DESCRIPTIVEDescriptive StatisticsStatistics Calculator
๐Ÿ“Š

Average Rating โ€” Weighted Avg, Bayesian Avg, Wilson Score

Compute weighted average, Bayesian average, Wilson score for star ratings. Distribution, mode, median, standard deviation. Used by IMDb, Reddit, and product reviews.

Concept Fundamentals
ฮฃ(wแตขยทxแตข)/ฮฃwแตข
Weighted Mean
Weighted average formula
(Cยทm + ฮฃx)/(C+n)
Bayesian Avg
Prior-smoothed rating
Lower confidence bound
Wilson Score
Small-sample correction
Reviews & rankings
Application
Product rating systems
Compute RatingEnter count for each rating level

Why This Statistical Analysis Matters

Why: Raw average is misleading with few reviews. A single 5-star review gives 100%. Bayesian average and Wilson score penalize low sample sizes for fair ranking.

How: Enter counts per rating. Weighted avg = ฮฃ(rร—count)/n. Bayesian = (Cร—m + ฮฃ(rร—count))/(C+n). Wilson gives a conservative lower bound for ranking.

  • โ—Bayesian shrinks toward prior when few reviews
  • โ—Wilson used by Reddit for comment ranking
  • โ—IMDb uses Cโ‰ˆ25,000, mโ‰ˆ7.0
โญ
STATISTICSDescriptive Statistics

Average Rating โ€” Weighted Avg, Bayesian Avg, Wilson Score for Star Ratings

Compute weighted average, Bayesian average, Wilson score. Distribution, mode, median, standard deviation.

Real-World Scenarios โ€” Click to Load

For educational and informational purposes only. Verify with a qualified professional.

Key Takeaways

  • โ€ข Weighted average: Rฬ„ = ฮฃ(rating ร— count) / ฮฃ(count) โ€” the standard way to combine star ratings
  • โ€ข Bayesian average shrinks toward a prior mean when you have few reviews โ€” prevents new items from ranking too high
  • โ€ข Wilson score lower bound gives a conservative ranking that accounts for sample size โ€” used by Reddit, IMDb
  • โ€ข Standard deviation and variance measure spread of ratings; mode and median complement the mean
  • โ€ข Distribution percentages show how ratings are spread across star levels

Did You Know?

โญIMDb uses a Bayesian-weighted rating formula to prevent new movies with few votes from outranking classics.Source: IMDb FAQ
๐Ÿ“ŠReddit uses the Wilson score to rank comments โ€” a comment with 10 upvotes can beat one with 100 if the proportion is higher.Source: Reddit
๐ŸŽฏThe Wilson score interval was developed by Edwin Wilson in 1927 for binomial proportions.Source: Wilson, 1927
๐Ÿ“ˆBayesian average: BR = (Cร—m + ฮฃ(ratingร—count)) / (C + total). C is the confidence parameter; m is the prior mean.Source: Wikipedia
๐Ÿ”ขA product with 5.0 from 2 reviews ranks lower than 4.5 from 500 reviews when using Bayesian or Wilson methods.Source: Evan Miller
๐Ÿ“ฑApp stores often use Bayesian averages to balance new apps (few reviews) against established ones.Source: Industry practice

Expert Tips

Bayesian Prior

Set m to your scale midpoint (2.5 for 5-star, 5.5 for 10-point). C: higher = more weight on prior.

Wilson for Ranking

Use Wilson lower bound when ranking items. It penalizes low sample sizes and prevents gaming.

Scale Choice

5-star is universal; 10-point allows finer granularity (IMDb, games). Custom scales for specialized contexts.

Distribution Shape

J-shaped (many 5-stars) vs bimodal (polarized) โ€” distribution % reveals more than the average alone.

Why Use This Calculator vs Other Tools?

FeatureThis CalculatorSimple AvgExcel
Weighted averageโœ…โœ…โš ๏ธ Manual
Bayesian averageโœ…โŒโŒ
Wilson scoreโœ…โŒโŒ
Distribution %โœ…โŒโš ๏ธ Pivot
5/10/custom scaleโœ…โœ…โœ…
Mode, median, SDโœ…โŒโš ๏ธ Multiple
Copy/share/AIโœ…โŒโŒ

Frequently Asked Questions

What is the Bayesian average?

A prior-weighted average that shrinks toward a default (e.g., 3.0) when you have few reviews. Prevents new items from outranking established ones.

What is the Wilson score?

A confidence interval lower bound for a proportion. Used for ranking: conservative estimate that accounts for sample size.

Why not just use the simple average?

Simple average ignores sample size. 5.0 from 2 reviews is less reliable than 4.5 from 500. Bayesian and Wilson methods correct for this.

How do I choose C and m for Bayesian?

m = scale midpoint (2.5 for 5-star). C: start with 10โ€“25. Higher C = more shrinkage; lower = faster convergence to observed mean.

What counts as a positive rating for Wilson?

Typically ratings โ‰ฅ 60% of max (e.g., 4โ€“5 stars on 5-star scale, 7โ€“10 on 10-point).

Rating Systems by the Numbers

5
Star Scale
10
IMDb/Games
1.96
Wilson z (95%)
1927
Wilson Paper

Disclaimer: Bayesian and Wilson parameters are configurable. Results depend on your choice of C, m, and positive threshold. Adjust for your use case.

๐Ÿ‘ˆ START HERE
โฌ…๏ธJump in and explore the concept!
AI

Related Calculators