Archery Statistics & Analysis

Most Consistent Archers: How to Measure Consistency

August 20, 2026 / 10 min read

Most Consistent Archers: How to Measure Consistency

Consistency is the single most useful performance concept for anyone who follows competitive archery closely — athletes, coaches, statisticians and bettors alike. But “consistent” can mean different things depending on the format (72‑arrow ranking round vs head‑to‑head matchplay), the discipline (recurve vs compound), and the sample size available. This guide explains practical, competition‑level metrics for measuring consistency using primary sources (World Archery / Ianseo results), explains how to avoid common statistical pitfalls, and shows how bettors should interpret consistency when markets appear (usually around major events only).

Why format and sample size matter

World Archery competition formats create very different statistical environments. The typical Olympic/recurve ranking round is 72 arrows at 70 metres (maximum 720 points), which gives a long, low‑variance sample that’s appropriate for measuring an archer’s baseline scoring level. Matchplay, by contrast, uses sets (recurve) or short cumulative matches (compound) with far fewer arrows, producing much higher variance per contest and making short‑term “consistency” harder to estimate from match results alone. The ranking‑round / matchplay distinction is central to any valid consistency metric. (worldarchery.sport)

Practical consequences

72‑arrow ranking rounds (max 720) are best for per‑arrow and per‑end variance analysis — they approximate an archer’s baseline accuracy. (worldarchery.sport)
Short match formats magnify luck and situational factors: a single wind gust, one nervous arrow, or a single shoot‑off arrow can decide a match. Use match outcomes to supplement, not replace, ranking‑round measures. (worldarchery.sport)
Compound and recurve scoring nuances matter: compounds often post many 9s/10s at 50m while recurve 70m scores contain wider spread; inner‑10/X counts are tracked differently and are useful separate metrics. (scribd.com)

Core metrics for archer consistency (what to compute)

Below are metrics that are practical to compute from competition scorecards (World Archery / Ianseo output) and that provide complementary views of consistency.

Metric	What it measures	Why it helps	Limitations
Mean score (per arrow / per 72 arrows)	Average points (e.g., 700/72 = 9.72 per arrow)	Baseline performance level; easy to compare across archers	Ignores spread/variance; affected by outliers and small samples
Standard deviation (SD) of per‑arrow or per‑end scores	Typical spread around the mean	Direct measure of score volatility; lower SD = more consistent	Needs raw arrow or end data; per‑arrow SD can be very small on compound events
Coefficient of variation (CV)	SD divided by mean	Normalises volatility for different mean levels (useful across divisions)	Less intuitive units; unstable for very small means
% Tens and % Xs (inner‑10s)	Share of arrows hitting the 10 ring or inner‑10	Directly tied to competitive scoring and tiebreak conventions	X usage differs by discipline; tied to scoring face used. (scribd.com)
Rolling mean rolling SD (e.g., last N arrows or last M ranking rounds)	Form over recent windows	Captures short‑run trend and volatility	Window choice is arbitrary and can overfit
Match‑outcome Elo / Glicko	Head‑to‑head competitive strength accounting for opponent quality	Useful for predicting match results; updates after each match	Less sensitive to arrow‑level precision; influenced by pairing frequency
“Reliability index” (mean ÷ SD — archery Sharpe‑style)	Ratio of central tendency to volatility	Synthesises scoring level and steadiness into a single ranking	Arbitrary scale; needs comparison across similar formats

How to calculate these from primary sources

World Archery publishes official ranking‑round totals and match results, and event result systems (Ianseo) provide end‑by‑end and arrow‑level scorecards for many events. Use those official outputs as the data source for calculations rather than second‑hand summaries. (ianseo.net)

Recommended workflow:

Collect ranking‑round scorecards (72‑arrow totals) and, where available, arrow/end breakdowns from Ianseo or the World Archery competition result pages. (ianseo.net)
Compute per‑arrow averages and SDs when arrow‑level data exists; otherwise compute per‑end averages (6 arrows) and SDs as a proxy.
Compute % Tens and % Xs from scorecards; record separate values for ranking rounds vs matchplay.
Apply a small‑sample shrinkage (Bayesian or empirical Bayes) when the number of arrows or events per archer is small — this reduces over‑ranking of lucky small samples. (Details below.)

Small‑sample correction (why shrinkage matters)

Imagine an archer who shoots one competition and posts 710/720: the raw mean is excellent, but with only one 72‑arrow sample you can’t be confident the archer will keep that level. Shrinkage (empirical Bayes) pulls extreme estimates toward the population mean by an amount proportional to sample size. In practical terms, use a hierarchical model or a simple James‑Stein shrinkage for mean and variance estimates to reduce false confidence from tiny samples. This is standard practice in sports analytics when samples are unequal.

Why correlation ≠ predictive power

High consistency measures correlate with long‑term success, but correlation is not causation and does not guarantee future wins. Factors that weaken predictive power:

Format switch: dominance in ranking rounds does not always translate to head‑to‑head match wins (different pressure and set scoring). (worldarchery.sport)
Environmental sensitivity: wind and venue conditions can change the relative advantage between archers.
Small samples and luck: single shoot‑offs and one‑arrow events have high variance.
Injuries, equipment changes, and selection decisions introduce non‑stationarity that past scores don’t capture.

For bettors and modelers, this means consistency metrics should be used as inputs to probabilistic models (odds estimation, implied chance), not as absolute predictors. Always quantify uncertainty (confidence intervals, posterior distributions) around any forecast derived from these metrics.

