Friday 2 November 2012

Forget IBU, think about BU:GU


What does 10 IBU, 50 IBU or 100 IBU actually mean to the drinker?

Not a lot.

Why?

Because it’s all about context. And I’m about to get geeky...

50 IBU (International Bitterness Units) will make a light lager brutal with bitterness. 50 IBU will make a strong stout seem cloyingly sweet.

It’s more relevant to look at the balance between bitterness and sweetness, so let’s throw away IBUs and bring in the BU:GUs (Bitterness Units: Gravity Unit).

BU:GU is the bitterness ratio and it was first introduced by Ray Daniels in Designing Great Beers (which is a great book!). It measures hop bitterness against malt sweetness where the higher the number, the higher the bitterness is to the sweetness in the beer. It’s calculated by dividing the IBU by the original gravity of the beer. So, with a beer of 25 IBUs and an original gravity of 1050 (you drop the first two digits of the OG unless you’re starting with a monster beer that’s in the 1100s, in which case you just drop the first digit) you get 0.5. On the BU:GU scale, that’s balanced – a score of 1 would be a beer with a very high perceived bitterness whereas 0.1 would be very low. (There’s a list of BU:GUs here and there's a colour chart of BU:GUs here)

Image from here

But then, to make it less straight-forward, another factor comes in: final gravity, or the apparent attenuation of the beer (the amount of sugars which are left in the finished beer). It’s fine knowing the original gravity but obviously a beer with a lower final gravity (or higher attenuation) will be drier in the finish with less residual sugars, which will then influence the perception of bitterness.

The Mad Alchemist has come up with a formula for figuring out the RelativeBitterness Ratio (RBR) against the same numerical scale as BU:GU (here's another colourful chart). Here’s what they say:

“ADF = Apparent Attenuation. 0.7655 is the average ADF of all beer styles. Since the Relative Bitterness Ratio takes into account balance relative to all beer styles, it uses this as a constant. You are comparing your beer's ADF against the average ADF (0.7655), then adjusting the standard Bitterness Ratio accordingly (it goes up if your ADF is higher than average, down if your ADF is lower than average). Just like BU:GU, higher numbers mean more bitter, lower numbers mean less bitter, and 0.5 is roughly average balance.”

Here’s the sum for a beer with 25 IBUs, OG of 1050 and 80% apparent attenuation:

RBR = (BU:GU) x (1 + (ADF - 0.7655))
RBR = (25/50) x (1 + (0.8 - 0.7655))
RBR = 0.5 x (1 + (0.0345))
RBR = 0.5 x 1.0345
RBR = 0.51725

So if you have two beers with the same IBU and same OG but one has a high apparent attenuation (85%, for example) and one with a low apparent attenuation (55%), they perception of bitterness will be different in both. Get it?

We know the IBU scale and understand that 10 is very low and 100 is very high, so it’ll be just as easy for drinkers to take notice of a BU:GU scale where 0.1 is very low and 1.0 is very high. While IBUs are an interesting piece of information, it means little without measuring it against the sweetness in the beer. That’s where a BU:GU score, worked out according to the
Relative Bitterness Ratio, is more relevant and revealing.


Wonderful top photo from here

28 comments:

  1. Two beers with the same apparent attenuation can be made very differently. One could be very dextrinous and another very sweet. The former is done via mash temperature or carapils/dextrine malt, and the latter via heavy doses of caramel malt or an underattuative yeast. Two totally different things, perceptively.

    ReplyDelete
    Replies
    1. Not sure that can be thumbed into the equation, can it?!

      Delete
    2. Nope. Just pointing out another variable. I should have said two brews with the same OG and same apparent attenuation.......
      Daniels spoke of this ratio in the mid 90s, but we assume that people are unaware of this concept. Most just see high IBU levels, gravity, etc. Balance, as a generic adjective, has always been about BU:GU.

      Delete
  2. You're right that IBU in isolation can be very misleading, the Oxford Companion to Beer is very good on this :"...the IBU is a laboratory construct that was never meant to leave the laboratory".

