MODEL PREDICTION · 2026-09-24
● HARD

J. Duckworth vs L. Sonego — prediction

Chengdu
✗ Missed
DUCKWORTHWIN PROBABILITYSONEGO
55%
model prob.
@2.13
odds · 47% impl.
⚔H2H 1–2 Duckworth⏱Rest 5d vs 16d◐Hard 47%🎾Serve 65%📈Form 7/10 · 2✗
CONDITIONS OF THE MATCH◆ in the model◇ context
Surface◆
Hard

Consistent bounce, medium-fast: neutral conditions, no style favored.

Temperature◇
25°C

Warm: the ball flies a little more and fitness counts.

Humidity◇
73%

Humid air: the ball loses some speed.

Wind◇
6 km/h

Light wind: no noticeable effect.

Altitude◇
500 m

Moderate altitude: the ball flies a little more than at sea level.

Surface feeds the model (surface specialization is one of its factors). Weather and altitude are context we publish for you — they do NOT move the probability.

OUR MODEL'S REASONING

›Ranking: #65 vs #86 (better ranked)

›Head-to-head: 1-2 against

›Model 55% vs market 47% → the model sees it as MORE likely than the odds

›Recent form: 7/10 in recent matches

WATCH FOR

!Unfavorable head-to-head record (1-2)

Calibrated model probability (~65% out-of-sample accuracy). Not a guarantee: the model ≈ the market on average, so the odds already capture almost all the edge. 18+ · gamble responsibly.
@1.82
fair odds
+16.9%
expected value
BREAKDOWN · MODEL ANALYSIS

The model makes J. Duckworth the favorite with a 55% win probability, against L. Sonego's 45% — a tight match, without a wide margin. Converted to odds, that probability is worth about @1.82; the offered odds are around @2.13 (a 47% implied), slightly above the market, so the model is a touch more optimistic.

Several factors explain the number: #65 vs #86 (better ranked); 1-2 against; model 55% vs market 47% → the model sees it as more likely than the odds; 7/10 in recent matches.

Read it with perspective. Our probability is calibrated — when the model says 55%, that outcome happens roughly that percentage of the time, with ~65% out-of-sample accuracy — but being the favorite is not being the winner: roughly 45 out of every 100 times Sonego wins. The model also tends to agree with the market, so the odds already capture almost all the edge: don't take it as a sure value. Watch out for: Unfavorable head-to-head record (1-2). This is informational analysis, not a betting recommendation. 18+ · play responsibly.

Analyze today's matches →