Volt Projections
A full esports projection model for CS2, League of Legends, Valorant and Dota 2 — every line priced, every lineup built, every pick graded.
Data Driven
Model Focused
Results Proven
A full esports projection model for CS2, League of Legends, Valorant and Dota 2 — every line priced, every lineup built, every pick graded.
Four things separate a projection model from a pick sheet. Volt is built on all four, and the fourth is the one nobody else shows you.
Multiple data sources. Real time updates.
Custom projections. Built to find value.
High confidence picks. Long term profitable.
Every pick tracked. Transparent performance.
The model runs continuously. Lines move, projections recompute, lineups rebuild.
Prop lines ingest continuously from every supported book, each movement time-stamped so opening and closing numbers are both preserved.
Projections compute on arrival against the player, map and team databases — form windows, roster changes, map pool and matchup context all feed the number.
Correlation-aware construction searches the slate for combinations that clear the confidence threshold, instead of stacking legs independently.
Kills and headshots projected per player per map, priced against the book line, with the discrepancy surfaced rather than hidden.
| Player | Matchup | Map | Market | Line | Projection | Edge | Lean |
|---|---|---|---|---|---|---|---|
| ropz | FaZe vs NAVI | Mirage | Kills | 17.5 | 19.2 | +1.7 | OVER |
| donk | Spirit vs G2 | Inferno | Kills | 20.5 | 22.8 | +2.3 | OVER |
| m0NESY | G2 vs Spirit | Inferno | Headshots | 9.5 | 8.1 | -1.4 | UNDER |
| ZywOo | Vitality vs MOUZ | Ancient | Kills | 19.5 | 21 | +1.5 | OVER |
| NiKo | Falcons vs Astralis | Nuke | Headshots | 8.5 | 9.6 | +1.1 | OVER |
Illustrative layout. The live board renders from the model API.
Stacking legs independently overstates a lineup’s chances. Volt scores each build against measured correlation between players, maps and matchups — so the number on the slip is the number that matters.
Lineups rebuild as lines move, not once at lock.
Pairwise relationships measured from graded history.
Nothing publishes below the configured threshold.
PrizePicks, Underdog, Sleeper, Betr.
Every projection is settled against the actual result — wins, losses and pushes alike. Opening and closing line value are both recorded, so performance can be judged on process and not only on outcome.
Volt publishes a graded record, not a highlight reel. No performance figure appears on this site until the model behind it is live and its results have been graded in full. This panel turns on with the record, not before it.
The model is theirs. The pipeline, the infrastructure and the grading are built around it so the numbers you see are the numbers it produced.