A Beginner’s Approach to DraftKings CSGO Picks and Strategy
As the COVID-19 quarantine drags on day after day I have found myself looking for additional ways to entertain myself, get that degen fix, and test my brain in a problem-solving competition for real money. DFS industry friend Adam Daly turned me onto Counter-Strike: Global Offensive, or CSGO, a first-person shooter game that pits a team of terrorists against counter-terrorists on a variety of different maps. Each matchup is a best of three contest on up to three different maps with a first to 16 outcome on that map.
Do I understand the game? No, not really. At best marginally.
Has it been a fun sweat? Hell yes.
Have I been lucky enough to bink? You betcha.
On DraftKings, the game features points for kills, assists, and negative points for deaths, in addition to a variety of bonuses for “clutches” which are kills when you’re outnumbered. DraftKings also features a tilting game-level adjustment (GLA). Since some games only go two maps when there is a sweep, the GLA adjust games to reflect the same number of maps played.
On the recently launched Fanduel game, the scoring has been adjusted to add additional team points for one of the primary objectives, bombs defused and planted, as well as rounds lost and won.
I am still teasing this out a bit but there are some correlations:
Kills / Assists / Deaths: Both platforms featuring scoring for kills and assists and deaths. For every kill, there is obviously a death for an opposing team player as a big piece of negative correlation. In addition, some kills are assisted by team members which bring in some same team correlation. This correlation exists, but is small, as an upside kill game may be 80+ kills (what up Niko) while an upside assist game may be 15+ assists.
Last for newbies like myself, this “assist” is not like an assist in the NBA or NHL or Soccer. It isn’t somebody passing it off, but rather two people shooting the same player and injuring him before one person grabs the kill. Because of this, you’ll often see players with high kill and assist totals because they are always involved in that role.
Anyways, I have over-tilted kills or kills per round in my projections to try to target players that have the upside in the scoring system.
GLA: The game adjustment increases the value of 2-0 blowouts with same team correlations.
Should we Stack?
I’m still adjusting here. One of my GPP wins featured a 3-2-1 roster construction while the other had a 1-1-1-1-1-1 lineup with no players from the same team. There is a finite number of points available in a given game and for a given team, and some players will have a high market share of that team’s kills.
My gut right now is that on 3-game slates it is okay to run (but no need to FORCE) 3-man “stacks” if they happen to project well naturally. On a 6-game slate, I think its less likely that one team can put up 3 of the top 6 scorers on the slate, and you can restrict to two players per team with maybe one or two exceptions.
Similarly, despite the negative correlations, allowing players on opposing teams of the same game can work if the game goes three maps and those players command a high market share of the team’s offense. My 1-1-1-1-1-1 win contained DAPR of FunPlus Phoenix vs Steel of Chaos, as well as MarKE of Swole Patrol vs yuurih of Furia.
TLDR: I think it is bad to include opponents vs a 3-man onslaught but you can include opponents in rosters that have 2-man stacks or one-offs.
Fanduel Team Correlations: Fanduel definitely looks more stack friendly with heavy team scoring elements. My initial inclination is that running 3 or 2-man mini stacks would be an optimal approach.
April 9th Slate Thoughts
MAD Lions are a -370 favorite versus FunPlus Phoenix and the most likely team to win 2-0 with a huge GLA. They are the only team I would consider 3-man stacks for on this slate, though I likely will restrict it to 2-man max given the slate size. acoR is my top value on the slate at $5800 and I have him projected as the 3rd highest scorer on the team. sjuush and Bubzkji are the top-end options and sjuush has the highest kill rate on the team so far this season. One thing I don’t know how to reconcile (being a noob and all) is that over the last three months acor, sjuush and bubzkji have similar kill rates, but acrR’s is way down (21.6 -> 17.1) to start the newest season. A change in his role / strategy or some variance we can capitalize on? Winning teams today will likely have at least one MAD Lion.
FAZE is the other big favorite on today’s slate and NiKo should be one of the highest owned players. He is the 2nd highest projected player for me (behind s1mple) and is #1 overall if you simply sorted the DraftKings data by average fantasy points per game. NiKo is averaging more than 24 kills per 30 to start the season and leads his team in assists as well. He has three sub 60-point games and four games where he’s averaging more than 100-points. Along with NiKo, coldzera is my 2nd projected player on the Faze team and at $6200 he’s one of the stronger players on the slate. He probably doesn’t have 100+ upside and may not make sense as a captain but he’s put up 70+ in onslaughts and correlated well with NiKo in those matchups.
Other Top Options: s1mple, ropz, KSCERATO
Other Top Values: electronic, dapr, KRIMZ
My Five Highest Owned: s1mple, sjuush, acoR, NiKo, ropz