P. Carreno-Busta vs L. van Assche — prediction

Carreno-Busta's Elo advantage (1942 vs 1895) is the foundation of his favorite tag, translating to a 57% model probability. That gap is real but not large — about 47 rating points, which in Elo terms is a modest edge rather than a dominant one.

The market, however, prices him much higher at 66% implied probability. That disconnect suggests the market is factoring in something beyond pure rating — likely the recent form and head-to-head data below — while the model stays more conservative.

Van Assche's 8-2 record over his last 10 matches is a clear signal of current sharpness, contrasting with Carreno-Busta's more uneven 6-4 stretch that includes two losses. The favorite does have notable quality wins over Lehecka (Elo 2028) and Tirante (Elo 1906), showing he can beat higher-rated players, but his overall trend is less steady.

The single head-to-head meeting also tilts toward van Assche, who won their only prior match this year at Challenger level. With just one data point this carries limited statistical weight, but it does confirm he has already found a way to beat this exact opponent.

On paper, Carreno-Busta holds the better service numbers (65% vs 61%), which should give him control on his own serve games. But van Assche's return game is sharper (47% vs 41%), meaning he converts more return points than Carreno-Busta does — enough to slightly outweigh the serve gap when the two numbers are combined (108 vs 106 total points won).

This near-balance suggests neither player has a clear stylistic advantage; the match may hinge on a handful of break points rather than one player's serve dominating throughout.

The core honest read here is a gap between model and market: Carreno-Busta is favored at 57% by Elo, but the market prices him at 66%, and the odds of 1.52 produce a negative expected value of -13.8%. That means, using this model's estimate, backing the favorite at this price does not represent value — the market is more confident in Carreno-Busta than the data-driven model is.

This is a case where being the favorite does not equal being the smart bet. The rating edge is real but slim, recent form and the head-to-head favor the underdog, and the price already assumes a stronger favorite performance than the numbers support. No profit is promised either way — this is simply a case where the market appears to be paying less than the actual risk warrants.