A new way of thinking of the BU:GU ratio

General Homebrew Discussion
1

I was recently sent an email from Ryan Shwayder, who has kinda reconceptualized the way he thinks of the BU:GU ratio.  He calls his concept the “Relative Bitterness Ratio”.  Basically, the big change it to take the attenuation of the beer into consideration.  His way of thinking is that the higher the attenuation of the beer, the more the blance will veer towards bitterness, so he’s come up with a formula to account for that.  I’m hoping. that we can start a discussion about the concept and its applicability.  Part of the email I received from him reads…

“The short version is that I feel like BU:GU (Bitterness Ratio) doesn’t consider something that is easily quantifiable; attenuation.  I did a lot of work coming up with a good formula to adjust BU:GU based on your beer’s apparent attenuation relative to all other beers (in the BCJP style guidelines), and finally came up with something worth sharing.  You can find all the pertinent information linked here:
http://www.madalchemist.com/archives/relative-bitterness-ratio/”

I’ve sent him a link to this discussion so he can join in, too.  So, what do you guys think?  I’m gonna withhold some of my own comments until we get a discussion going.

2

very interesting. i am not sure it will matter a ton unless attenuation really changes significantly from about 0.76.  in the final taste of the beer.  in terms of comparing sweetness to bitterness, if indeed most of the residual sweetness can be related to the final gravity, then it may make more sense to develop a number based on the final gravity rather than the original gravity.  in addition think 'bout a beer with attenuation as low as 60 or as high as 90.  this accounts for about a 14 % shift in the bu/gu ratio. big change, but those are big shifts in the attenuation from the norm of about 75%. so anything that is between 66-86 % attenuation is within 10 percent of predicted based on bu:gu.

3

I think it’s an improvement over the BU:GU ratio, and I don’t want to diminish the quality of Ryan’s work, but it will have most of the same limitations.

  • IBU estimates can vary wildly depending on the formula used.
  • Attenuation is only one factor influencing the mouthfeel: carbonation can have just as much - or more - of an effect on perceived bitterness.
  • Different hops (and different boil times) have different perceived bitterness, despite having the same IBU.
  • I don’t personally feel that there’s a linear relationship between attenuation (or FG) and mouthfeel. In some beers a six-point variation is undetectable, while in other two points is a night-and-day difference.

I don’t bother to calculate a BU:GU ratio when designing a recipe, because I don’t think it will yield any information I didn’t already have.

4

BU:GU is a significantly flawed concept, but it can be somewhat useful and its darn easy to evaluate.  I’ve contended for almost a decade that attenuation needed to be in the mix somewhere.  I’m OK with the concepts that have been put forward to include attenuation.  I have no idea if they will prove useful in practice or if they agree with my personal perceptions, but its probably a move forward.

5

Thanks for getting the discussion started, Denny. My intent with spending all the time working on this article was largely to start a conversation on finding ways to predict and target beer balance. I’m glad the conversation is starting.

Ultimately, the best way to make the beer you want is to brew it once then make corrections in pursuit of your intent, but I’m always interested in methods to frontload that effort with calculations (the fewer repetitions needed to perfect a recipe the better).

The Mad Fermentationist (not me, I’m the Mad Alchemist!) emailed me and proposed that we start thinking along the lines of the residual extract (which I believe is the same thing as real extract, so that’s what I ran with).

To push this discussion further along, here’s some thoughts I sent him about that approach.

I added columns for Real Extract in both Plato and Specific Gravity to the far right of the data chart.

I originally used RDF (Real Attenuation) instead of ADF (Apparent Attenuation) for similar reasons. Ultimately, since I decided to adjust the Balance according to a beer relative to all other beers, using ADF vs. RDF became inconsequential (because your beer’s RDF in comparison to the average is the same as your beer’s ADF in comparison to the average). I switched back to ADF so people didn’t have to bother with RDF = ADF * 0.8192.

Real Extract is ultimately a similar approach, though might be more accurate in terms of perceived sweetness of the final beer.

If we were to decide to create a new number scheme from BU:GU entirely, I think it might yield some useful results (part of my decision to go with what I did for RBR was to make it more accessible since many people are familiar with BU:GU already, and the numbers are essentially the same).

Anyway, I also dropped in a thought on a potential method to calculate this under “Perceived Bitterness?” in the last column.

It is IBU/RE (°P)

International Bitterness Units / Real Extract (degrees Plato)

RE = (0.1808 × °Pinitial) + (0.8192 × °Pfinal)

Interestingly, this (IBU/RE (°P)) gives numbers roughly on a scale from 1 to 10, with the outlier being Imperial IPA at 13.4. It’s actually the only beer style above 10 using this equation.

I went ahead and colored the columns as well, which allows us to see that beers relative to one another look roughly the same as BU:GU and RBR.

6

Just an observation, but if you normalize to the value your equation produces for the high end of the Imperial IPA (1.3368) by dividing by that value, the numbers range from 0 to 1 with the low end of the American Pale Ale corresponding exactly to 0.5.

Interesting stuff.

7

Tough concept to try and put in a formula.  Since ADF tends to run between 70 and 80%, you’re only talking about a 5% difference with your formula.  I suppose theres the occasional 60% attenuation and those are more likely to come from a lower OG beer like common bitter where you want to leave some flavor and sacrifice alcohol to do it.