    Graham Wheeler's Home Brewing book from 1990 had a similar idea to the BU:GU ratio called "Wheeler's Bitterness Constant".

    ReplyDelete
  3. That's quite sophisticated and a bit complicated. A homebrewers rule of thumb is that the IBU should match the last two figures of the OG reading, so that a 1040 OG should be buttered to 40 INC. It's a pretty good starting point.

    ReplyDelete
    Replies
    1. That seems high - it means you're aiming for a RBR of 1. Using the charts provided, only Imperial IPA gets to 1. Plus, you won't want the same pre-set value of bitterness in a Hefeweizen and an IPA, would you?

      Delete
    2. You're right, but it's just a rule of thumb. And it turns out fairly balanced beers in my experience.

      Delete
    3. Hmm, looking at the sums again, and the stats for Sierra Nevada Pale Ale, makes my experience looks like bollocks: http://www.sierranevada.com/beers/paleale.html

      Delete
    4. Beer should never be buttered, regardless of OG.

      Delete
  4. That's probably not a good rule of thumb for a mild.

    ReplyDelete
    Replies
    1. No, it's a good rule of thumb, because you know that mild is less hoppy, so you shoot for less IBU vs OG.

      Delete
    2. So that would be a rule of thumb which we tend to ignore? Some rule. The point of the BU:GU ratio which Daniels popularised (in his good book, which you should get, if you haven't) is that the numbers fall into fairly tight groups for beers broadly classified together. i.e. if you're designing a mild (for instance) you pick your starting gravity, plug that into yr simple formula and out pops the BU you'll be aiming at. This is precisely the same point that Wheeler made in his books - for a rather narrower range of styles. Of course, given the variation in hop supplies, measuring and utilisation we get when brewing on a nano (home) scale, we won't be particularly close. But for brewing to style, it's probably a useful tool. And since attenuation is (broadly) consistent for a style, "correcting" for attenuation with any more complicated calculation is probably a waste of time. If you're not brewing "to style" then a rule of thumb like your BU=GU will give you reliably bitter beers. Perhaps rather too bitter for some, particularly at the low gravity end.

      I'm not sure there's any single number, no matter how fiendishly cunning the calculation, that can adequately sum up perceived bitterness or balance, so I shouldn't think we'll ever see beers labelled
      RBR = 0.51725

      Delete
    3. I'll second your point about correcting for attenuation. In practice it rarely adjusts the ratio outside of the error that you can expect from measuring equipment (at least homebrewing measuring equipment, fancy pro lab kit is another matter)

      This, plus I think BU:GU is at its most useful when considering the difference between styles, rather than between two beers of the same style, makes me think that in most instances where BU:GU is useful, doing the apparent attenuation adjustment is not going to change the outcome, it simply adds complexity.

      Delete
    4. Stringers - On your last point: a considerable number of breweries list the IBU of their beer on bottles or online. Because we've come to learn what the scale means, these are used and mean something in theory. If drinkers start understanding that a BRB of 0.9 is high, then it'll be just like listing IBU or colour units or even ABV.

      Ben - Interesting... It's good to see the difference between styles, I agree, but it's also interesting to compare beers within the same style range. IPA would be most interesting, I reckon: line up 10 popular American IPAs and figure out the BU:GU and I bet there'd be a fair range...

      Delete
    5. I gave it a quick go... and, as it happens, no.
      The top 10 (according to Beer Advocate FWIW) American IPAs (excepting 2 outliers), range from 65 to 77 IBU and by strength (excepting 2 outliers) from 6.9 to 7.2 (abv).

      Attenuation, where you can work it out, is about 82% (i.e. pretty much exactly what you'd expect from US-05)

      BU:GU (excepting two I couldn't find IBU for and one clear outlier) of the top 10 = 1.1 +/- 0.1
      i.e "RBR" = 1.15 +/- 0.1

      Im my arithmetic's right, it's a pretty tight group of beers, apart from the odd one.

