Lager Fermentation and Brülosophy Exbeeriments

[lupulus]

[Techbrau]

Correct me if I'm wrong, but I think what you're getting at with respect to the objective score argument is illustrated by this picture I drew to make a similar point a while back on HBT (forgive my lack of artistic talent)

brewing_space.png (355.82 KiB) Viewed 12306 times

Imagine you're walking around this desert blindfolded and trying to reach the highest point of elevation. You can't see the whole landscape at once, so instead you resort to doing a bunch of "experiments": you take a single step from your current position (in some direction) and see if you go uphill or downhill.

Obviously, whether or not you go up or downhill depends on where you already are. In certain parts of the space, no matter what direction you step, you don't seem to really go up or down all that much. In other parts of the space, the slope is much steeper.

So, when somebody tells you "I did an experiment by taking a step Northwards and I didn't go uphill or downhill, so it looks like walking North won't move you uphill or downhill", it would be useful to know what part of the desert they're in.

[lupulus]

I read you post yesterday morning and made me think quite a bit Yes, I think it is a fantastic way to stress the concept that "you need to know in what part of the desert you are" when doing a beer experiment. I may "steal" it one day (with proper credit of course).
If the control beer is in the desert floor or even in the plateau of mediocrity, it would be much harder to detect a significant difference, than if your control beer is in Valhalla, and as a researcher it is one's obligation to try to know where you are with that beer. In research lingo, it is one of the steps one must take to minimize the odds of a Type 2 error.
Ways to operationalize this obligation are for example entering the beer in 1-3 competitions, in Germany, use the DLG beer test, assemble a panel of judges that you are certain will evaluate the beer objectively. It is not perfect but it is at least a measure of control.
I would call this to VALIDATE YOUR MEDIA.
In addition to this media validation, sometimes you need to VALIDATE YOUR METHOD. Although I never force carbed a beer in my life (natural carb or spunding), I can concede that for an ale it is a widely used process. However, when your media is a lager beer, if you want to use an uncommon method, at minimum you need to validate your method against one of the many lager brewing methods. In these experiments, the BRU team not only used their common ale method but used a lager fermentation temperature schedule that is unproven.
BRU has not even started addressing the RELIABILITY issue. Can you make the same beer over and over again (yes, the same GREAT beer). If your media changes often, and if you let your media be dictated by what you need to brew for a given event, reliability becomes an afterthought.
Last but not least, DESIGN TO FIND DIFFERENCES. Your media (beer style) and your treatments must be carefully selected to maximize the odds of finding a difference, should there be one. This is a controversial topic, even among scientists, but to scientifically prove similarities requires an effort that is orders of magnitude larger than the effort required to prove similarities; so unless you are willing to undertake the effort to prove a similarity, design to find differences.
Finally, DO NOT IGNORE THE LITERATURE. Scientists do not reinvent the wheel. Many times, scientists hypothesize that the literature is wrong, and that is part of science, but to prove this wrong, they repeat the experiment/s they think is/are incorrect; scientists do NOT start all over again from zero.
Last but not least, a comment on the "we are just homebrewers" argument for bad science. The scientific process has nothing to do with the complexity of the instruments at your disposal. Indeed, there are experiments one cannot do if one does not have the instruments, time or resources, but one can still do great science with "homebrewing" equipment, time and resources. Experiments we reproduce in high school chemistry or physics lab are no less science than experiments run at the CERN supercollider; science is just about following the scientific process, nothing more.

[Weizenberg]

I explain the very same graph to novice programmers. The only difference in mine is: There is no mountain of greatness. We just deal with the unacceptable and the mediocre (randomness, combinatoric explosion etc... all things the human brain appreciates little).

I wonder whether the lesson learned would not apply to the subject of brewing. After all, there is similarly plenty of irrationality there too.

[Owenbräu]

I don't think consideration for thresholds is considered enough in their experiments. If you are willing to accept their single data point that simply reflects the results of what they did, then that is fine. However, it has very limited relevance for the purported purpose of the experiment. I find this applies to both recipe and process.

The analytical data for the changes in ester and fusel production for different lager strains at different temperatures is well established. Increasing temperature results in increased levels. Now, whether one can detect those compounds at those levels, and how the levels of one compound affect your sensitivity to another compound are not as well fleshed out. Then, multiply that by hundreds of compounds in beer. Point being, focusing on the temperature differential is meaningless if both temperatures produce beers with compounds above threshold. No difference is (likely) expected, unless extreme concentrations are produced. This is a case where the absolutes may be more important than differential and where having a classic fermentation profile to serve as control is essential.

I had this debate with a homebrew friend who wanted to do an experiment on ester production in hefe. He kept saying that as long as there was "enough" temperature difference between the two fermentations, then you should notice a difference. My argument was, if both temperatures product detectable levels, then you may commit a type II error and incorrectly reject the null. This all depends on the hypothesis, too. If the hypothesis is temperature affect esters production, then you may falsely reject. If the hypothesis is ester production will be affected between two temperatures, then it may be valid to reject the null. Again, some questions are very limited and some have broad applications; careful consideration of experimental design is warranted based on the hypothesis being tested.

Apply this to both recipe and other processes. If you mask your beer with high hopping rates, then are you really going to detect differences in the variable you are testing? Or, are you simply asking if a variable makes a difference in a very specific case? Enter Tech's argument on nuance. Similarly, if you introduce oxygen during the cold phase, then you essentially both erase a percentage of the flavor compounds you may have preserved during the hot side as well as generate higher levels of oxidized compounds. Was it enough to eliminate threshold differences? Or, did you simply generate a limited data point?

I'm certainly not against saying "with this style, this process and set of parameters, variable X does or does not make a noticeable difference". But, I also don't necessarily see the value in asking such limited questions, unless one is really dialing in a recipe to their own tastes.

[lupulus]

It seems we are all in the same page
To summarize, here are our suggestions for future experiments:
- Research/ read the literature
- Most likely effects would be esters and fusels, sulfur?
- Your control beer must follow a classic lager fermentation profile and must minimize esters and fusels.
- Give yourself the best chance to pick up differences (ie, brew a helles, do not brew a pils) (see Tech comment on nuance )
- Avoid flavor and aroma hops for maximum effect (pick up ester and or fusel differences)
- Avoid cold-side oxidation
- Do not use/add processes uncommon to lager fermentation (eg, force carbonation, gelatine fining)
Anything else?

[Techbrau]

To be honest, I think these kinds of criticisms and ideas just go in one ear and out the other with them.

Their goals are just completely different from our own. We’re in our own little corner of the web trying to make 49 or 50 point beers that hold their own against the best Bavarian examples served vom fass. When your benchmark is Firestone Walker Pivo Pils (to my taste a 31 or 32 point beer at best, and completely misses the style mark), what motivation do you have to improve your process, if you’re already making something you’re satisfied with? At that point, why not focus your efforts on trying to figure out what matters and what doesn’t with respect to your own process, so you can simplify your brew day?

I think they’re more than happy as is with all the attention and endorsement deals they get. The reality is that we represent the minority point of view in the homebrew/craft brewing world.

[lupulus]

Amen

[bjanat]

The latest one with märzen and fermentation temperature confirms what has been pointed out here. Comparing warm and cold, but not following a traditional cold fermentation process.

[bjanat]

News are in: step mashing matters. Good to know http://brulosophy.com/2016/04/04/single ... t-results/
EDIT: old post, saw it just now.