T. Pereira vs J. Hrazdil — prediction

The 328-point Elo gap (1653 vs 1325) is the largest structural edge in this match, and it's reinforced by the ranking picture: Pereira sits at No. 371 while Hrazdil has no ATP ranking at all. Together these numbers produce the model's 87% win probability for Pereira — a gap wide enough that, on rating alone, this looks like a mismatch rather than a close contest.

Recent form sharpens the class gap rather than softening it. Hrazdil has lost all ten of his last matches and is riding a 14-match losing streak with zero quality wins logged. Pereira's own form is far from sparkling — 4-6 over his last ten with a current 2-match losing streak — but even a mediocre run for him looks strong next to Hrazdil's complete skid.

Pereira's own numbers show a 62% rate of service points won, a tour-typical figure that should let him hold routinely, paired with a 38% return rate that suggests he isn't primarily a break-point threat. There is no serve or return data available for Hrazdil, so this can only be read as a description of Pereira's game, not a direct head-to-head comparison — but a 62% service number on its own is a meaningful baseline of control.

Rest is close to even — 7 days since Pereira's last match versus 6 for Hrazdil — so there's no clear scheduling edge there. The one flag worth noting is workload: Pereira has played 4 matches in the last 14 days against Hrazdil's 2, which, combined with his 2-match losing streak, is a mild fatigue consideration, though not one large enough to offset the class and form gaps above.

The market prices this at 1.04, implying a 96% win probability for Pereira, while the model — an Elo-based estimate in a soft, less-analyzed ITF market — puts him at 87%. That gap produces a -9.6% expected value, meaning the price is asking bettors to pay more certainty than the model supports.

Pereira is a legitimate favorite on rating, ranking, and form, but favorite does not equal value here. With the Elo edge unproven in this tier and the EV already negative, this is a case where the model agrees with the market's direction but not its price — there is no backed edge to act on.