D. Dzumhur vs H. Gaston — prediction

Elo rates both players identically at 1822, and their rankings sit a single spot apart (#108 vs #107), so neither the rating system nor the rankings alone tell you much. The gap opens in the baseline model, which gives Dzumhur 41% against Gaston's 24% — a signal that factors like recent quality wins and head-to-head are pulling the projection toward the favorite even though the raw level metrics are flat.

Both players hold serve at a similar clip — Dzumhur 60%, Gaston 61% — so neither is dominant on his own delivery. The separation shows up on return: Dzumhur wins 43% of return points against Gaston's 40%, meaning Dzumhur is statistically the more likely of the two to generate break chances. It's a narrow edge, not a mismatch, but it points to Dzumhur applying slightly more pressure on Gaston's service games than the reverse.

This is the clearest asymmetry in the data. Dzumhur is playing on 3 days' rest after 7 matches in the last two weeks, including a final at Umag just 3 days ago. Gaston, by contrast, has had a full week off and played one fewer match in that span. Congested scheduling and a deep tournament run this recently are real physical costs over best-of-three or five sets, and they weigh against Dzumhur regardless of his edge in the return column.

Gaston's 7-3 record over his last 10 matches is nominally better than Dzumhur's 6-4, and both players arrive having lost their last two. Dzumhur's quality wins — over Molcan (Elo 1933) and Svajda (Elo 1908) — are individually stronger than Gaston's best result (Diaz Acosta, Elo 1918), which helps explain why the model still leans his way despite the head-to-head win-loss gap. The single prior meeting, won by Dzumhur in 2025, adds a small additional tilt but shouldn't be overweighted given the tiny sample.

The model puts Dzumhur at 57%, only four points above the market's implied 53% at odds of 1.89, producing a modest 7.4% expected value. That's a real but not large edge, and it comes from a factor model with roughly 65% out-of-sample accuracy rather than a live-market signal — so this should be read as a mild statistical lean, not a confident price mismatch. Being the favorite here does not mean value is large, and the rest disadvantage is a genuine offsetting risk the market may already be pricing in.