MODEL PREDICTION · 2026-07-22

A. Bublik vs F. Diaz Acostaprediction

✓ Correct
BUBLIKWIN PROBABILITYACOSTA
78%
model prob.
@1.54
odds · 65% impl.
🌡22° · 33% hum760 m altitudeRest 5d vs 2d🎾Serve 72%📈Form 5/10 · 2✗
THE MODEL'S REASONING

Ranking: #11 vs #110 (better ranked)

Recent form: 5/10 in recent matches

Model 78% vs market 65% → the model sees it as MORE likely than the odds

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.
@1.28
fair odds
+20.4%
expected value
HOW EACH FACTOR MATTERS
Level (Elo/ranking)▸ Bublik●●●
Bublik is #11 (Elo 1980) vs #110 (Elo 1918); baseline model gives him 62% to 36%, reflecting a clear class gap.
Serve/return▸ Bublik●●
Bublik holds at 72% vs Diaz Acosta's 64%, but Diaz Acosta returns better (42% vs 31%) — serve control is the stronger lever.
Form▸ Acosta●●
Diaz Acosta is 8-2 in his last 10 with a rising ranking (+41), while Bublik is 5-5 with a two-match losing streak.
Rest▸ Bublik●●
Bublik has 5 days rest and only 1 match in 14 days; Diaz Acosta played 5 matches in 14 days on just 2 days rest, risking fatigue.
Altitude▸ Bublik
At 760m the thinner air speeds the ball, which tends to help the stronger server — Bublik's 72% hold rate vs 64%.
Weather= Even
Mild, dry conditions (21°C, 37% humidity, 7km/h wind) add no clear edge to either player's game style.
LEVEL AND MOMENTUM

Bublik's ranking (#11, Elo 1980) and Diaz Acosta's (#110, Elo 1918) support the model's 62%-36% baseline split, a meaningful class gap that anchors the favorite's standing.

Recent form runs counter to that gap, though: Bublik is only 5-5 over his last 10 with a two-match losing streak, while Diaz Acosta is 8-2 with a live one-match win streak and a ranking trend of +41 versus Bublik's -1. The class edge is real, but form narrows the practical distance between them.

SERVE-RETURN DYNAMIC

Bublik's 72% service-points-won rate is well above Diaz Acosta's 64%, giving him more control when serving and a higher floor in tight sets.

Diaz Acosta counters with a better return game (42% vs Bublik's 31%), meaning he generates more break opportunities than Bublik typically allows. The serve differential is larger in absolute terms, so this dynamic leans toward Bublik, but it is not a one-sided mismatch.

SCHEDULE AND CONDITIONS

Rest favors Bublik clearly: 5 days off and just 1 match in the last two weeks, versus Diaz Acosta's 5 matches in 14 days on only 2 days of recovery — a congested stretch that can show up physically as matches lengthen.

The 760m altitude and mild, dry weather (21°C, 37% humidity, low 7km/h wind) modestly speed up the ball, a condition that generally favors the better server — here, Bublik at 72% versus 64%.

VALUE READ

The model sets Bublik at 78%, clearly above the market's implied 61% (odds 1.64), producing a nominal +28.2% expected value. That gap is large enough to flag, but it should be read with caution.

Diaz Acosta's momentum (8-2, rising ranking trend) and Bublik's recent dip (5-5, two-match skid) are real signals the model may be underweighting relative to the strong ranking/Elo gap. This is a case of model-versus-market divergence, not a confirmed edge — treat the projected value as a signal to weigh, not a promise.

Impact and analysis from real match data (Elo, form, head-to-head, rest, surface vs baseline, weather, altitude). The model ≈ the market on average; the odds already capture almost all the edge. 18+ · gamble responsibly.

Analyze today's matches →