Beer Review Methodology

With each beer review I try to judge the beer based only on beers of its same style. Sometimes that is very difficult (Kvasir, for example).

If one goes back and looks at all my ratings you can probably see which beer styles I like the best and which I like the least. I try to only review beer styles that appeal to me to counteract this bias.

Numerical Ratings

Each beer review is comprised of 5 numerical scores that are based on the 5 characteristics listed below. Each numerical scores ranges from 1 to 5. I include a brief description is each review category as well.

  • appearance = 5%, I note the beer’s color, carbonation, head and its retention.

  • smell = 20%, I note the beer’s aromatic qualities.

  • taste = 45%, I take a deep sip of the beer and note any flavors, or interpretations of flavors.

  • mouthfeel = 10%, I note how the beer feels on the palate.

  • overall = 20%, My overall impression of the beer.

At the end of the review I calculate the beer’s Final Rating by taking a weighted average of the numerical scores from each characteristic using the percentages indicated above. This gives a nice idea of the quality of the beer. It doesn’t really take into account the economy of the beer though.


There is a huge variety of really good beer out there. The “craft beer revival” that the US has experienced in the last 20 years, even in just the last 5 years, has resulted in a wide disparity between beers with respect to price and container volume. It is quite difficult to decide whether a great beer is “worth” its great big price tag. Everyone will have their own subjective opinion on whether Beer A is a better deal at price X than Beer B is at price Y, but my “aFR” or adjusted Final Rating (What can I say? I took a very literal approach to devising my acronym.) is my attempt to standardize beer ratings.

I calculate aFR as follows:

\[\begin{align*} aFR = \frac{Final_Rating}{\frac{Price}{Volume}} + sign(\beta)*\beta^2 \end{align*}\]


\(\beta\) = \(\frac{X-E[x]}{\sigma(x)}\)
\(\sigma(x)\) = \(\sqrt{Var(x)}\)
\(x\) = The set of all Final Ratings

See Wikipedia for more information.


This will result in a number that should make it a little clearer as to the relationship of a beer’s quality and price. My idea is that the aFR will tell one how “efficient” the beer is; how many Points you are buying with each Dollar (while also considering how much volume of that beer you’re getting). The aFR will increase as the Final Rating increases but it will decrease if the price increases without also increasing the quality above the average of all beers reviewed.  In theory this will help me decide if that Revolution Brewing Deth’s Tar at $15 will provide me a better bang for my dollar than the Toppling Goliath Brewing Kentucky Brunch Brand Stout at $22.

The trickiest (most valuable?) part of this rating system is that it is relative to all other beers reviewed. That means that as I review more beers and compile more data, the aFR of beers that I have already reviewed will change depending on how the new reviews affect the value of the average of all the Final Ratings. This is why the aFR on each beer review page is a dynamic field that will update every time I add a new beer to my database.

This system is clearly not perfect. The aFR is already on its second iteration. I’m always open to suggestions on how I might be able to better capture this idea mathematically.


For a list of all the beers I have reviewed (and their scores) check out the Summary of Beers Reviewed and Scores page.