What does the ProbWin Data Shop sell?

Ready-to-use historical sports datasets: odds (pre-match and closing), results, expected goals (xG) and pre-match machine-learning stats, for European football (15 leagues) and MLB. Every file ships as CSV + Parquet with a README documenting each column and its exact coverage.

How does the purchase work?

Secure Stripe payment from your ProbWin account. The download link arrives immediately (5 downloads included). Single-user licence: your models and analyses (commercial included) are yours; resale of the files prohibited.

Are the datasets updated?

Each product states its coverage (seasons included and cut-off date) at purchase time. In-progress seasons are sold as clearly identified season-to-date versions.

Verified data, ready to use

ProbWin Data Shop

Q: Is the data reliable for backtesting?

That is its whole point: authentic closing odds (including Pinnacle), stats verified leak-free by recomputation, zero missing values on core columns, and every coverage limit is quantified in the README instead of hidden.

The data we feed our own models, packaged for yours: historical odds with closing lines, results, xG and pre-match machine-learning stats. Every file matches its product page exactly — the README quantifies each column's coverage, season by season.

Built for people who model: point-in-time by design — odds are the ones recorded before kickoff, features frozen at prediction time (no future leakage), audited column by column. The same standards behind our public results journal.

600,000+Matches covered

3Sports

CSV + ParquetFormats

€9Starting at

Pick your sport

Football — 15 European leagues

27,109 matches over 6 seasons (2021-2027): 1X2 and Over/Under 2.5 with three levels of closing odds (Pinnacle, average, best), xG, and a Machine Learning dataset of 96 pre-match stats computed by ProbWin.

From €9 per season
Browse Football datasets →

MLB — 3 seasons (2023-2025)

7,070 games: Pinnacle pre-match totals and run lines, First 5 Innings, 65,350 timestamped line-movement snapshots, and 45 pre-match machine-learning stats with point-in-time xwOBA.

From €14
Browse MLB datasets →

Tennis — ATP + WTA

567,866 matches with results and pre-match odds (2015-2026, Grand Slams to ITF) + 3.19 million timestamped snapshots from our 2026 capture, the feed our own models use.

From €19
Browse Tennis datasets →

How it works

Pick your dataset

Every product page details the columns, per-season coverage, and shows a real data sample before you buy.

Pay via Stripe

Secure payment from your ProbWin account. One-time price, no subscription.

Download instantly

Immediate link — CSV + Parquet + full README in one archive. 5 downloads included.

Secure Stripe paymentCard payment, one-time price, no hidden subscription.

Instant downloadThe link arrives as soon as payment clears.

No surprisesReal sample and quantified coverage on every page, before you buy.

Personal supportQuestions? [email protected]

Frequently asked questions

What exactly does the Data Shop sell?

Ready-to-use historical datasets for backtesting and machine learning: odds (pre-match and closing), results, expected goals (xG — the quality of chances created) and pre-match stats. The same data that powers our own models.

Is the data reliable for backtesting?

Authentic closing odds (including Pinnacle), stats verified by strict-cutoff recomputation (no future leakage), zero missing values on core columns. Every coverage limit is quantified in the README — never hidden.

What is the licence?

Single-user licence: the models, backtests and analyses you build with this data are yours, commercial use included. Only resale and redistribution of the files themselves are prohibited.

A question before buying?

Write to us: [email protected] — we answer personally.