Tennis Predictions
Model win probabilities for ATP & WTA matches, with a confidence rating per pick and a public accuracy record over every settled match.
68Upcoming
71.9%Accuracy
4697Settled
Iasi, Romania
clay
✓ Correct
High · 68%
vs
2–1
Braunschweig, Germany
clay
✓ Correct
High · 71%
vs
2–0
Winston-Salem W01
hard
✓ Correct
High · 73%
vs
1–2
Winston-Salem W01
hard
✓ Correct
High · 68%
vs
2–0
Winston-Salem M01
hard
✓ Correct
High · 67%
vs
2–1
Contrexeville, France
clay
✗ Missed
High · 83%
vs
2–0
ATP Challenger Nottingham 3, Great Britain Men Singles
grass
✓ Correct
High · 72%
vs
1–2
Wimbledon, Men
grass
✓ Correct
High · 89%
vs
3–0
Bastad, Sweden
clay
✓ Correct
High · 78%
vs
0–2
Iasi, Romania
clay
✓ Correct
High · 82%
vs
0–2
Trieste, Italy
clay
✓ Correct
High · 70%
vs
0–2
Braunschweig, Germany
clay
✓ Correct
High · 68%
vs
0–2
ATP Challenger Liege, Belgium Men Singles
clay
✓ Correct
High · 77%
vs
1–2
Trieste, Italy
clay
✓ Correct
High · 75%
vs
0–1
Iasi, Romania
clay
✓ Correct
High · 79%
vs
2–1
Iasi, Romania
clay
✗ Missed
High · 72%
vs
1–2
Iasi, Romania
clay
✓ Correct
High · 81%
vs
2–0
Bastad, Sweden
clay
✓ Correct
High · 69%
vs
2–0
Bastad, Sweden
clay
✓ Correct
High · 72%
vs
0–2
Trieste, Italy
clay
✓ Correct
High · 72%
vs
0–2
Trieste, Italy
clay
✓ Correct
High · 78%
vs
2–1
Braunschweig, Germany
clay
✗ Missed
High · 83%
vs
0–2
Braunschweig, Germany
clay
✓ Correct
High · 75%
vs
2–0
ATP Challenger Liege, Belgium Men Singles
clay
✗ Missed
High · 69%
vs
1–2
Contrexeville, France
clay
✓ Correct
High · 74%
vs
2–0
Contrexeville, France
clay
✓ Correct
High · 78%
vs
0–2
Contrexeville, France
clay
✓ Correct
High · 84%
vs
0–2
Braunschweig, Germany
clay
✓ Correct
High · 72%
vs
0–2
Trieste, Italy
clay
✓ Correct
High · 70%
vs
2–0
Bogota, Colombia
clay
✓ Correct
High · 84%
vs
2–1
⚡ This page, via API Pro
Everything on this page is available as JSON via the Tennis API (Sports Addon, $5/mo):
GET
/tennis/api/v2/predictions/?upcoming=true
How to read this: the bar shows our model's win probability for each player; the confidence pill rates how much signal the model had for this match-up. Accuracy is measured over every settled prediction — not a curated subset. Not betting advice; no outcome is guaranteed.