The Brew Matrix Experiment #2 - Mash Fines in the Boil x Kettle Trub in the Fermenter

[CascadesBrewer]

There was some discussion way up in this thread about the pros and cons of taking the visual away from the tasting. If the focus was only flavor, that would make sense but everyone drinks beer in clear glasses. How does process affect the beer in the glass with all things considered? I'm not suggesting that the clear vs hazy appearance didn't affect the overall rating. It probably did, because it generally does. I have enough of the beer left to have all the club's judges taste them blind next month and I'll publish that data when it happens.

Yeah, not questioning the process for this round. I tend to think that factors like appearance and fermentation performance are valid factors that should be taken into account. I am more curious if the impact of the Clean Boil is more a slight reduction in haze or actually makes a perceptible impact on the character of the final beer. It would be great if you collected and shared that data.

I have been brewing BIAB almost 5 years. My process is often not much different than your DBDF batch. Sometimes my beers drop crystal clear quickly, and sometimes they hang on to a bit of haze that takes several weeks to drop out, though they do tend to be more clear than what many of my friends make.

 

[Bassman2003]

Thanks for the great experiment and video Bobby. However anecdotal, tests like these push the conversation forward while possibly washing away some dogma. With my pursuit of low oxygen brewing, I have gone very far into the clear everything camp. The sad thing about homebrewing compared to pro brewing is that we do not brew often enough to really see trends and fix them. I have had some fermentation challenges with lagers but lag time was not the issue. I am totally sold on clear wort into the kettle. The use of Brewtan-B in the mash helps this even more. I am a little more open on less clarity into the fermenter, but I pause when it comes to stale flavors for longevity and yeast repitching.

Since going down the rabbit hole, I have noticed every beer has a more "clear" flavor. Meaning less mud. I have to attribute some of this to focusing on clarity through the process. If the difference is throwing in some yeast nutrient vs carrying over trub, I think I would still choose the yeast nutrient until more testing was done. Logically, the idea of just clear wort and yeast sounds like it would produce the more natural beer. But like you said, it is best to follow the data.

 

[mashdar]

It's also interesting that your results align with the much maligned Brulosophy Exbeeriment showing a net benefit of trub in the fermenter when it comes to finished beer clarity, and subtle differences in taste with some tasters preferring the high trub beer, and some preferring the low.

The result doesn't seem particularly surprising, though. Dumping a bunch of cold break into the fermenter is almost the same as dumping a bunch of gelatin in. (Especially if kettle finings were added.) Precipitating protein complexes tend to agglomerate with whatever's at hand. I'd be more worried about head impacts if you lose too many small proteins.

In the vein of another hot thread, I wonder if some of the improvements are due to Zn in the trub.

 

[Bobby_M]

Thanks for the great experiment and video Bobby. However anecdotal, tests like these push the conversation forward while possibly washing away some dogma. With my pursuit of low oxygen brewing, I have gone very far into the clear everything camp. The sad thing about homebrewing compared to pro brewing is that we do not brew often enough to really see trends and fix them. I have had some fermentation challenges with lagers but lag time was not the issue. I am totally sold on clear wort into the kettle. The use of Brewtan-B in the mash helps this even more. I am a little more open on less clarity into the fermenter, but I pause when it comes to stale flavors for longevity and yeast repitching.

Since going down the rabbit hole, I have noticed every beer has a more "clear" flavor. Meaning less mud. I have to attribute some of this to focusing on clarity through the process. If the difference is throwing in some yeast nutrient vs carrying over trub, I think I would still choose the yeast nutrient until more testing was done. Logically, the idea of just clear wort and yeast sounds like it would produce the more natural beer. But like you said, it is best to follow the data.

I think it's easy to slowly convince yourself that efforts are worth it, especially if you spend any money. I was pretty confident that the CBCF was going to be a better beer.

 

[Bassman2003]

I agree. We are all slow boiling frogs!! I thought the same thing but when I step back and think what is happening here, it appears to be a yeast nutrition thing. For me, the tasting was the major part of the experiment. Beer clarity is achieved in many forms but taste is either there or not. It would stand to reason that the trub is giving the yeast something that it needs/wants resulting in a better ferment & tasting beer.

The low oxygen/clarity nut folks add zinc to the fermenter and or yeast nutrients. They/I have also learned to add a lot of oxygen for the yeast. The idea being to strip away everything that is negative and if necessary, add in what is needed. Imho, ferment quality=flavor. So I would say the dirty fermenter had a better ferment.

