On raw level, Zverev is the stronger player: his Elo of 2204 and #3 ranking sit clearly above Fritz's 2064 and #9. That gap is real and explains most of the model's lean toward Zverev as favorite.

But this match isn't played on a neutral canvas — it's grass, a surface where the ranking gap has historically flipped. The broader level advantage has to be weighed against a surface and history that both point the other way.

The surface numbers are stark: Fritz plays grass at 83%, a full 16 points above his own 67% baseline — a significant positive deviation that suggests grass genuinely elevates his game, likely through a bigger, more direct service game rewarded by the surface's low bounce and pace.

Zverev shows the opposite pattern: 69% on grass versus a 73% baseline, a 4-point dip. That's not a collapse, but it's a real signal that grass blunts some of his baseline-built advantages rather than amplifying them.

Fritz has won 8 of their 11 meetings, including the last four in a row — most recently in 2026 and 2025, plus a win listed among his current quality wins. That's not a small-sample quirk; it's a sustained pattern across multiple years and surfaces.

Combined with the grass splits, the history suggests Fritz has found a way to match up against Zverev's game specifically, which tempers how much weight the raw Elo/ranking gap should carry in this particular pairing.

Serve numbers are identical at 66% for both, so the hot, dry, low-wind conditions that typically favor the server don't create a lopsided edge here — it's a wash. Zverev's return rate (38% vs Fritz's 36%) offers a marginal edge on break points, but the gap is small.

Rest slightly favors Fritz, who has one extra day (2 vs 1) despite both playing four matches in the last two weeks. None of these secondary factors are large enough to override the surface and head-to-head signals on their own.

The model prices Zverev at 63% against a market-implied 47%, generating a large theoretical EV of +35.4%. That gap is worth noting, but it should be read with caution: this is an ATP factor model with roughly 65% out-of-sample accuracy, not a certainty, and it appears to weight the Elo/ranking gap heavily while the head-to-head (8-3, four straight to Fritz) and surface splits (Fritz +16 vs baseline, Zverev -4) pull in the opposite direction.

In practice, this looks like a case where the model's favorite and the market's favorite diverge more than the underlying match-specific evidence comfortably supports. Being priced as the model favorite is not the same as being the more likely winner on this surface against this specific opponent — treat the implied edge as unproven rather than a promised value bet.