrelevant options.
To begin, I grabbed the weekly performances of stacked QBs and WRs; I only included the top two WRs on each team. Here are the results:
Let’s unpack this chart, starting with the black line. This line just gives us the distribution of outcomes for the total score of an un-stacked QB and WR on your roster (QB and WR on different NFL teams). You can see that this distribution centers around about 27 points. I’ve shaded the ‘Boom’ area green for anything greater than 40 points, and the ‘Bust’ area red for anything less than 20 points.
The blue line is similar: it’s the distribution of outcomes for a stacked QB and WR. All of these scores are the sum of QBs and WRs on the same NFL team. As an aside, the best stack in this dataset was Week 5 of 2019, when Deshaun Watson tossed three TDs and over 200 yards to Will ‘The Flying V’ Fuller.
What do you notice about these two lines? Well, the blue line is underneath the black line in the middle of the chart, but higher than the black line in the Boom and Bust zones. Simply put, stacked QBs and WRs have a higher probability of booming and busting than non-stacked QBs and WRs.
This is intuitive – it’s what we were expecting – but the question remains: how big is this effect? Ultimately, in this dataset, a stacked QB/WR combo has a 15.3% chance of booming vs. a 12.0% chance for a non-stacked QB/WR combo. Similarly, there is a 24.0% chance of busting for stacked vs. an 18.8% chance of busting for non-stacked. It might not seem like much, but that difference can add up over a season; in addition, fantasy football is about finding every edge you possibly can!
Employing the Stack
How can you use this information? Decisions to stack and unstack will depend entirely on your team and situation. Imagine that you’re taking a flyer on a waiver wire QB heading into the Sunday afternoon slate of games. If at that point, your fantasy squad is down significantly, it might make sense to find a similarly ranked QB that stacks with a WR on your roster.
This will give you a higher chance of booming to overcome the deficit. It will, of course, also increase your probability of busting, but who cares: you were losing anyways! By the same token, if you have a comfortable lead heading into the final group of games for the week, it could be smart to avoid a stack and lower your probability of a disastrous outcome.
Perhaps more importantly, you might find yourself with the opportunity to stack on draft day. Say you picked up Calvin Ridley in the second round and, near the end of the draft, have the option to pick Ryan Tannehill or Matt Ryan as your QB. Of course, Matt Ryan will complete the stack with Calvin Ridley; is that sort of boom/bust profile something you want? Are you feeling confident about the rest of your team outside of these two players and want to go for it, or did you call your shot on some RB sleepers and want to stay lower risk with your QB and WR combo?
Wait… There’s More!
If you’re intrigued by the strategy of stacking, good news: there’s an even more aggressive stack possible. Here is the same chart as above, but for the top two scoring WRs on an NFL team:
The probability of ‘booming’ (defined here by scoring 35 points or more) is 11.1% for stacked teammate WRs and 7.6% for non-stacked WRs. Likewise, busting (10 total points or less) has a 9.7% chance for stacked WRs and a 4.6% for non-stacked WRs.
The discrepancies between stacked and non-stacked WR pairings are pretty large, which makes sense. If ‘the pie’ is big – the offense throws for lots of yards and touchdowns – everyone gets to eat plenty. Similarly, if the QB has a rough day, all of the wideouts go hungry. Even though WRs are ‘splitting the pie’ – competing for targets – the size of the pie is more important, and the result is a positive correlation.
Stacking WRs isn’t as common in fantasy but can be deployed just like we saw above. Really want to go for it in your draft? Try grabbing Cole ‘The Beasles’ Beasley in the later rounds to supplement your early Stefon Diggs pick. Down by a bunch heading into Monday Night Football? Consider rostering two real-life teammate WRs – within reason, it is of course still important to start players with solid projections – and hope that they ‘bring the boom.’
We also see this effect, albeit a bit smaller, with QBs on opposing teams. Two QBs playing against each other in real life have a 13.2% chance of booming together in fantasy (45 total points or more) vs. 11.8% for non-opponent QBs. Similarly, the bust probabilities (less than 20 total points) are 6.4% and 5.3%. This makes sense: QBs facing each other in shootouts will likely both put up big numbers and vice versa for defensive slogs. Identifying these relationships can be valuable in a Superflex league, and potentially crucial when you are taking a flyer for a week in a single-QB league (want to try to ‘cancel out’ your opponent’s stud QB? Might be a good idea to stream the opposing signal-caller).
It’s easy to forget that fantasy football is still a ‘team sport’ in a sense. It’s not just the individual players on your roster: context matters, and the whole of your team can be greater than the sum of its parts. Even if, on average, stacked players will score the same amount as unstacked players, leveraging the boom/bust profile in specific situations can be the difference between a win and a loss. Study up on stacking to give yourself the best possible chance of victory week in and week out.