Player ratings and market edges.
For Counter-Strike DFS pick'em and Kalshi prediction markets.
CARNAGE
Every relevant player on the slate, rated 35 to 99. Blends form, opponent, win probability, and BO format. Calibrated for DFS ceilings.
MARKET
Every CS2 game with the model's fair value next to live Kalshi. The gap is the edge.
The Methodology
Three layers. Every projection runs through all of them.
Player Skill Calibration
Every CS2 player carries a Skill Calibration score from continuous form tracking. The score updates after every game, weighting recent performance over historical baseline. Role context built in.
Opponent & Game Context
A player's projection shifts with opponent strength, implied win probability, and BO format. The same player against Vitality projects differently than against a tier-3 roster.
Market Fair Value
Model fair value gets compared against live Kalshi prices. Gaps above the noise floor flag as edges. Below it, signal goes ignored.
Stats describe. Analytics decide.
Most CS2 data tells you what already happened. Season averages. Historical baselines. The shape of a player across months.
Game of Skill projects. Forward-looking numbers calibrated for DFS ceilings and Kalshi trading. Not who was good this year. What to expect from this player tonight, against this opponent, at this price.
Built by an analyst with eight years running fantasy sports analytics at RotoGrinders, applied to Counter-Strike.