Chances are, the only people that enjoyed my last post on the “Best” Beers in America relative to their respective styles were stat geeks, completist Aleheads fans, and Doc. Nevertheless, I enjoyed delving into the “numbers” and what they revealed about the way “we” rank and rate different styles of beer.*
*And by “we”, I mean “the good people at BeerAdvocate that do all of the work which allows me to waste so many hours sifting through their data for no apparent reason.”
The gist of the last post was, ostensibly, to figure out which beers surpassed the other top-ranked beers in their respective styles by the widest margin. My thinking was that this approach would help eliminate the notorious style-bias for high-profile beers like Imperial Stouts and IPAs. It’s impossible to normalize the data to eliminate those kinds of biases (after all, the raw data is entirely subjective to begin with), but I figured if a brewery made a truly outstanding version of a generally plebeian style of beer, they should get some recognition for their accomplishment.
It was a fun little project and it left me with, what I thought, was a solid list of very good beers that demonstrated that any style of beer can be made “great” in the right hands. While my goal was to determine a Top Ten list of American beers, there were some other little statistical items that jumped out at me while I was working with the data. I decided to share some of those tidbits with our readers just in case you can’t get enough beer-related number crunching…
I completely ignored lagers since the numbers were so skewed if you included them. Because craft brewers generally give short shrift to lagers (for reasons too numerous to elucidate here) and because Aleheads tend to look down on them, most lagers receive fairly average ratings on BeerAdvocate. As such, when a brewer DOES take the time and effort to produce a serviceable lager, it quickly rises to the top of the charts. Let’s face it, if your favorite craft brewery takes the time to produce a drinkable, flavorful Light Lager, there’s a pretty good chance it will blow away the competition. Even if it’s not particularly good, it’s still probably better than the rest of the brews within that style.
But maybe you’re interested in which lagers “would” have made my list if I had included them in my original post? No? Well tough shit. I’m telling you anyway. The Top 5 beers in that original list would have all been lagers. In order, they are:
- Sam Adams Light (21.12% above the average of the Top 50 Light Lagers)
- Cisco Sankaty Light (19.71% above the average of the Top 50 Light Lagers)
- Moonlight Death and Taxes (17.85% above the average of the Top 50 Euro Dark Lagers)
- Schlitz Gusto (16.87% above the average of the Top 50 American Adjunct Lagers)
- Moonlight Reality Czeck (15.75% above the average of the Top 50 Czech Pilsners)
Rounding out the Top Ten Lagers respective of their styles are:
- Short’s Black Licorice (13.55% above the average of the Top 50 Euro Dark Lagers)
- Pretty Things American Darling (12.71% above the average of the Top 50 American Pale Lagers)
- Zion Canyon Jamaica-Style Lager (12.62% above the average of the Top 50 Light Lagers)
- Snoqualmie Summer Beer (11.15% above the average of the Top 50 American Adjunct Lagers)
- Brooklyn Lager (10.77% above the average of the Top 50 American Amber Lagers)
Just like the ales, there were only a handful of American beers (just 16 actually) that were 10% (or higher) better than the average of the Top 50 beers within their style. That seems to be a pretty good cut-off point to keep in mind. If a brewery makes a beer that rates 10%+ higher than the other “best” beers in a style, chances are they’ve made an incredible version of a generally less-respected style of beer. So if you usually shy away from Czech Pilsners, you should probably sample something like Moonlight’s Reality Czeck before you claim that you “hate them all”. Just because there aren’t a lot of great Czech Pilsners out there doesn’t mean that there aren’t “some” amazing offerings in that style.
All that said, you can probably see why I avoided lagers in my initial list. Sure, Sam Adams Light is spectacular for a Light Beer, but it’s still not very good. I suspect the same is true for the other Light Beer representatives and the American Adjunct Lagers on their list. As pompous as this sounds, I just don’t see any way to make a “great” version of either of those styles. Better than average…sure. But great? I don’t think so. That’s not to say some of the beers above (like the Death and Taxes, Reality Czeck, Black Licorice, American Darling, and the always-serviceable Brooklyn Lager) aren’t worthy beers. But to include them on my list, I would have to make room for the Schlitz Gusto and Zion Canyon Jamaica-Style Lager. It just wasn’t worth it.
In order to calculate which beers outperformed other offerings in their style by the widest margin, I had to calculate the mean for each style. I used BeerAdvocate’s Weighted Ranks as my raw data (to see how BA makes these calculations, click here and scroll down to the bottom). While the means themselves were just a “means to the end” (I’ll see myself out, thank you) for uncovering my eventual Top Ten list, they’re actually fairly illuminating in their own right.
What are the Top Ten styles in terms of the mean for the Top 50 beers on Beer Advocate?
