This page explains how we measure beer quality and popularity so that highly rated beers with few check-ins can be compared fairly to widely distributed beers with tens of thousands of ratings. The system is inspired by baseball’s Wins Above Replacement (WAR). The metrics wOBAR (weighted Overall Beer Average Rating) and Beers Above Replacement (BAR) originated from Beergraphs.com; we use and extend that framework here.
count = floor(0.70 × listCount), minimum 1).So replacement level is “the worst beer in the set of beers that get most of the check-ins in that style.”
How much better than the style floor a beer is:
wOBAR = Average Rating − Replacement Level
Example: If the replacement level for American IPA is 3.4 and a beer’s average rating is 4.1, then wOBAR = 4.1 − 3.4 = 0.7.
We use base-10 log of check-ins so that volume doesn’t dominate and small beers still matter:
Volume = log₁₀(# of check-ins)
We scale so that a “perfect” beer (5.0 rating, 1,000 check-ins) equals 15.0 BAR (analogous to Babe Ruth’s 1923 WAR). Scale depends on the style’s replacement level:
Scale = (15.0 / log₁₀(1,000)) / (5 − Replacement Level)
Since log₁₀(1,000) = 3:
Scale = 5 / (5 − Replacement Level)
If (5 − Replacement Level) ≤ 0, the scale is guarded (e.g. no division by zero).
BAR = wOBAR × Volume × Scale
This is the main “quality × popularity” metric before any adjustment for low check-ins.
To give low-check-in beers a fairer shot, we add an Adjustment Factor (AF) to volume.
Adjustment Factor (AF):
AF = (C / log₁₀(check-ins + E)) + 1
Defaults: C = 5, E = 1.
Adjusted Volume = log₁₀(# of check-ins) + AF
Then the Hall Rating (TAP / ABR) is:
ABR = wOBAR × Adjusted Volume × Scale
So: same wOBAR and Scale as BAR, but Volume is replaced by Adjusted Volume.
Interpretation of AF:
log₁₀(check-ins) is the raw volume.(C / log₁₀(check-ins + E)) + 1 is larger when check-ins are low, so low-volume beers get a volume boost.Example (C = 5, E = 1):
| Beer | Check-ins | log₁₀(check-ins) | AF (approx) | Adjusted Volume (approx) |
|---|---|---|---|---|
| Beer A | 10 | 1 | 5.81 | 6.81 |
| Beer B | 1,000 | 3 | 2.67 | 5.67 |
So a beer with 10 check-ins can have a higher adjusted volume than one with 1,000 check-ins, illustrating how the adjustment “levels the playing field” for underdogs.
BAR (and similarly ABR/TAP) can be interpreted using the following ranges:
| TAP range | % of population (approx) | Descriptor |
|---|---|---|
| 10+ | 10.26% | Hall-of-Famer |
| 8 to 10 | 1.01% | MVP |
| 6 to 8 | 0% | All Star |
| 4 to 6 | 0% | Very Good |
| 2 to 4 | 0% | Above Average |
| 0 to 2 | 86.13% | Useful to Average |
| Below 0 | 2.6% | Not Good |
ABR = wOBAR × Adjusted Volume × Scale.BAR = wOBAR × Volume × Scale. See also the BAR scale table above.Scale = 5 / (5 − Replacement Level) (with guard when denominator ≤ 0).Style+ = (Beer’s rating / Average style rating) × 100. Good for finding lesser-known beers that rate highly within their style.wOBAR × Adjusted Volume × Scale. This is the metric used for Hall of Taps rankings.log₁₀(# of check-ins).Average Rating − Replacement Level. Quality relative to the style floor.If you want to compare a personal rating to the style average:
Relative Beer Score = (Your Rating / Average Style Rating) × 100
Example: You rate an American IPA 4.2; average for the style is 3.8.
Relative Beer Score = (4.2 / 3.8) × 100 ≈ 110.5 (you rated it about 10.5% above the style average).
This formulation keeps the Hall of Taps math consistent across the site and any backend calculations.