Testing Your Hypothesis: Experiments you should do

In the always-ongoing quest to improve the beer they produce, homebrewers have become enamored with experimentation. Experiments about hops, experiments about malts, experiments about water, maybe even experiments about experiments! But there’s a hole in the boat . . . homebrewers are reading about experiments, but not many are actually doing experiments. Why let someone else tell you how to brew your beer?

Denny started doing homebrew experiments over 20 years ago, long before Experimental Brewing or our good friend Marshall Schott’s Brülosophy existed. And Denny certainly wasn’t the first homebrewer to experiment with processes and ingredients. Both of us remember sheaves of experiments being reported on the old “Home Brew Digest” and various USENET groups. Heck, the Tinseth hop bittering calculation that most of us use started as a backyard experiment!

When we started Experimental Brewing five years ago, our concept was to test the conventional homebrewing wisdom and find what worked and what didn’t. Brülosophy was started with the same goals. Homebrewers immediately glommed onto the findings from both of them in an effort to improve their beer. And while that’s a good thing, there’s a downside, too. It’s great that people are doing these experiments for us, but how do you know if the conclusions actually relate to your brewing? For example: That Tinseth formula – Glenn Tinseth admits the number is accurate for his homebrew system, in the 1990s, using whole hops. Its mathematical relationship to the rest of us is a bit fuzzier.

See, as enlightening and convenient as it may seem to have somebody experimenting for you, you need to keep in mind that it’s “citizen science,” not real science. Someone trying something and relating the results constitute a data point, not a scientific conclusion. Real science demands repeated trials under the same circumstances. Findings need to be confirmed by multiple researchers. I mean, have you read a scientific report that doesn’t end by saying “more study is needed”?

With Experimental Brewing, we tried to simulate this by having a group of experimenters each doing the same experiment before presenting the results to panels of blind tasters. The good part of that is that you had a group of people doing the experiments on their own equipment, giving kind of a real world point of view. The bad part is that you had a group of people doing experiments on their own equipment, which meant that it was difficult to maintain a controlled situation. In addition, while it’s great to have a wide array of data points, it’s also difficult to compile data given the variables.

Brülosophy takes a different approach. They have one person doing the experiment before presenting it to a panel of blind tasters. While it reduces the number of trials of the hypothesis, it also allows them to maintain much better control over experimental variables. It also makes interpreting the data much easier, but the lack of replicates does just make it a single data point.

Again, it’s very important to stress – that regardless of the approach, most of these homegrown experiments, regardless of source, are loosely scientific at best and serve as interesting pieces of data, but not rock solid valid experimental conclusions.

This is why it’s so important for you to do your own experiments. Only you know what the variables are in your own homebrewery and your palate. So why aren’t more people doing their own experiments?

It seems that one of the biggest issues is time. We’ve heard over and over again that someone doesn’t have much time to brew, so they want to go with what (they think) they know. They don’t want to take a chance of possibly producing beer they don’t like as a result of an experiment. Also, people have the fear of wasting money. Remember, a batch of beer is relatively cheap and learning is part of the fun!

Another reason is equipment and space limitations. Many experiments require brewing a large batch of wort and then splitting it to examine the effects of something like different yeasts or hops. That can require two brewing and fermentation setups, which need to be as identical as possible. Most homebrewers aren’t set up to do that.

Finally, let’s get real . . . another reason people don’t do the experiments themselves is laziness. If someone has already done the experiment and told you the result they got, why should you bother? It’s simply human nature to conserve effort. Let’s not forget that homebrewers can’t completely ride the science done by professional brewers — our small setups and volumes change a number of the variables from their efforts. It’s both freeing and frustrating simultaneously.

But if you truly want to improve your brewing, you need to see how things work out in your own brewery. Here’s an example . . . Does mash temperature make a difference in the finished beer? Maybe, it seems to depend to a great extent on what malt you use and how the maltster treated it. As far as we know, no one who has done this experiment has tried it on multiple brands of malt. That’s simply unworkable for homebrew experimenters. But you can do it to find out about the malts you use. In that sense, “science” becomes a personal thing and you can learn more about your favorite malts.

We have a few tips to consider before you start designing your experiment. First, you need to eliminate as many variables as possible. Unless you have perfect brewing control – Ha! — you’re going to need to find ways to eliminate brew day replication. If you’re doing a split boil, it’s a great help to have two identical kettles and heating sources. You also need to have identical fermentation setups. This might look like overkill, but it really does help eliminate the wibbly factor. If you have a truly automated repeatable system, you can do back-to-back brewing, but remember – wibbly is a thing.

Here’s our guide to finding your own truth:

Things to Consider Carefully

In designing an experiment, there are a few steps to go through before you whip out the burner and start brewing. The first, and most important, is “What do I care about?” Crafting a hypothesis is deceptively tricky. You need to think about, “Is my hypothesis falsifiable?” E.g. it’s clear when you’re wrong. And keep in mind that just because you get a positive result doesn’t mean you’re right! It may be something else.

Here’s an example of the steps in designing an experiment you can run:

Start with the Question: “Do decoctions matter?”
Then the Hypothesis: “Decoctions do not create a noticeable sensory difference over an infusion mash.”
Then the Protocol: Do two mashes – one decocted, one just taken through the rests. Ferment and package the same (unless one of these are your variable).
Evaluation: Check out our December 2019 column for recommendation on best practices here.

Got that? OK, then, here are some experiments we think you might find valuable. We’re sure you’ll come up with some of your own variables to test too.

