C. Ugo Carabelli vs T. Droguet — prediction

The picture here is genuinely mixed. Carabelli holds the clearly better ranking (#59 vs #116) and is trending up (+9) while Droguet trends down (-9), which is the kind of momentum signal that tends to matter over a single match. Yet Droguet's Elo rating (1924) is actually higher than Carabelli's (1890), and his baseline model win rate (58%) outpaces Carabelli's (45%) before any match-specific adjustments are applied.

This tension explains why the model's final number (57% for Carabelli) is not a blowout projection: it is pulling the ranking/trend signal against a baseline that leans the other way, landing on a moderate edge rather than a lopsided one.

Both players enter with similar recent records — six wins and four losses in their last ten — so raw form does not separate them. The gap shows up in the quality of those wins: Carabelli has beaten Tiafoe (Elo 2031) and Berrettini (Elo 1965), both notably stronger opponents than Droguet's best recent result, a win over Blockx (Elo 1998).

That difference in the level of victories suggests Carabelli has performed better against tougher competition recently, even though the overall win-loss tallies look comparable.

Both men are playing on the same three days' rest, so there is no advantage there. The real gap is workload: Droguet has played 8 matches in the last 14 days compared to Carabelli's 3, a much heavier recent schedule that can leave legs and focus more taxed by this stage of a tournament.

Separately, both players reached the quarterfinals in Umag just three days ago, so the deep-run fatigue flag applies to each of them equally and does not tilt the match either way on its own.

The serve and return profiles are essentially a wash. Carabelli wins 62% of service points against Droguet's 63%, and Carabelli's 39% return rate is only a point above Droguet's 38%. Neither player has a meaningful mechanical edge in the exchange of serve, so this factor does not meaningfully push the projection in either direction.

The model gives Carabelli a 57% chance to win, versus a market-implied 54% at odds of 1.85, producing a modeled edge of about +6.3%. That is a real but modest gap — this is not a case of the model wildly disagreeing with the market, and the ranking-versus-Elo split noted above is a reminder that the underlying signals themselves are not unanimous.

Treat this as a small, honest edge rather than a confident pick. A positive expected value is not a guarantee of profit on a single match, and given Droguet's higher Elo and baseline win rate, this is closer to a coin-flip with a slight lean than a clear favorite situation.