Then theres the difference in sweetness between beers with a common FG based on whether the residual “stuff” is sugar or soluble fiber or whatever.

I think the difference in sweetness between a 1.010 and 1.025 FG, is more significant than that.  I think the factor ought to somehow account for absolute FG.  Yes a higher OG means higher alcohol and that probably accentuates the bitterness as well, but a 1.020 IPA and a 1.020 common bitter are both going to be more alike in terms of how a given OBU presents.

8

Doesn’t Gordon Strong talk about the BU:FG ratio as being more of what one wants to target in his book?

9

Regarding the outlying IIPA, is that based on some extravagent IBU estimate?  I’ve seen IIPAs with IBU quotes of well over 100.  Unfortunately, the saturation limit for alpha acids and iso alpha acids prevent that much bitterness from actually existing in the beer.  On the order of 80 ppm iso alpha acid (80 IBU) is about the limit.  That is why IIPAs are typically more ‘balanced’ than an IPA or PA.  The gravity is so much higher than the bittering in the IIPA and the balance tips more malty.

10

The outlier (IIPA) is based on 90 IBU, which is the average based on BJCP style guidelines.

I’ve added some data to the chart for us to look at:

  • “BU:RE (°P)”: Bittering Units/Real Extract in degrees Plato
  • “BU:REU (SG)”: Bittering Units/Real Extract Units in Specific Gravity (Real Extract Units = (SG - 1) * 1000)
  • “BU:FG (°P)”: Bittering Units/Final Gravity in degrees Plato
  • “BU:FGU (SG)”: Bittering Units/Final Gravity Units in Specific Gravity(Final Gravity Units = (SG - 1) * 1000)
  • “BU:FG (SG)”: Bittering Units/Final Gravity in Specific Gravity

Light American Lager and Gueze both get a little broken because their average FG is actually less than 1 °P (unless you look at the Real Extract numbers, which look a bit better).

My favorite of the bunch is BU:RE (°P). The numbers make my brain happy because it is basically on a 10 point scale. The average of all beer styles is 5.4, so I suppose it could be corrected to make 5 be dead average, but that’s just manipulating the data to make it prettier.

The formula looks a little nasty if you don’t break it up into its constituent parts.

Effectively it is just IBU / RE in °P.

Real Extract (RE) = (0.1808 × °Pinitial) + (0.8192 × °Pfinal)

If you want to write the entire thing out into a formula using FG in Standard Gravity (since most homebrewers use SG and not °P)… Well, this is going to look nasty in forum text, but I’ll attempt to write it out.

(IBU / ((0.1808 * (259-259/OG)) + (0.8192 * (259-259/FG))))

Yeah, looks nasty. Hooray for calculators. This includes the conversion of SG to °P right in the equation for both OG and FG. It also includes the conversion from FG to Real Extract in the equation. It certainly looks like a mess, but the results are elegant and seem pretty accurate to me on a 10 point scale.

This still suffers from the need to either measure your FG or try to predict it accurately if using it to formulate a recipe. Perhaps it would be more useful when communicating with each other as to the balance of our beers rather than in planning a recipe.

“Here, I brewed up this Irish-American Red. Its BU:RE is around 7, so it’s more bitter than your average Irish Red. I lovingly call it the South Boston Red. What do you think?”

“Try my Imperial IPA. It’s got a BU:RE of 9, so it’ll taste a little less bitter than you might be used to in such a bitter style. I call it Archduke IIPA. Got enough bite for ya?”

11

I thought more about this over the weekend and I am starting to like the idea of using this as a method to communicate a finished beer’s balance instead of as a planning tool.

As a planning tool, BU:GU is something you can do very quickly and is pretty accurate.

Using BU:RE seems like a good approach to start with when talking about actual perceived balance.

The wall I’m hitting now is that if we want to find a common language to use when talking about balance, that means we ought to incorporate many other factors like percentage of roasted grain, chloride/sulfate ratio of the water, even the mash profile.

There are too many inputs to create a simple equation.

Ultimately, maybe it’s best to just give up and go with what we’ve been doing for years; finding a way to distill the essence of a beer into a short description for those we meet.

“A well-balanced IPA with big malt character, a smooth-but-pronounced English hop aroma and flavor, and pleasant undertones of freshly-baked bread.”

This might really mean:

  • Mashed at 156 F
  • Balanced Chloride/Sulfate ratio with ample Sodium for rounding out the flavor
  • Used Dry English yeast and fermented to keep some residual sweetness in play
  • Medium carbonation
  • Maris Otter base with some Victory/Biscuit malt and medium Crystal

Perhaps, then, it is worth agreeing upon certain descriptive keywords or phrases that can really build a solid mental picture of a beer before you even taste it.

While I know a lot of these are established in terms of actual beer judging, I’m more interested in what you could communicate to a layman or a less advanced homebrewer.

12

the problem i sometimes think about also is the aroma and flavor ratios and not just the bitterness to fg or og.  for example i can bitter my beer with 60min utilization of a small amount of hops or with 20 minutes of larger amount of hops. the ibu may be the same but the flavor aroma perception in terms of bitterness to malt is different.  that i think is also hard to put in to an equation.  i have still been pondering this since denny first posted your ideas.