#1Selected
Natasha Mack
Phoenix Mercury vs Toronto Tempo
UNDER
PTS 10.5
Projection
8.31
Model
68.0%
Gap
-2.19
Price
-125
projection 8.31 vs line 10.50 / under probability 68.0% / gap -2.19 (0.39 sigma)
W
ULTOPS / WNBA
2 picks for 2026-08-29
FanDuel Live Mode: EXPERIMENTALAug 29, 2:41 PM ET
WNBA Prop Card
WNBA Correlation-Aware Player Prop Model V1 ranks the current slate by model probability, projection gap, book price, and portfolio risk. FanDuel-live mode is the primary card path, with unavailable props removed before model selection.
Selected
2
3 board rows
Avg prob.
67.0%
1 A / 1 B
SGP risk
MEDIUM
97.5% est. multiplier
Settlement
Pending
2 pending
Settlement Tracker
0 settled / 2 pending
Wins
0
Losses
0
Accuracy
Pending
#1 Natasha Mack
UNDER PTS 10.5
Actual n/aFanDuel
#2 Isabelle Harrison
UNDER P+R 22.5
Actual n/aFanDuel
Today's Picks
Ranked one per player, highest model score first.
2 selected
#2Selected
Isabelle Harrison
Toronto Tempo vs Phoenix Mercury
UNDER
P+R 22.5
Projection
18.79
Model
66.0%
Gap
-3.71
Price
-118
projection 18.79 vs line 22.50 / under probability 66.0% / gap -3.71 (0.40 sigma)
Next Up
Strong rows held by portfolio or slate gates.
CCandidate
Isabelle Harrison
Toronto Tempo vs Phoenix Mercury
UNDER
PTS 14.5
Projection
11.72
Model
63.4%
Gap
-2.77
Price
-114
Board Explorer
2 of 3 rows shown by current filters
Natasha Mack
SelectedPhoenix Mercury vs Toronto Tempo
Model lean
UNDER PTS 10.5
Model
68.0% / 0.780
Gap
-2.19
Listed
UNDER PTS -125
FanDuel
Single Side Price / Thin Market Count
Isabelle Harrison
SelectedToronto Tempo vs Phoenix Mercury
Model lean
UNDER P+R 22.5
Model
66.0% / 0.685
Gap
-3.71
Listed
UNDER P+R -118
FanDuel
Combo Market Correlation / Short Rest Or B2b
Model Stack
Projection engine
Blends player per-minute form, EWMA, last-3/last-10 production, season baseline, position fallback, opponent allowance, home/away splits, and expected minutes.
Probability layer
Converts projection gap into an over/under probability with market-specific residuals and a recency-weighted empirical hit-rate blend.
Price edge
Normalizes supplied American odds to no-vig fair probability, then compares book price against model probability.
Portfolio gates
Limits exposure by player, team, game, market, combo market, and same-team counting overs, then applies live-mode availability gates and SGP exposure accounting.
Latest Selector Proof
Six-pick coverage
41/41
Archive-ML walk-forward covered every evaluated slate
Settled parlays
14/27
51.85% archive/cross-source six-pick parlay hit rate
Settled legs
175/226
77.43% archive selected-leg accuracy
FanDuel shadow
3/5
60.00% limited FanDuel availability replay; sample is too small to drive the selector
Portfolio Gates
Max 6 picks
Max 1 per player per slate
FanDuel-live mode is the primary six-pick path for the website card
Archive-ML proof does not transfer to single-book availability by default
Unavailable user-confirmed FanDuel props are removed before model selection
Soft SGP tax lowers priority for additional same-game legs in FanDuel-live mode
Max 6 per team in live mode
Max 6 per game in live mode
Max 4 per market
Max 4 combo markets
Max 1 same-team counting over
Requires target-date ESPN slate and current ESPN roster match
FanDuel rows require playable side odds and current source context
Live forced fill requires at least 52% model probability
Archive-ML rerank uses only prior archived candidate outcomes
Current archive proof: 41/41 covered, 14/27 settled six-pick parlays, 175/226 legs
Limited FanDuel replay: 3/5 settled six-pick parlays, 73.33% legs
Data Contract
game_dateplayerteamopponentmarketlineover_oddsunder_oddssportsbook_countprojected_minutesstarter_expectedinjury_notesource_picksource_projectionsource_url
WNBA V1 ranks sourced player props with historical boxscore evidence, current FanDuel-labeled prices, source projection alignment, and portfolio gates. Archive proof and live FanDuel execution are separated; live use still requires player availability and book availability confirmation.