 

[balrog]

Have we just started the Dirty Fermenter subreddit?

 

[Bobby_M]

The low oxygen/clarity nut folks add zinc to the fermenter and or yeast nutrients.

I wonder what the flavor impact of the nutrient is. I mean directly, not via its utility on yeast health.

 

[Bassman2003]

I would hope zero impact on flavor...

 

[DavidWood2115]

I wonder what the flavor impact of the nutrient is. I mean directly, not via its utility on yeast health.

If you are adding just zinc, it's hard to imagine you could taste it. The recommended dosage is only 300 ppb.
https://www.homebrewtalk.com/threads/zinc-usage-amount-for-healthy-fermentation.732186/

 

[corncob]

I work with data a little in my day job--not usually data reported by people, but it's not miles different.

It would be reasonable, depending on how you frame the question "what does this indicate?" to interpret this data as inconclusive. One gripe many have with the brulosophy guys is their misuse of statistics, which is real. But the idea of assigning a threshold for declaring demonstration of difference is valid, even if the execution is faulty or inconsistent. Without getting a crack at the raw data, and without having any other reference points, it is probably reasonable to say that this test showed little or no difference between the methods tested.

 

[Bobby_M]

I work with data a little in my day job--not usually data reported by people, but it's not miles different.

It would be reasonable, depending on how you frame the question "what does this indicate?" to interpret this data as inconclusive. One gripe many have with the brulosophy guys is their misuse of statistics, which is real. But the idea of assigning a threshold for declaring demonstration of difference is valid, even if the execution is faulty or inconsistent. Without getting a crack at the raw data, and without having any other reference points, it is probably reasonable to say that this test showed little or no difference between the methods tested.

I accept that many academic endeavors do not track with common sense at all. Can you explain how this doesn't clearly show that beer D was better than beer A? Zero people rated A an 8, 9 or 10 but Five people did for beer D. Conversely beer A received 7 ratings of 1, 2, or 3 while beer D received no ratings that low. I'm open to better understanding.

 

[corncob]

I accept that many academic endeavors do not track with common sense at all. Can you explain how this doesn't clearly show that beer D was better than beer A? Zero people rated A an 8, 9 or 10 but Five people did for beer D. Conversely beer A received 7 ratings of 1, 2, or 3 while beer D received no ratings that low. I'm open to better understanding.

View attachment 863217

To do it by the book you would have to show (using more data and analysis) that the differences in the datasets you have are meaningful, given the normal variation in this data. Basically: what if you repeated this test 9 more times--same beer, same tasters, the whole thing, and compiled all the data? How certain are you (based only on this one test) that the big data set would look identical to this one? That question is what the father of statistics was asking himself when he was looking at data from batches at the Guinness brewery.

In not saying this test is not indicative of anything real we can take away--it definitely is useful. I'm just saying that it's not "statistically significant." we would be hasty to toss out our brew bags and corona mills based on this test alone. On the other hand, when compiling this with other anecdotal data we possess, a somewhat clear picture is emerging. It would not be wise to dismiss this.

Statistics is a very divisive tool. It's actually only useful for a very narrow set of tasks, but only people on the inside know that. So numbers are abused and used as cudgels to batter the laymen and fleece the sheep. Do not ever trust a number you didn't calculate yourself. Point of fact: "statistically significant" never, ever means the same thing as "true"

 

[mashdar]

I don't think we're looking for a statistically significant result; it's possible the raw data achieves this (and I'd be intrigued to poke at the data). Even if it were significant, there'd be enough confounding that you could keep moving goal posts.

I'm happy to see the wholistic approach, as I'm sure that's much more fun for Bobby. I'd trade some statistical wizardry for more trials; the latter sounds much more interesting.

(My biggest gripe with Brulosophy is when they claim to have confirmed the null. Lack of evidence is not evidence of lack. edit: but I appreciate their work nonetheless)

 

[doug293cz]

I'm just saying that it's not "statistically significant."

This is an absolute statement. What statistical test(s) did you use to determine this? Please show your work.

Brew on

 

[corncob]

This is an absolute statement. What statistical test(s) did you use to determine this? Please show your work.

Brew on

Hey I'm not the one making a claim. You got me all wrong.

