Reclaiming Trub Yeast.

[Potatoes]

Especially when effect size is small. If it was beer compared orange juice, significant results would be easier to get. 

[Otto Von Blotto]

I think part of it is only testing one thing at a time, of itself not bad experimental practice, but one brewing process considered to not be best practice may not result in much difference at the glass. If you brew a batch and use several bad practices then it might have more effect. 

[BlackSands]

On 10/8/2018 at 9:12 AM, Otto Von Blotto said:

....to me the chances of an "insignificant" result are greater with a smaller sample size.

I'm no statistician and really don't know what sample size would be optimal in this situation to give a decent level of confidence in the results but I would have thought the level of confidence with a small sample size would apply to both 'significant' and 'insignificant' findings alike.  In other words if a trial yields a positive result (and of course many do) they too would share the same level of uncertainty.  I think the arguments against the sample size used works both ways.   

It's also worth noting that a number of their trials have been repeated many times.  Replicating results several times would tend to negate any argument relating to the small sample sizes they use.   

 

[BlackSands]

12 hours ago, Otto Von Blotto said:

I think part of it is only testing one thing at a time, of itself not bad experimental practice, but one brewing process considered to not be best practice may not result in much difference at the glass. If you brew a batch and use several bad practices then it might have more effect. 

"Short and Shoddy" (16 brews so far).    This series of trials break several "best-practice rules" at a time e.g. short mash times, short boils, trub into the fermenter etc.  

 

[Otto Von Blotto]

If they haven't already I think they should test these things with beer that has been aged for a period of time, or different periods of time. Some of the effects of bad practices may not show up early on. All the experiments I've read have had them tasting the beers about a week after kegging. 

As well as that, while it is only anecdotal I have had different experiences with kettle trub being allowed into the fermenter, as a result of not using kettle finings (unintentionally, and the only difference in process). The two batches where I did this resulted in cloudy wort into the cube, then into the fermenter, cloudy beer into the keg despite treatment with isinglass and polyclar, and it just didn't taste as good as batches brewed with kettle finings. Excess trub can also result in oxidation problems, but these probably wouldn't appear straight away. I didn't get that issue in those beers but they weren't aged for that long as I just wanted them gone.

[BlackSands]

16 minutes ago, Otto Von Blotto said:

If they haven't already I think they should test these things with beer that has been aged for a period of time, or different periods of time. Some of the effects of bad practices may not show up early on. All the experiments I've read have had them tasting the beers about a week after kegging. 

Yeah, I'm aware of the argument  that "time may tell".  That in time the effects (if any) of 'shoddy' practices might begin to show.  I also suspect it's a non-issue for many in practice because most brewers I know are not into long-term storage of their beer!  I rarely have a batch exceed 6 weeks!   

 

Quote

..kettle trub being allowed into the fermenter, as a result of not using kettle finings (unintentionally, and the only difference in process). The two batches where I did this resulted in cloudy wort into the cube, then into the fermenter, cloudy beer into the keg despite treatment with isinglass and polyclar, and it just didn't taste as good as batches brewed with kettle finings. 

I was interested to note their trub exbeeriments where incidentally they found faster fermentation, greater attenuation and clearer beer when the trub was thrown into the fermenter. 

I do it all the time now. I don't bother faffing about trying to filter/exclude the trub -  I just chuck the whole lot in. In my partial-mash case though it's only trub from around 2kg grain and whatever hop additions.  One thing I've noted (anecdotally) is that by throwing hoppy trub into the fermenter I'm now getting a LOT more out of my late hops than I was previously.     

 

 

[Otto Von Blotto]

Depends on what long term storage is. Mine don't often last more than 6 weeks at the moment either, but I have left beers for months depending on style. Who knows if they had been as nice if I'd used a number of shoddy practices instead of what I did do. 

I think the biggest issue isn't as much the practices themselves but that some brewers will read it and think that it doesn't matter what you do, the beer will always turn out well, and it's BS. You only have to look around brewing forums to see the endless questions about why a beer turned out badly, and it's almost always down to using a shitty practice somewhere along the line. They'll start getting lazy, and when the time comes that they do end up with a shitty batch they won't have a clue why. 

[Potatoes]

On 10/9/2018 at 10:23 AM, BlackSands said:

I'm no statistician and really don't know what sample size would be optimal in this situation to give a decent level of confidence in the results but I would have thought the level of confidence with a small sample size would apply to both 'significant' and 'insignificant' findings alike.  In other words if a trial yields a positive result (and of course many do) they too would share the same level of uncertainty.  I think the arguments against the sample size used works both ways.   

