PythiaFootballPythiaFootballBETA Free Picks Fixtures League Stats Team Stats Player Stats Pro Picks PRO Research Head-to-Head Validation Pricing Log in
Research

Which European league is hardest to predict?

We took every match in our database that carries a closing bookmaker price, 52,476 of them across 22 leagues in 11 countries, and asked a simple question: how often does the favourite actually win? Then we turned the same prices around and asked how well the market knows its own business.
Last updated 04 September 2026 · PythiaFootball research
52,476
Matches analysed
settled, with a closing price
22
Leagues
across 11 countries
50.6%
Favourite wins
the market's shortest price
26.2%
Draws
about one match in four
43.0%
Home wins
12.2pt edge over away
30.8%
Away wins
visitors take all three
1. The finding
Italy - Serie B is the hardest league in Europe to call: the favourite wins just 44.8% of the time. At the other end, the favourite in Portugal - Primeira Liga wins 56.5%, a gap of 11.7 percentage points. Italy - Serie B also produces the most draws anywhere in this sample: 31.6% of its matches end level.
2. The bigger finding: chaos lives in the second divisions, everywhere
54.0%
Top divisions
11 leagues
47.7%
Second tier and below
11 leagues
6.4pt
The division gap
±0.9 at 95%
12/55
Top-flight pairs that separate
of 4.5pt total spread
Rank Europe's top flights against each other and most of the ordering is noise: of the 55 possible pairings, only 12 separate at the 95% level, and those are the extremes. The division gap is a different kind of number. It is 6.4 points, give or take 0.9, which is bigger than the entire 4.5-point spread across the top flights, and measured tightly enough that nothing about it is in doubt. Which division you are watching tells you more about how predictable the football will be than which country you are watching it in.

The likeliest reason is squad depth rather than anything romantic about lower-league football. A top flight contains a handful of clubs that can afford to be reliably better than everyone else; a second tier is a flatter distribution, and a flatter distribution produces more matches the market cannot separate.
3. Outcome rates, by league
The Predictability Index: how often the shortest price came in. Lower means harder to predict. Bars are scaled across the 43-58% band the data occupies, so the differences are visible; figures are ± the 95% confidence interval.
Serie BItaly
44.8%±1.9
Ligue 2France
45.1%±1.9
ChampionshipScotland
46.2%±2.4
46.4%±1.7
League TwoEngland
46.4%±1.7
League TwoScotland
46.9%±2.9
47.3%±1.7
48.0%±2.1
League OneEngland
50.4%±1.7
50.5%±1.7
51.9%±2.3
League OneScotland
52.2%±2.9
Ligue 1France
52.4%±2.0
52.7%±2.1
BundesligaGermany
52.8%±2.1
53.0%±2.0
La LigaSpain
53.8%±1.9
Serie AItaly
54.4%±1.9
PremiershipScotland
55.0%±2.5
55.3%±2.0
EredivisieNetherlands
56.2%±2.1
56.5%±2.1
4. Does the market actually know?
Every selection, grouped by the probability its closing price implied once the bookmaker's margin is removed, against how often it actually won. A perfectly calibrated market would print the same number in both columns. Grey is what the market said, blue is what happened.
Implied band
Said
Actual
Diff
0-10%
7.5%
5.3%
-2.2
10-20%
16.0%
14.7%
-1.3
20-30%
26.0%
25.4%
-0.6
30-40%
34.1%
34.2%
+0.1
40-50%
44.6%
45.2%
+0.7
50-60%
54.5%
56.0%
+1.5
60-70%
64.5%
68.0%
+3.5
70-80%
74.4%
77.9%
+3.5
80-90%
83.5%
87.6%
+4.1
It is close, and closer than most people expect. But the error is not random: it runs the same way at both ends. Selections priced at around 7% win only 5.3% of the time, while those priced at 83% win 87.6%. Longshots are systematically overpriced and strong favourites systematically underpriced. This is the favourite-longshot bias, first documented in racing markets in the 1940s and still measurable in football closing prices today.
5. Home advantage, season by season
Share of matches won by the home side, across the same priced sample, scaled across the 39-46% band the data occupies.
2019-204,675 matches
43.8%±1.4
2020-217,645 matches
40.8%±1.1
2021-227,821 matches
42.5%±1.1
2022-238,089 matches
44.1%±1.1
2023-248,031 matches
43.0%±1.1
2024-257,932 matches
43.8%±1.1
2025-267,806 matches
43.0%±1.1
2026-273,025 matches
43.5%±1.8
The weakest season for home sides is 2020-21, at 40.8%. That was the season played almost entirely in empty stadiums. Home wins fell to 40.8% from 43.8% the year before and recovered to 42.5% once crowds came back, while away wins moved the other way, from 29.7% to 32.9%. It is the closest thing to a controlled experiment football has ever run on its own home advantage, and the honest reading is that the effect was real but modest: about three points of win rate, not the collapse it was reported as at the time. Note too that 2019-20 itself finished behind closed doors, so the pre-pandemic baseline here is if anything slightly understated.
6. The longest prices that came in
The biggest upsets in the sample, by closing price. The market gave every one of these results less than a 5% chance.
Result
Price
Premiership · 04 Mar 2020
25.90
Liga II · 21 Aug 2026
24.29
Bundesliga · 21 Jan 2024
24.26
Bundesliga · 05 Oct 2019
23.32
Süper Lig · 25 May 2025
23.21
Eredivisie · 30 Aug 2025
21.19
7. The leagues we could not include, and why
We hold results for 143 competitions across 54 countries. This study needs a closing price on every outcome, and that history is thinner than the results are: 113 competitions have at least one priced match, and only 22 clear the 400-match floor a stable figure requires.

