Props audit trail — WNBA

Props ledger.
Different rules.

Last updated: 2026-09-07 10:00 AM ET  ·  WNBA only  ·  Deduplicated

WNBA player props (points, rebounds, assists) are tracked separately from the main Soccer & MLB audit trail. The pricing methodology is different, and we want subscribers to understand exactly what that means before drawing conclusions from these numbers.

Why this is a separate page: The main EdgeLeverage trail prices every pick by devigging the sharp consensus from six marketmaker books — Pinnacle, Prophet, Bookmaker, Sporttrade, LowVig, and BetOnline. That methodology produces a reliable fair-price benchmark.

WNBA props use a different source: Player prop markets for WNBA are priced off the SportsGameOdds (SGO) consensus — an aggregated market across multiple sportsbooks. There is no Pinnacle WNBA props market, so we cannot benchmark against the same sharp reference we use for soccer. The implied edge is real in the sense that it beats the consensus price, but it is not the same claim as "this beats Pinnacle's closed line."

CLV is not tracked the same way: Because there is no sharp closing line to compare against, closing line value is not reported for these picks. Win/loss outcome and ROI are the primary metrics here.

What this data does tell you: Whether the model-based selection is finding +EV opportunities in the consensus prop market. Over a large enough sample, a positive hit rate above the implied breakeven, or consistent ROI, would validate the approach.

Sample so far

Aggregate stats (1308 graded picks)

Graded picks

1308

627W – 681L – 0P

Hit rate

47.9%

ignoring pushes · sample too small

ROI (flat 25% Kelly)

+2.1%

no Pinnacle benchmark · sample too small

Avg Odds

+118

2.18 decimal · 45.9% implied

Day-by-day log

Raw ledger

DateRecordGames
Sep 01 8W–6L Connecticut Sun @ Dallas Wings
Aug 31 9W–11L Los Angeles Sparks @ Seattle Storm
Aug 30 10W–15L Toronto Tempo @ Phoenix Mercury
Aug 29 37W–55L Portland Fire @ Atlanta Dream
Aug 28 10W–9L Washington Mystics @ Phoenix Mercury
Aug 27 25W–20L Toronto Tempo @ Seattle Storm
Aug 26 38W–36L Golden State Valkyries @ Minnesota Lynx
Aug 25 3W–2L Washington Mystics @ Portland Fire
Aug 24 58W–65L Seattle Storm @ Dallas Wings
Aug 23 47W–40L Indiana Fever @ New York Liberty
Aug 22 63W–68L Las Vegas Aces @ New York Liberty
Aug 21 24W–18L Atlanta Dream @ Los Angeles Sparks
Aug 20 14W–24L Toronto Tempo @ Washington Mystics
Aug 19 26W–30L Dallas Wings @ Golden State Valkyries
Aug 17 9W–22L Indiana Fever @ Atlanta Dream
Aug 16 5W–8L Los Angeles Sparks @ Washington Mystics
Aug 15 20W–14L Dallas Wings @ Indiana Fever
Aug 14 5W–7L Atlanta Dream @ Connecticut Sun
Aug 13 17W–9L Minnesota Lynx @ Portland Fire
Aug 12 20W–25L New York Liberty @ Indiana Fever
Aug 11 6W–7L Golden State Valkyries @ Los Angeles Sparks
Aug 10 11W–2L Phoenix Mercury @ Washington Mystics
Aug 09 3W–2L Las Vegas Aces @ New York Liberty
Jul 20 13W–15L New York Liberty @ Las Vegas Aces
Jul 17 146W–171L Toronto Tempo @ Atlanta Dream

How to read this

Methodology notes

Pricing source

WNBA player prop odds are sourced via SportsGameOdds (SGO), which aggregates lines from multiple sportsbooks into a consensus market. The scanner identifies picks where the SGO implied probability is below what the Poisson player model projects, adjusted for pace, matchup, and recent form.

This is not the same as the Pinnacle devig methodology used in the soccer and spreads scanners. The consensus market is softer than Pinnacle's sharp line, so the implied edge is a different kind of claim.

Why no CLV

Closing line value requires a sharp, liquid closing line to compare against. For soccer totals and spreads, Pinnacle fills that role reliably. For WNBA player props, no equivalent sharp market exists, so CLV is not tracked. The primary validation metric here is long-run hit rate vs. implied probability and ROI vs. flat staking.

What this is and isn't

This is an honest record of a model-based overlay on a softer market. It is not a Pinnacle-benchmarked edge claim. The grading is transparent, the timestamps are pre-game, and the methodology is disclosed in full. Draw your own conclusions — and watch the sample grow before drawing any.