 

Sample size depends on the size of the effect. So if it’s a small effect, (small difference in taste), then you need a larger sample size to find a significant result. 

After reading the website, I see they use a binomial test with a one tail distribution. This calculator allows you to estimate the number of correct responses you need to get a significant result. https://www.socscistatistics.com/tests/binomial/Default2.aspx

Assuming a 0.3333 chance of guessing correctly, I think ( calculate) the following:

A sample size of 21 means you need 12 correct responses to be statsically significant. 

A sample size of 42 means you need 20 correct responses to be statsically significant.

A sample size of 63 means you need 28 correct responses to be statsically significant.

i think that is right... happy to be corrected.

 

overall, I really like the brulopshy website and approach. I think most people reading the website understand how to interpret the results. 

 

[Otto Von Blotto]

Take them with a grain of salt is the way to interpret the results  

I have never said that using bad brewing practice will definitely result in a bad beer or something going wrong 100% of the time. All I've ever said is that doing so risks that kind of outcome. 

It's just illogical to me that anyone would spend the time and effort creating wort either from scratch with AG or doing partials, and then risk ruining all that work by essentially being lazy with one or more processes. If you use good practices every time then you'll get consistently good or great results. If you cut corners you risk inconsistent quality. Not worth it to me just to save a few minutes here and there on a 6 hour brew day.

[BlackSands]

On 10/9/2018 at 1:26 PM, Otto Von Blotto said:

 

I think the biggest issue isn't as much the practices themselves but that some brewers will read it and think that it doesn't matter what you do, the beer will always turn out well, and it's BS. You only have to look around brewing forums to see the endless questions about why a beer turned out badly, and it's almost always down to using a shitty practice somewhere along the line. They'll start getting lazy, and when the time comes that they do end up with a shitty batch they won't have a clue why. 

I think the exact opposite is true.  What these process-related trials show is how robust the process really is and that it would seem there is in fact plenty off leeway.  They show you can be 'lazy', or 'time-efficient' as I would prefer to call it.

Within reason you can be pretty loose with boil times, pitch rates, fermentation temperatures etc etc all at the same time and still make great beer.  What you CAN'T be relaxed about however is sanitation and I think ultimately oversights in that area is probably the main source of problematic brews. 

[BlackSands]

10 hours ago, Potatoes said:

Sample size depends on the size of the effect. So if it’s a small effect, (small difference in taste), then you need a larger sample size to find a significant result. 

After reading the website, I see they use a binomial test with a one tail distribution. This calculator allows you to estimate the number of correct responses you need to get a significant result. https://www.socscistatistics.com/tests/binomial/Default2.aspx

Assuming a 0.3333 chance of guessing correctly, I think ( calculate) the following:

A sample size of 21 means you need 12 correct responses to be statsically significant. 

A sample size of 42 means you need 20 correct responses to be statsically significant.

A sample size of 63 means you need 28 correct responses to be statsically significant.

i think that is right... happy to be corrected.

 

overall, I really like the brulopshy website and approach. I think most people reading the website understand how to interpret the results. 

 

Your numbers seem consistent with Brulosophy.  However I think the issue raised is not to do with 'statistical significance' as such but rather the reliability or confidence in the result, whether it be significant or otherwise, when such a small sample size is used. 

If it's close that could mean just one taster makes the difference between achieving significance or not.   With a larger sample size it could be say 5 tasters and that then would presumably mean a more reliable and trustworthy result?

[BlackSands]

7 hours ago, Otto Von Blotto said:

.If you use good practices every time then you'll get consistently good or great results. If you cut corners you risk inconsistent quality. Not worth it to me just to save a few minutes here and there on a 6 hour brew day.

'Good practices' vs 'pointless practices'.  

'Good practice' according to who based on what evidence?  

You misrepresent the time potentially saved by eliminating time-wasting unnessary 'good'practices' which can in fact amount to hours.  

[Otto Von Blotto]

The trials don't show much at all in relation to the overall homebrewing community, all they show is what happened in old mate's garage. Just like what happens in my brewery isn't necessarily indicative of the wider brewing community. I've read countless posts about fermentation taking way longer than it should, stalling, simply not starting, or the beer tasting off as a result of it. That's from underpitching and/or fermenting too cold, not being slack with cleaning and sanitation. But according to Brulosophy, underpitching doesn't matter and doesn't make any difference to anything. I've also read numerous posts about beers that were fermented too warm tasting off. But again if you believe Brulosophy that can't be the cause because their expert panel couldn't tell the difference. It doesn't prove anything, and it certainly doesn't disprove the general consensus on best brewing practices. At the end of the day what people do is up to them but I figure if you are going to use best practices with cleaning and sanitation then you might as well do it for all brewing processes.