I ain't seen nothin.

I don't know nothin.

 

[doug293cz]

Hey I'm not the one making a claim. You got me all wrong.

I ain't seen nothin.

I don't know nothin.

Um, I quoted directly from one of your posts. You don't get to say that you didn't say what you are on record as saying.

Brew on

 

[corncob]

Um, I quoted directly from one of your posts. You don't get to say that you didn't say what you are on record as saying.

Brew on

I retract the statement. I yield. You win the internet today.

 

[Bassman2003]

I don't think it helps to pick apart the statistics when Bobby said up front that this was a casual experiment/test. It discourages people from putting in the effort to do things like this. Seriously. Corncob, I am not trying to tear you down, what you said is probably true but I think a little bit of a filter might be a better approach. If you do not like the data, either put the effort in and make your own video & test or just let it pass. What this test showed is that there is a difference to people in the tasting and by Bobby himself. How much weight that holds for each of us is up to the reader.

If you were in person I do not think you would have given the statistical analysis. I am on a crusade for some more respect and manners on the internet. It takes a lot of time and effort and willingness to offer something to do a multi-week test like this. I get the impression that true statistics are kind of a mythical absolute that is almost never reached. It is a homebrew test and just don't squeeze the bag. Sorry and please carry on with your insightful posts...

 

[marc1]

I accept that many academic endeavors do not track with common sense at all. Can you explain how this doesn't clearly show that beer D was better than beer A? Zero people rated A an 8, 9 or 10 but Five people did for beer D. Conversely beer A received 7 ratings of 1, 2, or 3 while beer D received no ratings that low. I'm open to better understanding.

View attachment 863217

You can put all the individual scores into an online ANOVA calculator for stats. It will compare all the groups against each other. Run it once for the clarity scores and once for the taste scores.

 

[corncob]

Bobby sells biab equipment for a living. The test he graciously put together and graciously documented seemed to indicate that biab might tend to produce less than the best beer. By suggesting that the results aren't black and white, I was not not trying to tear our guy down, but to build him up.

I'm not sure anova applies anyway. Multiple testers VS. multiple trials VS. multiple variables. The point I was trying (but obviously failing) to make is that the numbers in this case are only qualitative. It would be a mistake to do a rigorous statistical analysis.

I am a big fan of these kinds of experiments.

 

[Bobby_M]

I don't think it helps to pick apart the statistics when Bobby said up front that this was a casual experiment/test. It discourages people from putting in the effort to do things like this. Seriously. Corncob, I am not trying to tear you down, what you said is probably true but I think a little bit of a filter might be a better approach. If you do not like the data, either put the effort in and make your own video & test or just let it pass. What this test showed is that there is a difference to people in the tasting and by Bobby himself. How much weight that holds for each of us is up to the reader.

If you were in person I do not think you would have given the statistical analysis. I am on a crusade for some more respect and manners on the internet. It takes a lot of time and effort and willingness to offer something to do a multi-week test like this. I get the impression that true statistics are kind of a mythical absolute that is almost never reached. It is a homebrew test and just don't squeeze the bag. Sorry and please carry on with your insightful posts...

I definitely don't want this thread to death spiral. I can read some criticisms about the approach and mildly defend and possibly disregard some things that are over my head. I know that true data nerds have the best intentions and don't want anyone to misunderstand reality.

 

[Bobby_M]

If anyone wants to play with the numbers..

 

[tdf]

Thanks for sharing the raw data. I’ve been enjoying this thread and I’m glad you put in the effort.

 

[pers]

Out of curiosity, was there any correlation between being able to correctly identify the 5th sample and the taster level of training?

 

[Bobby_M]

Out of curiosity, was there any correlation between being able to correctly identify the 5th sample and the taster level of training?

A taster accuracy of 3 means they scored the E cup the exact same as its duplicate sample. If they were one apart, that accuracy score went to 2, and so on.
I wasn't particularly impressed here. There were five non-BJCP judges that matched perfectly.

I think judges think too hard. I did hear one of them suggesting that maybe each one was dosed with Siebel calibration compounds. I suppose after tasting the diacetyl in A, it was a primed assumption.

Perhaps one thing that is interesting is that no matter how I applied weighting, whether it be extra weight for judge rank or E-cup accuracy, or BOTH, none of that changed the final ranking of the beers. That's why I left most of that out of the video.