On DraftKings, you can roster up to five batters from the same team — a strategy known as stacking. On FanDuel, you can stack up to four batters. Over the past couple of years, stacking has become a popular strategy in guaranteed prize pools (GPPs), and it’s very common for the winner of a DraftKings tournament to have rostered four or five players from the team that scored the most runs in the slate. Take the massive 15-game slate from August 8. The winner of the MLB $500,000 Mega Full Count (10,330 entries) had four Twins:
In the bigger MLB $130,000 Four-Seamer (38,235 entries), the winner had five Twins:
Stacking is seemingly popular and effective in GPPs, and for good reason. I’ve written a lot about stacking in the last couple of months; here’s a passage from my DraftKings arcade mode MLB strategy guide:
Stacking is popular because of the upside it gives your lineup — if a team goes off for 15 runs, then your chances of accumulating a lot of points are higher when you’ve stacked the team — but, honestly, stacking is also important because it limits your downside. The odds of hitting on every batter in a lineup are small; if you’re taking batters from different games with different pitchers and different dynamics, the odds get smaller. Stacking eliminates variables and correlates outcomes. All you need is one team to go off, and you will have five of your spots perform well. Stacking gives you upside, but perhaps more importantly it decreases downside.
That said, a strategy is only as effective as the public allows it to be. There could be a ton of value in stacking the top five batters of the Colorado Rockies when they’re home at Coors Field, but that value diminishes quickly if everyone in the tournament does so. In daily fantasy sports, sports betting, or just life, there can be some value in fading the public. So it’s a fair question to ask: If stacking is so valuable, have we reached the point where it makes sense to fade the stacking public in GPPs? To answer that question, we need to know how many people actually fully stack in GPPs. That has been a difficult question — until now. With our new DFS Contests Dashboard, which allows Pro subscribers to query the ownership of any players together, we can analyze the exposure rates of various groups of players, including stacks.
So let’s look at a large slate — that 15-game main slate on Aug. 8 — and see how highly owned each top-of-the order five-man stack was. Here’s the data: The green bars show a team’s implied run total at game time, and the red line shows the combined ownership of each team’s 1-5 batters in the Mega Full Count:
No, you’re not seeing incorrect data. Indeed, the combined ownership of the highest-owned 1-5 stack was just 0.66 percent — or 68 lineups. The Houston Astros had the highest implied total at 6.0 runs, and their 1-5 batters were owned together in just 0.35 percent or 36 total lineups. The Minnesota Twins — the winning stack due to their slate-high 11 runs against the Brewers — were stacked 1-5 in just 0.09 percent of lineups. The Washington Nationals, who are second this season with a .341 team wOBA and were fourth on the slate with 5.2 implied runs, had ZERO percent ownership of their 1-5 stack in the Mega Full Count. They scored just three runs against the Marlins, but zero percent ownership for one of the most powerful offenses is notable.
That said, baseball managers don’t always put their five best players at the top of the order, and DFS players don’t always want to stack the top five batters: For instance, Joey Gallo of the Texas Rangers has hit in the bottom third of the order for most of the season. So what does the combine ownership look like not for the top of each lineup but for the five batters on each team with the highest ownership rates?
Not much changes. The highest-owned five-man stack was still the Mets at 0.66 percent. Most teams’ five highest-owned guys weren’t rostered together at a high rate. In this data, I find two main takeaways:
- DFS players overestimate the extent to which everyone else will be chalky.
- DFS players overestimate the extent to which everyone else will stack.
On the first point: Before we discuss how low all the stacks are in terms of combined ownership, it should be noted that the ownership is relatively equal in its distribution among teams. Part of that is likely because there are so many choices. With 30 teams in the slate, a team’s ownership can get only so high, even if it’s the chalk. Even so, there should still be a mostly linear relationship between implied run total and stack ownership. That relationship, at least on this slate, is almost nonexistent, which you can see for yourself in the charts above. Additionally, the correlation coefficient between implied run total and the ownership of 1-5 stacks is 0.1738.
As a result, there is likely an edge to be had in GPPs: The Twins, who were implied for 5.5 runs — the second-highest total in the slate — had their 1-5 batters owned together in less than 0.10 percent of lineups. It’s important to be contrarian in tournaments, but it seems that rostering even a should-be chalky-ish five-man stack accomplishes that goal. You don’t have to dip down to low-upside offenses to benefit from low ownership.
On the second point: The highest combined ownership of any five-man stack was only 0.66 percent — the top five batters from the New York Mets. Some teams did not have even one five-man stack with the top of the order or the highest-owned batters — including the Washington Nationals, who were implied for a whopping 5.2 runs. Why is this the case? My first theory, which I’ll look at next week, is that perhaps DFS players are stacking four batters instead of five. If that’s true, that could provide a huge edge for people who stack five batters in GPPs, both in terms of upside and the probability of reaching that upside. This data from the new DFS Contests Dashboard is surprising and deserves further investigation. We’re always looking for edges in large-field tournaments. Part 2 is coming soon.