How it works
CS2 Lab rates CS2 Kills and Headshots props for Maps 1-2 with a statistical model, then tracks every pick in public so you can judge it.
Every PrizePicks, Underdog and Sleeper CS2 Kills / Headshots line is matched to the player's last ~20 series on Maps 1-2: kills, maps, rounds, overtime, blowouts, LAN, opponent and the map veto.
Expected rounds on Maps 1-2 × his kill rate on those maps, adjusted for how close the match should be — every adjustment learned from real results.
Maps 1-2 are simulated thousands of ways (stomps, close maps, overtime, good and bad nights) to get the chance he clears the line, then trimmed and capped so it stays honest.
A play only passes with enough edge, positive EV at the app's real payout, recent form that agrees, a fitting game script and a big enough sample. Stand-ins are never a Take.
Two passing plays on the same app, from different matches that haven't started, ranked by EV. Unproven Unders and adjusted lines stay out until results prove them.
Every result is graded from official stats. Each night the model checks itself on days it hasn't seen and only keeps changes that make it more accurate.
Every pick is graded from official match stats, wins and losses alike. Past results don't guarantee future ones.
