Part 1: GPP Lineup Construction in 4+ Game Slates
Having good projections is only half the battle in the 2020 landscape DFS. Even with the best projections, which I know you have since you are reading this on TheEsportsDepartment.com, winning GPPs is not as simple as plugging those projections into an optimizer and clicking go. If it was everyone would be doing it. Lineup construction is often equally as important, if not more so, in creating GPP winning lineups.
Before we dive in I want to make it very clear that I am only referring to large field, multi-entry, GPPs in this article. All of my data is taken from the big $10/12/15 tournament of the slate. Also, everything in this article is talking about how to win 1st place as often as possible, not get a min cash. Tournaments these days are super top heavy, often giving out 20, 25, even upwards of 30% of the prize pool to 1st place. If you are MME’ing you are looking for wins, not to cash 146/150 lineups all at min cash, as that’s next to impossible to do anyways. Even if you are entering only a few lineups, your goal should still be 1st place, so you want to give yourself the best possible shot at getting there.
One last thing I want to make clear. As the title indicates, this information is ONLY for slates that have 4+ games, which is what I consider a “large” slate given the landscape of CS:GO (if anyone at Draftkings reads these, please, please, please start doing more slates of combined tournaments so we aren’t stuck with 2 game slates all the time). Now that that’s out of the way let’s dive in.
From May 1st through July 7th, there were 30 Draftkings slates that contained 4 or more games. In the winning lineup from those 30 slates, the largest stack from one team has been 2 players 25 times, while it has been a 3 stack from a single team only 5 times. Only 16.7% of the time does a lineup with 3 players from the same team win 1st place.
If you look at how scoring works now in CS:GO, this makes perfect sense. There is a finite number of points a team can get, every kill Player A gets is a kill Player B can’t get. In a win, a team is typically going to score in the neighborhood of 300-350 fantasy points as a whole.
The average score for the winning lineup in the 26 Best of 3 slates since May (there were 4 bo1 slates) was just over 540. This comes out to needing about 83 non-captain points per player in your lineup, with the captain’s 83 becoming 124.5 after the 1.5x multiplier. Now of course you may have a 90 or 100 point outlier in your winning lineup, so let’s go with 75 as our target for a player to be in a potentially GPP winning lineup.
If a team scores 325 total points in their win to have 3 players reach that target of 75 points means:
Player A: 75 out of 325 points = 23%, only 3 percentage points more than an even split of 20% per player, not too bad. There are now 250 points left for the other 4 players
Player B: 75 out of 250 points = 30%, now up to 5 percentage points more than the even split of 25% per player, getting a little tougher but still not too bad. There are now 175 points left for the remaining 3 players.
Player C: 75 out of 175 points = 43%, now all the way up to about 10 percentage points more than the even split of 33% per player, this is tough to do.
If player A and/or B goes above 75, which you are looking for particularly in your captain spot, it becomes even harder for Player C to hit 75 points as there are less for everything else.
You’re probably wondering why I spelled out that out since the data already shows that a 2 man as the largest stack wins over 83% of the time. Well the reason I wanted to spell that out, is because far too many people are running 3 stacks on these slates. In the same 30 tournaments where a 3 man stack only won 16.7% of the time, 26.4% of lineups entered had at least 1 3 stack (many were run as a 3/3). The number of 3 stacks entered was 58% higher than the rate that they win at!
“Well Chris, the scoring system was still relatively new in May and people hadn’t quite figured out how to make optimal lineups yet…”
Wow that’s a good point, thanks Reader, let’s look at what happens when you take May out, assuming that players have gotten sharper as time has gone on.
19 of the 30 4+ game slates have occurred since June 1st. Of those 19 slates, 17 have been won by a lineup with the largest stack being a 2 stack, only 2 (two) have been won by a lineup with a 3 stack. That comes out to 10.53% of the time that a 3 stack wins. Yet, in those same tournaments, 23.99% of lineups have still contained a 3 stack. That’s over double the rate that they win at! At the same time, max 2 stacks are being used 24% under the rate that they win at.
Long story short: Stop playing 3 stacks in GPPs if there are 4 or more games on the slate.
3 stacks are vastly over-represented in the field, and it’s much harder for the 3rd fragger on one winning team reach a GPP winning score than it is for the 2nd fragger on another winning team, there simply aren’t enough points to go around for a team to reliably have 3 player reach GPP winnings scores. People are very used to playing 3 stacks on 2 or 3 game slates (we’ll get to how to win those later), and it appears many of them are using the same rules to build lineups on 4+ game slates. By limiting your lineups to a maximum of 2 players per team, you will increase your likelihood of winning a CS:GO GPP!