- American Double / Imperial Stout – Mean: 4.34
- Russian Imperial Stout – Mean: 4.30
- American Double / Imperial IPA – Mean: 4.28
- American Wild Ale – Mean: 4.26
- American IPA – Mean: 4.22
- American Barleywine – Mean: 4.14
- American Porter – Mean 4.13
- Belgian Strong Dark Ale – Mean: 4.12
- Saison / Farmhouse Ale – Mean: 4.10
- American Strong Ale – Mean: 4.09
Why is that interesting? If you asked a reasonably knowledgeable Alehead to name the ten most “prestigious” styles (or, in a more negative light, the ten most “over-rated” styles), chances are their list would almost perfectly match this one. Imperial Stouts, Imperial IPAs, Wild Ales and IPAs easily topping the charts? Absolutely no surprises there. So why is that illuminating if I already assumed that was going to be the case? Because sometimes it’s nice to have the numbers confirm my uninformed opinion…that’s why, asshole. I’m so rarely correct in my assumptions, let me have this moment.
You can also make some interesting assumptions by looking at the standard deviations within each style. For those of us who haven’t taken a stats class in a few years, standard deviation measures the variation between the mean of a data set and its data points. So a higher standard deviation indicates that the data is spread out a lot more while a lower standard deviation indicates that the data points are generally closer to the mean. In the case of beer rankings, standard deviation gives a good sense of the consensus (or lack thereof) regarding a particular style.
For example, the lowest standard deviation within any style of ale was for Winter Warmers. This probably shouldn’t come as a huge surprise for any Aleheads. Winter Warmers are generally good, but not transcendent, and most Aleheads enjoy them without loving them. You would be hard pressed to find any Alehead who would list the Winter Warmer style as either their favorite or least favorite type of beer. As such, those beers tend to be rated within a fairly narrow range of grades on BeerAdvocate. How narrow? Every beer in the Top 50 list for Winter Warmers received either a B+ or an A-.
Compare that to Rye Beers. Rye beers are wildly hit or miss. Some breweries absolutely nail the style (like Founders and their phenomenally sessionable Red’s Rye), while others produce a peppery, unbalanced mess (try a Sam Adams Revolutionary Rye for a poorly-executed example). In between there’s a ton of variance. The Top 50 Rye Beers range between a straight A (for Great Lakes Rye of the Tiger) to a straight C (for Weston’s SunRyes Ale).
Standard deviations matter in terms of my analysis because it was far more likely that one of my Top Ten beers respective to style would come out of a style with more statistical variation. The top-ranked Winter Warmer just didn’t vary a whole lot from the mean for the style. Ditto IPAs and Saisons….just not a lot of variability there.
But of the twelve beers on my Top Ten list (just read the post, it will make more sense), 11 of them came from styles that fell in the top third of ALL styles of ale in terms of standard deviation. In other words, almost every beer I included on my list came from a style with a ton of variation in terms of quality. The only beer that fell outside of that group was the Tröeg’s Nugget Nectar, which again proves how far superior that beer is to other American Ambers. Though, in defense of the American Amber style, the Nugget Nectar is pretty much an Imperial Red Ale in all but name. It’s like comparing Dartmouth College to other small, liberal arts schools…they may technically be in the same category, but one of these things is not like the others.
Of course, considering my methodology, it would have been surprising if my top beers DIDN’T come from styles with a lot of inherent variability. I was comparing against the mean, so that’s exactly what you’d expect. What I enjoyed most about comparing standard deviations was finding styles with relatively “low” means that still had small standard deviations. You would expect low standard deviations for a style like American Double/Imperial Stouts…after all, almost every beer on that Top 50 list is rated A+ or A. I was more interested in finding styles like the aforementioned Winter Warmer which aren’t necessarily highly-rated but seem to fall within a tiny range.
Saisons for instance, had a microscopic standard deviation amongst the Top 50 beers. A quick check on BA reveals that of the Top 50 Saisons, 5 are rated “A”…and ALL the rest are rated “A-”. Seems like Saisons are a style that Aleheads absolutely love, but don’t LOVE. Witbiers are in the same camp…other than one sparsely-rated (just 12 reviews) straight “A” Witbier from The Alchemist Pub, every Witbier is an A- or B+. Perhaps my favorite example of this trend is the Scottish Ale…a style not generally given much respect in the beer world. There are two very sparsely-rated brews (14 and 11 ratings respectively for the Ballast Point Piper Down and COAST Red Legs…just above the 10-rating cut-off for the Top 50 lists) that receive A- ratings. Otherwise it’s a B+ or straight B for every Scottish Ale on the list. Just as I suspected…Scottish Ales are perfectly adequate, and generally no more.
I hope you find this stuff as interesting as I do. If not, I’ve just wasted another 15 minutes of your precious time. Ah well, when it comes to our site, you get what you pay for. Thanks, as always, for reading. Until next time…