Mash Experiments

Question: Does step mashing affect our beer flavor in a perceivable manner?
Hypothesis: Step mashing does not alter the body and flavor of a beer compared to a single infusion mash.
Brewing Sessions Needed: 2
Protocol:

1. Mash one batch of beer with a stepped mash, with rests of 145 °F (63 °C) for 30 minutes and 158 °F (70 °C) for 30 minutes.
2. Mash a second batch at 152 °F (67 °C) for 60 minutes.
3. Ferment both batches the same way: Yeast strain, temperature, and fermenter geometry must be consistent for both.

Evaluation: Perform the triangle and ranking tests, asking the tasters to rank the samples in order of most to least body, head formation, and head retention. Ask them which they prefer and why.

Question: Does adjusting your water chemistry for style produce a more enjoyable beer than the same recipe using unadjusted water?
Hypothesis: Changing the water profile for your favorite style of beer will not produce a different tasting beer.
Brewing Sessions Needed: 2
Protocol:

1. Choose a recipe that you enjoy and are familiar with.
2. Brew one batch with your normal water.
3. Get a water analysis and plug the values in the water calculator of your choice.
4. Choose a target water profile. Find them listed in your water calculator, a book, or on the Internet. Calculate your adjustments in the program.
5. Brew a second batch with your new, tweaked brewing water profile.
6. Ferment both batches the same: Same yeast strain, temperature, and fermenter geometry.

Evaluation: Perform the triangle and ranking tests, asking your tasters which sample most closely tastes like the style it’s supposed to be, as well as asking about body and sweetness. Ask them which they prefer.

Boil Experiments

Question: Will adding sugar to the kettle negatively impact beer flavor or attenuation?
Hypothesis: Adding simple sugar to the boil will cause yeast health and attenuation problems.
Brewing Sessions Needed: 1 (split with two kettles)
Protocol:

1. Brew a simple recipe that you trust and is balanced. We recommend a Belgian tripel possibly for this one. Split the wort evenly into two kettles. Add the sugar to one kettle at the beginning of the boil.
2. Treat the beers the same through to the middle of fermentation.
3. After 4–5 days, add sugar to the batch that didn’t receive it in the kettle. Let both batches ferment for another 1–2 weeks. Be sure to take a final gravity reading for each batch.
4. Package each batch the same way.

Evaluation: Perform the triangle and ranking tests, asking the tasters about sweetness and dryness as well as off-flavors that could be attributed to poor fermentation, such as sourness, phenolics, and diacetyl.

Question: How does the bittering from first wort hops (FWH) compare to the bittering from a 60-minute addition?
Hypothesis: Some studies have shown that FWH actually produces about 10 percent more measureable IBUs than a 60-minute addition, but it tastes less bitter.
Brewing Sessions Needed: 1 (split with two kettles)
Protocol:

1. Evenly split your wort into two kettles. Add your nominal 60-minute addition in one kettle as FWH before adding the wort. Note: We recommend Cascade hops for testing due to their noticeable, but non-dominating, character.
2. Steep the hops in the kettle while you sparge. Evenly split the sparge runoff between the two kettles.
3. Bring both kettles to a boil.
4. Add the same amount of the same hops as a 60-minute bittering charge to the other kettle after it comes to a boil. Boil both kettles for 60 minutes with no other hop additions.
5. Cool, pitch, ferment, and package.

Evaluation: Perform the triangle and ranking tests, asking tasters about the quality of bitterness (harsh, neutral, smooth) and hop flavor.

Fermentation Experiments

Question: Does the way you use multiple yeasts change the flavor?
Hypothesis: Pitching more than one yeast strain at a time yields unpredictable results. Pitching strains sequentially gives each a chance to develop flavors independently and improves repeatability for future brews.
Brewing Sessions Needed: 1 (split with two fermenters)
Protocol:

1. Split a batch of wort between two fermenters.
2. In one fermenter, pitch two different yeast strains. In the other, pitch only one of the two strains used in the first fermenter. Be sure to pitch the same amount of yeast into each. That means that the total pitched of the two-yeast batch will equal the amount in the one-yeast batch.
3. After three days to a week, pitch the other strain into the second fermenter (the fermenter that originally got only one strain). Pitch the same amount of this yeast as the original strain so that one will have less chance of overwhelming the other.
4. Let the ferments finish and package.

Evaluation: Perform the triangle and ranking tests, asking tasters to describe any differences in flavor, aroma, or mouthfeel. Does the flavor of one beer show more or less of a yeast flavor characteristic than the other?

Question: What is the effect of yeast pitch rate on yeast’s ester production?
Hypothesis: Pitching less yeast will result in fewer esters due to lack of acetyl-coA to create esters while it builds yeast cells.
Brewing Sessions Needed: 1 (split with two fermenters)
Protocol:

1. Save the yeast slurry from a 5-gallon (19-L) batch of beer in two sanitized containers. The easiest way to do this is to weigh the slurry so you have about 2⁄3 of the total in one container and 1⁄3 in the other.
2. Produce a batch of wort and split it evenly between two fermenters. Pitch one container of slurry into each fermenter.
3. Ferment, package, and serve. Take periodic specific gravity readings to compare the fermentation profiles.

Evaluation: Perform the triangle and ranking tests, asking the tasters about their perception of fruity esters in both the aroma and the flavor of the beer.

OK, that should be enough to get you going. Remember, the pursuit of these experiments isn’t empirical, universal truth — it’s the deeply personal truth of what fits you, your needs, and your desires. Think of the fast warmer lager techniques — it may work well for a great many (Drew: It does . . . for me!), there are plenty for whom there’s always a nagging sulfurous, lingering doubt that robs any chance to enjoy the beer. Maybe a 30-minute boil works for us, but your system has a gentler boil and you need longer to achieve boiled goodness. Those are questions we, and others dabbling in experimentation, can’t answer for you!

So . . . 3, 2, 1 — Let’s science!

Written by Drew Beechum & Denny Conn