Data sources
Summary + sample · full document is 1,708 words
Source × coverage × cost × access difficulty. Use this to decide what to integrate first and where the binding constraints are. Companion to data-sources.md (which has narrative context).
Costs are approximate and as of mid-2026; treat as orders of magnitude. Access difficulty is a 1–5 score: 1 = "git clone, you're done"; 5 = "vendor partnership required."
| Source | What it covers | Cost | Access difficulty | Notes |
|---|---|---|---|---|
| Football-Data.co.uk | Top European leagues 1993–present, results + pre-match odds, partial closing odds | Free | 1 | The starter dataset; CSV per league per season. |
| Betfair Historic | Top-flight English football back to 2008+, sub-second exchange order books and prices | Free | 2 | Registration + download of large compressed JSONs. Best public source of fillable closing prices. |
| Pinnacle archives (3rd-party) | Pinnacle pre-match + closing prices, partial coverage via OddsAPI archive | Paid (mid-three to low-four figures/year) or free with self-scrape | 3 | First-party export not offered; self-scrape and accumulate is the cheapest long-run path. |
| Smarkets API | Top-flight English football, exchange order books in real-time | Free | 2 | Lower liquidity than Betfair; useful cross-check. |
…
Pro
Data Sources Matrix runs 1,708 words. Pro members get every research note in full, plus the arbitrage feed, model outputs, and weekly updates.