TENNISEDGE.GG

A. Holmgren vs F. Diaz Acosta — Prediction & Analysis

French Open · ATP French Open - Final · May 21, 2026 · 11:00 UTC

TennisEdge's AI model finds no betting value in A. Holmgren vs F. Diaz Acosta (French Open on 2026-05-21) — the market is efficient, with F. Diaz Acosta favoured at 1.39.

Surface clay Odds A. Holmgren 3.06 · F. Diaz Acosta 1.39

Verdict: no bet

The model found no value at current prices — the market has this one priced about right. That verdict is part of the public record too.
Profiles: A. Holmgren · F. Diaz Acosta · head-to-head record · French Open

Model probabilities

Independent estimates of who wins, each from a different method. The blend is what the bot prices against the market.

Method A. Holmgren F. Diaz Acosta
Elo
ratings-based, surface-weighted
24.4% 75.6%
Machine learning
29-feature gradient-boosted model
19.9% 80.1%
Serve / return
point-by-point simulation
51.3% 48.7%
Blend
what the bot actually prices off
31.9% 68.1%
Market implied
de-vigged from the opening odds
31.2% 68.8%

We publish how well these score against the closing market, including where they lose: how accurate are tennis predictions?

Elo ratings

PlayerSurfaceEloRecordMatches
A. Holmgren hard1624 118–73191
F. Diaz Acosta clay1816 188–98286

Our own ratings, computed from match results since 2018. Players with fewer than 5 rated matches are omitted rather than shown at a provisional starting value. Last updated July 29, 2026.

Head to head

F. Diaz Acosta leads 1–0 over 1 meeting. Full record.

DateEventSurfaceWinnerScore
May 21, 2026French OpenclayF. Diaz Acosta2 - 1

Recent form

A. Holmgren

L B. Shick 1 - 2 July 29, 2026
L L. Draxl 1 - 2 July 23, 2026
W G. Johns 2 - 1 July 21, 2026
L A. Vukic 2 - 1 July 19, 2026
W J. Boulais 2 - 0 July 17, 2026

F. Diaz Acosta

W M. Poljicak 2 - 0 Aug. 2, 2026
W V. Gaubas 2 - 0 Aug. 1, 2026
W J. C. Prado Angelo 2 - 1 July 31, 2026
W F. Forti 2 - 0 July 29, 2026
W T. Compagnucci 2 - 0 July 28, 2026

All May 21, 2026 results & picks →

Full analysis

Math-only prediction: no LLM calls used