Tennis Challenger vs ATP: are favourites less reliable?

At Challenger level, is the favourite less reliable than on the ATP Tour? The idea is widespread: the rung below the ATP Tour mixes rising players, veterans on their way down, smaller events and tighter rankings. It seems logical that upsets are more common there, and predictions therefore shakier.

We took 86,865 singles matches, played between January 2022 and September 2026, including 42,047 at men’s Challenger events, and compared each result with what the Elo rating of the two players predicted. The answer is clear, and it runs against intuition.

Favourites win as often at Challenger level

The simplest measure first: how often does the player ranked higher by Elo win? To tell whether they win more or less than expected, we compare their win rate with the one Elo leads us to expect, calculated on ATP and WTA matches only.

64.0% at men’s Challenger events, 64.5% on the ATP Tour: half a point apart, which cannot be told from chance (z = 1.24). For the women, 65.1% against 65.7%, same conclusion. The favourite loses more than one match in three on all four tours.

The size of the gap between the two players is the same too: the average favourite is at 64.0% in men’s Challenger and 64.7% on the ATP Tour. We are not comparing tighter matches with more lopsided ones.

At equal probability, no gap

A 60% favourite and an 85% favourite are not comparable. The measure that matters is the gap to what is expected: does a 70% Elo favourite win as often at Challenger level as on the ATP Tour? With the expected rate removed, the gap between the two tours is +0.2 point for the men, with a 95% interval from −0.6 to +1.0 point. For the women, +0.3 point, from −0.9 to +1.6. In both cases the interval contains zero, and it is narrow: if there were a Challenger effect, it would be under one point.

Band by band, the two columns overlap. A 75% Elo favourite wins about three matches in four, whether the match is a Challenger or an ATP event.

The gaps that stand out, and why to be wary of them

Cutting the data further, a few cells stand out. These cuts were made after seeing the total, so they are hints, not proof:

• Grass disappoints favourites everywhere. The favourite wins 62.8% on the ATP Tour (2.7 points under expected) and 61.0% at men’s Challenger events (2.2 points under expected, on only 920 matches). The direction is the same on both tours, which points to an effect of the surface rather than the tour. Our study of surfaces goes into more detail.
• Women’s Challenger, favourites between 70 and 80%. They win 74.5% against 71.2% expected (1,426 matches, z = 2.7). Across sixteen band cells crossed with tours, a z of that size turns up by chance about once in ten: something to watch, not to believe.
• ATP, favourites at 80% and above. They win 82.6% against 84.2% expected (3,039 matches, z = −2.5), with no equivalent at men’s Challenger events. Same caveat.

One last point of method: over each half of the period, 2022 to 2024 then 2024 to 2026, the ATP Tour and men’s Challenger move together against the expected rate. That shared movement is therefore not a Challenger effect.

Bookmakers are not fooled either

On the matches for which we have bookmaker odds, about one match in two, the market favourite wins 67.0% of men’s Challenger matches (21,863 matches) against 65.4% implied by the margin-free odds. On the ATP Tour: 67.9% realised against 66.8% implied (9,723 matches). The odds slightly underrate the favourite in both cases, by 1.7 points at Challenger level and 1.1 on the ATP Tour, and that 0.5-point difference between tours cannot be told from chance (95% interval from −0.6 to +1.6).

The market beats Elo alone on both tours, by 2.3 points at men’s Challenger events and 3.2 on the ATP Tour: odds gather far more information than a rating. But Challenger is not a zone where the odds would be weaker. No obvious flaw to exploit here either: our betting guide explains why an idea like this is not enough to make money.

How to read a Challenger match

• Don’t treat it as a lottery. A favourite between 70 and 80% wins 71% of the time there, as on the ATP Tour. The displayed probability means the same thing on both tours.
• Don’t treat it as safe either. The favourite still loses more than one match in three, and a 60% favourite loses four times in ten.
• Look at what really changes: the surface (grass, rare at Challenger level), fatigue and rest days.
• Beware of missing odds. One Challenger match in two has no odds in our archive: with nothing to compare against, a model’s probability is all the more useful.

Pronocast covers men’s and women’s Challenger, in singles and doubles, with the same method as the ATP and WTA. Today’s Challenger predictions show each player’s probability, and the verified results page shows what they are worth once the matches are played.

Method

Singles matches on the ATP, men’s Challenger, WTA and women’s Challenger tours, from 1 January 2022 to 22 September 2026, in which both players had already played at least twenty matches: 86,865 matches. About 9% of the matches on these tours are dropped by that filter, the ones where a player’s history is too short for a reliable Elo. The ITF circuit is not in the study. The expected rate is the real win rate observed for a given Elo probability, in 2.5-point bands, calculated on ATP and WTA matches only: a Challenger favourite who did worse than that expected rate would be less reliable than a favourite of the same probability on the main tour. Gaps are judged with a z test, and the interval of the gap between two groups is that gap plus or minus 1.96 times its standard error. Odds are the bookmaker average, converted to a margin-free probability, available for 44,737 matches of the sample. The full script is kept with our data.