Model Center

v1.0-transparent — a weighted blend of named signals. No black box, no guarantees.

Model
Active version
v1.0-transparent
Every projection stores its version
Signals
6
3 listed but not yet wired in
Projections this week
0
Week 1
Bet-grade calls
0
0 leans · 0 passes
Versions in data
—
Loading

Signals and weights

The whole engine, in one table.

SignalWeight
Season baseline
Prior-season per-game production from the live stats feed. The anchor for everything else.
30%
Recent form
Last 5 games this season, weighted toward the most recent, to catch a changed role.
20%
Expected usageNot in v1
Snap share, route participation, target share and carry share. No live usage feed yet, so this does not move a v1 projection.
—
Opponent matchupNot in v1
Defensive efficiency, pressure rate and positional funnel. Not enough team game data synced yet, so it does not move a v1 projection.
—
Game environment
Posted game total against the week average, applied as a multiplier.
25%
Injury / role change
The player’s own designation on the live injury report. Teammate absences are not modelled yet.
10%
Home / away
Small venue effect, applied as a multiplier. Deliberately kept small.
2%
WeatherNot in v1
Game-time wind and precipitation are fetched and displayed, but are not yet an input to the number.
—
Market anchor
Pull toward the posted line, because the market prices information the model cannot see.
25%

Weights are fixed for v1 and applied identically to every player. Nothing is fitted after the fact. Signals marked “Not in v1” are shown on the site but do not change any projection yet.

Projection inspector

Pick a projection to see how the number was assembled.

No projections for this week

How a decision is reached

Edge, confidence and data quality are separate gates.

1. Edge

Model probability minus the probability implied by the price, in percentage points. Under 3 points is noise. Projection difference is reported separately because it answers a different question.

2. Confidence

Set by sample size and input completeness, never by the size of the edge. A large edge on thin data stays low confidence.

3. Decision

Bet needs 6+ points, high confidence and complete inputs. Lean is a real but modest edge. Pass is used freely and is not a failure.