Free during beta —Free during beta — to track favorite players + get game alertsto track favorites + alerts
Cross-sport·Model edges
Where the model
disagrees most tonight.
Every modeled league's clearest pick, priced as expected value at the current line. The headline names the side the model actually wants, then shows the EV% at that price.
Slate
13
games priced
Edges
12
surfaced
Top EV
+44.0%
MLB
Model · last 30
61.4%
221/360
Model accuracy · cross-sport
The receipts on tonight's leaderboard
Last 7
66.7%
96 games
Last 30
63.5%
416 games
Last 90
61.9%
1415 games
All time
+0.2vs mkt
1839/21571 priced · raw 66.8%
By confidence tier · all-time
★ Locks
80.8%
6698g
+20.8
Edges
67.8%
6123g
+12.8
Leans
57.4%
4810g
+5.4
Tossups
52.7%
3940g
+2.7
What does EV mean?
The percentage edge the model has over the posted price.
What does EV mean?
The percentage edge the model has over the posted price.
EV = Expected Value. A bet's long-run profit per unit risked, expressed as a percentage of the stake. It compares the model's win probability against the implied probability of the sportsbook's price.
+5% EV means $5 expected profit per $100 risked. Over a season, +EV bets sustained at the closing line are the bets that beat the market.
-EV means the price is shorter than the model thinks the side deserves; we don't surface those. Picks under +1.5% EV land in the noise floor and read as a pass.
Math: EV = (modelProb × payout) − (1 − modelProb) · payout from American odds at the posted line.
Last 14 days · edge ledger
How the surfaced edges actually played.
1 unit flat at the first available open moneyline · skipped rows when no usable open price was logged · pushes count as graded·0u.
Net units
+0.24u
ROI
+4.1%
Hit rate
50.0%
3/6
Sample
6
graded edges
2026-07-252026-08-07
Calibration · the receipt nobody else shows
Are the model's probabilities honest.
When the model says 60%, does the team actually win 60% of the time? Each row plots predicted probability against observed win rate across 2 leagues · 1,365 games · last 90 days. Closer to the dashed line is better.
-0.80pp
avg ECE drop · raw → calibrated
MLB
1,139 games
ECE2.21%→0.63%↓ 72%
biggest gap · over-priced 20%-30% · 11.1pp
0.63%
ECE cal
WNBA
226 games
ECE9.38%→9.37%
biggest gap · over-priced 20%-30% · 18.5pp
9.37%
ECE cal
ECE = expected calibration error · weighted avg gap between predicted prob and observed rate across deciles. Lower is better. Calibrated values use a Platt scaler refit on the rolling 90-day window.
EV stack · by league
How tonight's EV breaks down per sport
EV tiers · pass → sharp
<2%2–5%5–10%10%+
MLB
10 edgesTop 54.7% EV
3
3
4
WNBA
3 edgesTop 30.6% EV
1
2
Bar length scales with the league's edge count tonight; the longest bar belongs to the busiest slate. Segments inside each bar show how that league's edges break across the magnitude tiers — a league heavy on the cyan/emerald end is seeing meaningful price value, not just noise.
Slate distribution · cross-sport
Where the EV sits
5 sharp · 3 meaningful · 12 total
0
4
3
1
1
3
0–2
2–5
5–10
10–15
15–20
20+
Noise
Soft
Meaningful
Sharp
Top 3 · with the receipts
Why this edge.
Per-edge contributor breakdown · closing-line value · open → close prices.
Tonight · ranked
Top 12 cross-sport edges
Sorted by EV% across moneylines, ATS, and props
EV filter
Picks under +1.5% EV are passes; meaningful model value tends to live at +5% EV or better at the posted price. Lines refresh continuously from the odds snapshot ingest. New leagues fall into this leaderboard automatically as their model adapters ship.
Updated 10:24 AM ET · Lines refresh every 30 minutes; tomorrow's slate appears as the books post it.
How edges are computed
Each game gets a model win probability from our per-sport Elo derivative. We name the side the model actually picks, then price that pick as expected value against the posted American odds. The leaderboard merges across every league with a model running so you see the sharpest decisions across the night, not just within one sport.
Bet responsibly · 21+ · 1-800-GAMBLER · Models inform, they don't guarantee.