Optimal Strategy in Darts Cricket

A computational re-test of Frongello’s strategies under realistic skill and bull difficulty — and what changes once you drop the symmetric assumptions.

A computational study — 2026 · 36.8 billion simulated games across 1,936 tournaments. Methodology →

Three headline findings

11 / 11 The Shape Reader, a strategy that wraps a priority list in a pattern detector and a phase gate, ranks first at every one of the eleven skill levels we tested under matched conditions. At matched pro skill it averages 62.8% across the field, seven points above the next-best strategy, and never loses a head-to-head. Read the Shape Reader →

22 different strategies lead the field somewhere across the 1,936 asymmetric skill-and-bull cells we tested. No single universal optimum — matchup conditions reshape the answer, and the Shape Reader’s top-of-field ranking applies under matched skill, not when it is the weaker player. Explore matchups →

20 · 18 A reinforcement-learning agent, trained from scratch, discovered a turn-order-dependent opening: 20 when first, 18 when second. None of the 30 hand-crafted strategies adapt to turn order. The AI story →

If you’re new to darts cricket, start here

Cricket is played on seven targets — 15, 16, 17, 18, 19, 20, and the Bullseye — with three darts per turn. Each target needs three marks to close: a single counts one, a double two, a triple three. Once you close a target your opponent has not, each further hit scores points equal to its face value (15, 16 … 25 for the Bull) until they close it too.

Winning requires both closing all seven targets and having at least as many points as your opponent. That dual condition is the whole game: close too slowly and you lose to tempo; score too long and you run out of darts to close. When to switch between the two is the central strategic question, and the one every strategy on this site answers differently.

Frongello's Framework

In 2018, Andrew Frongello published a systematic analysis of darts cricket strategy that defined 17 distinct strategies based on three parameters:

By simulating 20,000 games per matchup between every pair of strategies in a round-robin tournament — at both equal skill and with a skill advantage — Frongello arrived at four key findings:

  1. Score first, then cover — Build a point lead before shifting focus to closing numbers. S2 (“take any lead, then cover”) is optimal for equally skilled opponents. Purely defensive strategies consistently lose.
  2. Use extra darts when more skilled — If you are the stronger player, S6 (S2 + extra darts) outperforms S2 by extending the game and leveraging your accuracy advantage. Against an equal opponent, S2 edges out S6.
  3. Never chase — Do not waste darts trying to close numbers your opponent has already closed. Every flavor of chasing is dominated by S2 in every scenario.
  4. Weaker players want short games — When outmatched, play S2 and hope your opponent does the same. Building large point cushions (S3–S5) extends the game and favors the stronger player.

What We Did

Three acts. Act 1: reproduce Frongello under realistic skill. Act 2: break the symmetric assumptions (bull difficulty and asymmetric matchups). Act 3: surface the conditional answer interactively so you can see what wins for your matchup.

Key Findings

1. Frongello’s Core Findings Hold Under Realistic Conditions

All four of Frongello’s central conclusions survive testing at 11 skill levels with probabilistic misses: score first then cover, never chase, extra darts don’t help at equal skill, and weaker players want short games. His framework is sound — but it is a symmetric framework, and real darts are rarely played under perfectly symmetric conditions.

2. The Best Strategy We Found Is a Stack, Not a Recipe

Everything in the 30-strategy tournament is a priority list with a parameter or two tuned. We later built one that wraps that priority-list core in two outer layers: a pattern detector that reads the opponent’s board, and a phase gate that flips scoring to covering only in the closing stretch. We call it the Shape Reader. In the full round-robin against the thirty originals, it ranks first at every one of the eleven skill levels we tested. At matched pro skill (MPR 4.9) it averages 62.8% across the field, seven points above the next-best strategy, and never loses a head-to-head (worst matchup: 51.1%). Within the thirty originals, E12 (“Finish Opponent-Closed”) remains the narrow top-tier leader at 55.4%, with the next four within a point. Read the Shape Reader →

Corrections · 2026-04-20

Two simulator bugs were found and fixed: the turn cap was truncating long casual-skill games, and the tournament tally was attributing those truncated games to the losing side. Together they inflated Phase Switch’s apparent dominance at casual/amateur skill. After the fix, PS and PS×E12 are mid-pack (#9 and #6 at MPR 2.0 respectively), not #1. See commit history for the fix. The current pages reflect the corrected results.

3. Bull Difficulty Reshuffles the Field

Frongello’s model treated the bullseye as equally easy to hit as numbered segments. Applying a 0.75× bull difficulty multiplier costs E1 (Early Bull) 1.56pp on average and moves it from the top five to #8–#9. The chase strategies (S11–S17) gain 1.2–1.4pp and become competitive. Phase Switch is essentially bull-invariant (−0.1pp). Under extreme bull difficulty (0.25×), the effect amplifies: E1 loses 5.8pp and falls to #16–#19, while chase strategies gain 5–6pp. Full analysis →

4. Asymmetric Matchups Rewrite the Answer

Once you let P1 and P2 differ in both skill and bull accuracy, 83% of matchup slots change their optimal strategy. Each +0.25 step in bull accuracy advantage is worth ~8pp of win rate, or about 0.11 MPR in skill-equivalent. Harder bull acts as an equalizer for the weaker player — worth up to +2.15pp at the highest skill gaps. There is no single universal optimum. The results page lets you set all four parameters and see what wins for your matchup.

5. Which Strategy Wins Depends on Your Skill

Under realistic bull, the top strategy shifts sharply across skill bands. The symmetric picture (equal skill, equal bull):

Amateur · MPR ~1

E12, E10, E3, E2, S2

Top five within ~0.5pp at 57–58%. E12 leads narrowly; any scoring-before-covering strategy gets you most of the way there.

Skilled · MPR ~3

E12, E10, E3, E2, S2

Same compressed top five, ordering shifts slightly as E10 and E2 step up. PS_E12 enters the top six; plain PS sits ~#9.

Pro · MPR ~5

E12, E3, E2, S2, E10

At pro skill the top five finish within 0.3pp at 59–60%. E3 edges E12 at MPR 5.6 by a fraction of a point.

6. At Novice Level, Strategy Barely Matters

When most darts miss their target (MPR < 1.5), luck dominates. The gap between the best and worst strategies narrows sharply. For amateur and higher skill, the right strategy is worth 5–10 percentage points — the difference between winning and losing a match.

7. A Reinforcement Learning Agent Discovered Something No Human Designed

We trained a neural network to play from scratch — no strategy rules, just wins and losses. After 19 versions, 12 training bugs, and millions of games against the hardest hand-crafted bots, it found a turn-order-dependent opening: always open 20 when going first, always open 18 when going second. None of the 30 hand-crafted strategies adapt to turn order. The agent also independently confirmed Frongello’s principles (score before cover, never chase). The full story →

Explore the Results