      Delete
  5. Some interesting and eminently sensible concepts here, which if adopted on product labelling could give the drinker an idea of what to expect with the beer.

    ReplyDelete
  6. I think IBUs are often very misleading though, is it not more just a rule of thumb that e.g. IPAs should have hops and Milds shouldn't. Isn't bitterness a more complex perception that the amount of iso alpha acids? Don't the beta acids play a role in the perception of bitterness? Thinking about Mikkeller's 1000 IBU light and some of Steel City's efforts - when I drink them they never actually taste that bitter. But next to those a Cannonball tastes horribly bitter, when probably even the actual IBUs in that are less than the higher ones.

    ReplyDelete
    Replies
    1. Beta acids are much more volatile, so hops used in early editions for bitterness will have very nearly all the beta acids blasted away in the boil, thus having little impact on bitterness. The difference between beers is probably down to perceived sweetness.

      Because bitterness is balanced on the palette with sweetness, when something is sweet and bitter in equal measure, it is not *perceived* as bitter. This is captured in the BU:GU ratio (though as an aside. '1000 IBU' is a bit of nonsense. There's a variety of lab work that has shown both that people can't perceive more than about 100 IBUs and that somewhere around that level you can't actually isomerize the alpha acids further...)

      Delete
    2. Brother - The whole science of tasting is obviously more complex than a graph of values can plot out. It's all about perception, as Ben says, which is where BU:GU becomes an interesting way of trying to put a number against a sensation. But it is just that - number vs experience.

      Delete
  7. Preach the gospel! I'm such a fan of BU:GU, it's one of my favourite ways to explain different styles to my less beer-soaked friends (I hang out with nerds, but not all of them are beer nerds...)

    One fiddly numbers point. You say:
    "(you drop the first two digits of the OG unless you’re starting with a monster beer that’s in the 1100s, in which case you just drop the first digit)"
    Strictly that's true, but it's unnecessarily wordy. To convert from Original Gravity (OG) to Gravity Units (GU), just drop the leading 1. That's all.

    1.050 -> 050
    1.110 -> 110

    That leading zero takes care of itself.

    Maybe that helps?

    ReplyDelete
  8. I think these measures are interesting for beer geeks but offer no real information to the casual drinker looking for something they might like.

    The cyclops information is more useful if you are looking at a label and wondering whether the grog is drinkable or not.

    ReplyDelete
    Replies
    1. Cookie - I agree. But hopefully one day those casual drinkers might pick up a beer they like and see that it says 50 IBU (or whatever) and that might pique an interest in learning more.

      Delete
  9. So what you are talking about is balance? Something that most of the so called 'craft' breweries in UK fail miserable at. Brewdog being at the top of the list!

    ReplyDelete
    Replies
    1. It's about measuring balance, not about the production of it. I make no comment on the relative balance and quality of beer! But I do happen to think that Punk is a well balanced IPA.

      Delete
    2. Good beer is about balance and any brewer worth his salt should strive for this.

      Punk good? the last four cans I've tried have been infected, not just with lactic but brett too, god knows how it got there. I actually agree Punk is ok if its not infected. We really need to stop thinking that just because something is 'craft' it's a fantastic beer.

      I really think the British craft scene is suffering from a severe case of emperors new clothes syndrome . Craft brewers can turn out infected and/or totally unbalanced beers or even lagers which aren't much better than Carling and the punters will still reckons its the best thing since sliced bread.





      Delete
  10. Funny you used as an example a chart which lists Balance Value (BV) but didn't go into BV at all. As a homebrewer, I always look at my BV in preference to BU:GU, as BV takes attenuation into account....

    ReplyDelete
    Replies
    1. I did talk about the RBR, which is the adjusted BU:GU with attenuation in mind. Makes it more complicated having a few similar things around the same measurement though, doesn't it!?

      Delete
  11. Great blog. Enjoyed reading your posts and surfing the blog.

    ReplyDelete