So the leagues missing from this table are missing for want of odds history, not on merit. We record closing prices every night, which means the floor is a moving target and this study has a natural next edition: leagues that cross it will appear in it. If yours is missing and you would like it measured, that is a good reason to get in touch.
8. What this study cannot tell you
Most of the ordering among top divisions is not real. Of the 55 possible pairings of Europe's top flights, 43 sit inside the margin of error, so for those pairs we cannot say one is harder to predict than the other. The 12 that do separate are the extremes of that group rather than neighbours in the table. Read the top-flight ranking as a rough banding, not a league table: the gap between 4th and 8th on this measure is noise.

None of this is a betting edge either. The calibration table shows where the market's prices sit against results once the margin is stripped out, and that margin is not stripped out of the price anyone is actually offered.

Coverage is a limit worth stating plainly: closing prices are available for 22 leagues in 11 countries, so this is a study of Europe's most heavily priced competitions, not of every league in Europe.
9. Method
The Predictability Index is the share of a league's matches won by the market favourite: higher means the favourite is more often right, lower means the league is harder to call.

For every settled match carrying an average closing 1X2 price, we removed the bookmaker margin (normalising the three implied probabilities to sum to 1), took the shortest price as the favourite, and recorded whether that selection won. The market is the benchmark rather than our own model on purpose: it is the best information publicly available before kick-off, and using it means this measures the sport rather than grading our own homework.

Every figure on this page describes the same set of matches. Leagues with fewer than 400 priced matches are excluded, as are the matches where the two shortest prices are identical and there is therefore no favourite. Probability bands holding fewer than 300 selections, and part-seasons of fewer than 3,000 matches, are not plotted. Confidence intervals are the normal approximation at 95%.

Closing prices come from football-data.co.uk, whose historical odds archive is what makes a study on this timescale possible at all. Results, fixtures and match statistics are our own.
10. Share a league
Every league in this study has its own share card. Pick one and the page shares as that league's result, ready to post or to attach to an email. Free to use with a link back.
Serie BLigue 2ChampionshipSegunda DivisiónLeague TwoLeague TwoChampionship2. BundesligaLeague OneNational LeagueSuper League 1League OneLigue 1Jupiler Pro LeagueBundesligaSüper LigLa LigaSerie APremiershipPremier LeagueEredivisiePrimeira Liga
11. Cite this
PythiaFootball (2026). Which European league is hardest to predict? https://pythiafootball.com/research/hardest-league-to-predict/
Charts and figures are free to reuse with a link back to this page.