Model probability
REPICOM's production estimate that a selection wins, built only from data available before the event. It is an estimate, not a certainty.
Resources
How to read REPICOM
The methods behind REPICOM's research, how to interpret probabilities and sportsbook odds, and exactly what a qualification decision means.
Research method
From real data to a decision
- 01
Pregame data
Player and team history, consistency (MCP), matchup evidence and captured market prices — only what existed before the event.
- 02
Model vs market
A validated production model estimates a probability; REPICOM compares it with the captured price.
- 03
Qualification gates
Confidence, edge, expected value and minimum historical samples must all pass. Thresholds never move to create content.
- 04
Locked, then verified
Qualified plays are frozen before the event, released by publication rules, and settled from verified results — wins and losses alike.
Reading the numbers
What each number means
Market implied probability
What the sportsbook price says, converted to a percentage. It includes the book's margin, so both sides of a market add up to more than 100%.
Model − market gap
The difference between the two, in percentage points. A gap alone never makes a play — REPICOM also requires confidence, positive EV and a sufficient sample.
Expected value (EV)
Average return per unit staked if the probability is right. EV is only as good as the probability behind it.
Historical sample
How many past games support an evaluation. Small samples are noisy; REPICOM shows the sample next to every number.
MCP consistency
Most Consistent Player: repeatability of a player's performance across games. Consistency is context, not the probability of a wager.
Educational tool
Odds & EV calculator
Convert a sportsbook price into implied probability and see what your own estimate would mean.
- Market implied probability
- 52.38%
- Decimal odds
- 1.909
- Fair price for your estimate
- -122
- Probability − implied
- +2.62 pts
- Expected value per 1 unit
- +0.050 u
Implied probability includes the sportsbook margin. A positive EV only exists if your probability estimate is correct — which is exactly what REPICOM's qualification gates and sample-size rules are there to test.
Decisions
Qualified, locked, NO PLAY
Qualified
Every production gate passed. Members see it before the event; it becomes public when its publication state allows.
Locked
Qualified but not yet publicly released. Public pages hide the side, line and numbers until release.
NO PLAY
Evaluated and rejected. The reason is published. A NO PLAY row can still show a statistical signal — it is research, not a pick.
Unavailable
REPICOM does not have verified data for that field yet. It is never estimated or filled in.
Results math
How the record is calculated
Hit rate = wins ÷ (wins + losses). Pushes and pending results are excluded and shown separately.
Units assume a flat 1 unit per published play. A win earns (payout ÷ stake − 1) units, or the price-implied profit when no payout is recorded; a loss costs 1 unit.
ROI = net profit ÷ total staked, using only rows with a recorded stake.
Every figure is computed from the public proof ledger on the Results page, so anyone can re-check it.
FAQ
Common questions
Why do some games show NO PLAY with a model lean?
REPICOM publishes its diagnosis for evaluated games. A lean that fails the gates is information, not a recommendation.
Why are player photos or team logos missing?
REPICOM only shows images it has verified rights to use. Until then you'll see REPICOM monograms and league abbreviations.
Can I see game logs or home/away splits?
Only when REPICOM has a validated feed for them. Where it doesn't, the page says so instead of estimating.
Does REPICOM place bets?
No. REPICOM provides sports intelligence and analysis. It does not place wagers, hold funds or access sportsbook accounts.