Example: what to do for a betting model

When archery markets are available (see note below on availability), a disciplined approach for model input is:

Primary features: last 3 ranking‑round means, long‑term mean, SD of per‑arrow and per‑end, % tens, % Xs, and Elo/Glicko based on recent head‑to‑head. Use format‑matched features (ranking‑round metrics for outright/ranking markets; matchplay metrics for match odds).
Adjust for conditions: if wind forecasts or venue altitude data exist, adjust expected variance upward.
Apply shrinkage for small samples and produce predictive intervals, not point estimates.
When markets are thin (few bets) prefer conservative staking: larger model uncertainty should result in smaller stakes. (Responsible gambling reminder below.)

Responsible‑gambling note: archery markets tend to be niche and, when offered, are often shallow; bettors should treat them as specialist markets and always bet within limits they can afford to lose. The editorial part of this guide is informational, not financial advice.

Bookmaker coverage: when and where archery markets appear

Archery fixed‑odds and in‑play markets are typically event‑driven: they appear around major World Archery events (World Championships, Hyundai Archery World Cup stages), Olympic Games and sometimes continental championships. Major bookmakers publish sport‑specific rules for archery and settle bets using official competition statistics when markets are offered. Examples: bet365 maintains an archery rules page which specifies settlement conventions (extra arrows count for settlement, podium presentation determines outright settlement), and specialist bookmaker guides note that live/in‑play archery markets are most commonly available during major events. Always check each bookmaker’s market list during an active event window — continuous archery markets are not guaranteed outside those windows. (help.bet365.com)

From an archery‑bettor’s perspective, that means:

Plan your data collection before events: bookmakers will only list archers/markets shortly before or during events.
Compare settlement rules across books (e.g., how they treat incomplete matches or shoot‑offs) because differences affect model outcomes. bet365’s archery rules are explicit about such cases. (help.bet365.com)

Format differences to control for in analysis

Account for these when measuring and comparing consistency:

Discipline distances/scoring faces (recurve 70m with 122cm face vs compound 50m with 80cm face and inner‑10 rules). (worldarchery.sport)
Indoor vs outdoor rounds — indoor faces and distances reduce environmental variance and change ideal metric baselines.
Team and mixed events combine arrows from multiple archers — use per‑archer contributions where possible.
Age/division effects — Masters/juniors have different variance baselines; compare like with like. (worldarchery.sport)

Common statistical traps and how to avoid them

Avoid using raw win/loss counts from short match series as a consistency proxy — low sample bias is severe. Use match records only with opponent‑quality weighting (Elo/Glicko) and combine with ranking‑round statistics.
Never compare per‑round totals across disciplines without normalisation (e.g., use per‑arrow mean or z‑scores relative to the event field).
Beware of survivorship bias: analysing only finalists will overstate consistency for the full competitor population.
When using historical World Archery rankings, remember the ranking formula uses best results and event weights; a high ranking reflects results selection as well as raw consistency. Check the World Archery rankings methodology and update schedule. (ianseo.net)

Worked example (conceptual)

The following outlines a lightweight reproducible example you can implement with publicly available results:

Download ranking‑round scorecards for the last 12 months for the event(s) you want to study from Ianseo or World Archery result pages. (ianseo.net)
For each archer, compute: mean per‑arrow (total/72), SD of per‑end totals (6‑arrow ends), % tens and % Xs.
Compute a reliability index: reliability = (mean per‑arrow) ÷ (SD per‑end). Rank archers by this index for a format‑matched leaderboard.
Apply empirical Bayes shrinkage to the mean and SD estimates; one simple approach is to blend the individual mean with the field mean weighted by sample size (N arrows). This reduces the rank volatility of archers with few events.
Validate: split data into train/test by events (e.g., use 8 events to compute metrics and test predictive power for match outcomes in the subsequent 4 events). Use calibration metrics (Brier score) for probabilistic predictions rather than raw accuracy.

Because arrow‑level data is necessary for true per‑arrow SD, many analysts use per‑end metrics when arrow logs are unavailable; that provides a reasonable proxy while remaining practical for scraping event PDF scorecards. (ianseo.net)

What this means for bettors and performance analysts

Consistency metrics can give a significant edge if used carefully: they help you identify archers who routinely hit tens under varying conditions (low volatility) versus “hot” archers with high short‑run peaks. But treat them as probabilistic signals, not certainties. When archery markets are offered (often only around major events), combine consistency measures with event‑specific factors (venue, wind, draw/pairings, recent travel or injury reports) and bookmaker settlement rules before placing stakes. (help.bet365.com)

FAQ

Q: Which is the single best consistency metric

A: There’s no single best metric. For long‑form measurement the SD of per‑arrow (or per‑end) combined with mean per‑arrow is most practical. A ratio of mean to SD (reliability index) is useful for ranking archers across events when the format is the same.

Q: Can ranking‑round consistency predict matchplay outcomes

A: Partially. Ranking round gives a low‑variance estimate of scoring skill, which improves probabilistic forecasts for match outcomes, but matchplay’s shorter format and set scoring increase variance — so combine ranking metrics with matchplay‑specific features and opponent Elo for best results. (worldarchery.sport)

Q: Are bookmaker archery markets always available

A: No. Archery markets are typically event‑dependent and most bookmakers only list markets for World Archery majors and the Olympics. Verify coverage on the bookmaker’s site during an active event. Bookmakers also publish archery settlement rules which should be reviewed before wagering. (livepanthers.com)

Q: Where do I get arrow‑level data

A: Best sources are World Archery competition result pages and Ianseo event result exports. These provide end‑by‑end and sometimes arrow‑by‑arrow scorecards for official events. (ianseo.net)

Q: How many ranking rounds are enough to judge consistency

A: More is better. Practically, 3–5 full 72‑arrow ranking rounds provide a usable estimate if you apply shrinkage; 8–12 is preferable for stable estimates across typical environmental variation.

Most Consistent Archers: How to Measure Consistency