Lab

Open research on 447,955 darts matches

Every study asks one question of the archive, answers it with a public Python tool that reads only through the API, and publishes the result here in full. The ones that fail are published too.

Open the live studyRun them with a free key
Studies
5
1 live, 4 next
Matches behind them
447,955
Every result the archive holds
Priced matches
10,397
with a closing price, 2016 to 2026
Every tool
Public
Python, on GitHub, API only
Live study

Ratings vs the market

A rating built from results alone, scored against the market's closing price on every match that had one. The market still wins, as it should: it knows about injuries, form and money that results never show. What the study measures is how close results alone can get, and what each extra source of results is worth.

  • 0.0172 between the market and our best rating, in log loss
  • 80% of the market's edge over a coin toss, from results alone
  • 19% of the gap closed by adding the live results layer, better in 9 of 11 years
Open the studyThe endpoint behind it

Log loss on 10,397 priced matches

Lower is better. Every figure is out of sample.

Closing pricethe market, the bar to beat
0.6062
Ratings, both sourcesarchive plus live results
0.6234
Ratings, archive onlyresults from the archive alone
0.6275
Coin tossfifty-fifty on every match
0.6931
0.560.630.7

Every study

One question each. The headline is what the tool found on the archive; open a study for the method, the full table and the code.

LiveNext
StudyWhat it foundRepository
Ratings vs the marketHow close does a rating model built only from results get to the betting market?
Market 0.6062, ratings 0.6234Log loss over 10,397 priced matches, 2016 to 2026
darts-ratings-vs-marketLive
Open
The throw-first advantageHow much is throwing first in a leg worth, and does it shrink as the standard rises?
Not run yetPlanned: Legs won throwing first against legs won throwing second, from the per-match splits, grouped by tier and by the pair's combined three-dart average.
darts-throw-firstNext
Preview
Who finishes under pressureWhich players convert doubles above what their scoring suggests, and does it last?
Not run yetPlanned: Each player's checkout percentage against the level their three-dart average predicts, then whether being above it in one season carries into the next.
darts-checkout-pressureNext
Preview
180s lines, testedIs the bookmaker line on total 180s beatable from averages alone?
Not run yetPlanned: The 180s prop model, run before each match, against the result, out of sample by season.
darts-180s-linesNext
Preview
Brackets, simulatedHow often does the favourite actually win a knockout, and how often is it an outsider?
Not run yetPlanned: Every major draw since 2016, simulated from the ratings before the first match, against who actually won.
darts-bracket-simNext
Preview

The rules every study keeps

Darts research is easy to flatter. These keep ours honest.

No look-ahead

Every figure uses only what was known before the match. Ratings are replayed in date order, never fitted on the future.

Out of sample

Anything tuned is tuned on one period and scored on another. A result that only holds in the data it was built on is not published as one.

Failures published

If a study finds nothing, the page says so, with the numbers. Not every question has the answer people hope for.

API only

The tools read the public endpoints, nothing else, so what you run at home is exactly what produced the page.

Got a question for the archive?

Suggest a study and we will tell you whether the data can answer it. The good ones become the next page here, with your question at the top.

Suggest a studyGet a free key
Run the live study yourself 
python -m ratingsvsmarket
From the darts-ratings-vs-market repository. Priced matches need a key on any paid plan.