[Otto Von Blotto]

Just now, BlackSands said:

'Good practices' vs 'pointless practices'.  

'Good practice' according to who based on what evidence?  

You misrepresent the time potentially saved by eliminating time-wasting unnessary 'good'practices' which can in fact amount to hours.  

Who says they're pointless practices? 20 guys in a garage in California vs. decades of brewing research? The only way you would save hours on a brew day is by going back to full extract. When I have the time I'm quite happy to go and find all the scientific literature that explains why the best practices are considered such. It might make you re-think whether some things are pointless or not.

[BlackSands]

2 hours ago, Otto Von Blotto said:

...when I have the time I'm quite happy to go and find all the scientific literature that explains why the best practices are considered such..

That's what I've been hoping to see for quite some time.  I'm not aware of anyone offering up any direct references to the science that supports many of what some consider 'best practices'. It seems time is always short when it comes to actually backing up such assertions. 

I imagine a fair bit of research is conducted in-house by commercial breweries and may not be made available to the public nor may even be all that applicable on a homebrew scale anyway.  But I assume there's some out there for all to see?  

Anyway, I remain open and would genuinely like to see some of the science data in support of current common practices many of which I suspect are actually just tradition-based, i.e. "we do it that way because that's the way we've always done it and it works so why change it?".     I don't expect to understand the technicalities of the science but the punch-line can usually be found in the conclusions/summary section at the end of scientific papers.  

 

[Otto Von Blotto]

I have linked to them before but maybe on a different forum. I'd agree not all things would necessarily translate to our level but chemical reactions would. It's the same stuff we're all using whether the batch is 20L or 20,000L. A lot of practices would have been tradition based but have since been found scientifically to be the best practice. Some have also been superseded by newer practices that have been shown to be better than the old ones.

[BlackSands]

5 minutes ago, Otto Von Blotto said:

Some have also been superseded by newer practices that have been shown to be better than the old ones.

Newer practices as shown by the great work done by the likes of Brulosophy and co!    

[Otto Von Blotto]

Not quite what I meant

[Greenyinthewestofsydney]

3 minutes ago, Otto Von Blotto said:

I have linked to them before but maybe on a different forum. I'd agree not all things would necessarily translate to our level but chemical reactions would. It's the same stuff we're all using whether the batch is 20L or 20,000L. A lot of practices would have been tradition based but have since been found scientifically to be the best practice. Some have also been superseded by newer practices that have been shown to be better than the old ones.

I would say with confidence that commercial breweries have put ALL the traditions to the test. They would have a scaled pilot plant based on there facility and would have tried all sorts of different brewing techniques to get the best quality and efficiency. Whether that translates to our 30L fermenters is another thing altogether.

[BlackSands]

17 minutes ago, Greeny1525229549 said:

I would say with confidence that commercial breweries have put ALL the traditions to the test. They would have a scaled pilot plant based on there facility and would have tried all sorts of different brewing techniques to get the best quality and efficiency.

Saying something with 'confidence' is of course not the same thing as providing supporting evidence   

However, I do agree that from a commercial perspective there would certainly be a keen interest in exploring anything that could potentially amount to a cost-saving without of course sacrificing quality or more specifically,  detectable quality.

[Otto Von Blotto]

I'd say some does translate to this scale and some doesn't. 

I know they don't often get statistical significance in those experiments but they're often only short by one or two tasters. That's still around half the tasters picking up a difference in most of them. It might not be significant statistically but I don't think it shows there's no difference either. If there was you'd expect no tasters to pick it up, or a few at the most, but not around half of them. 

 

[Otto Von Blotto]

http://www.ibdlearningzone.org.uk/article/show/pdf/493/
That one explains a few functions that wort boiling performs, which would ring true on any brewing scale, and suggests that complete removal of proteins etc. could lead to poorer quality beer, which makes sense, however it also doesn't suggest that throwing the entire hot break into the fermenter is beneficial either.

I also found this article which lines up with my experience with poor hot break separation, in the hot break separation section: https://www.researchgate.net/publication/215984266_Wort_boiling_today

[Potatoes]

14 hours ago, Otto Von Blotto said:

But according to Brulosophy, underpitching doesn't matter and doesn't make any difference to anything. I've also read numerous posts about beers that were fermented too warm tasting off. But again if you believe Brulosophy that can't be the cause because their expert panel couldn't tell the difference. 

This is what I was getting at with small sample sizes. There is a difference but it’s not detected at statistically significant levels because your sample size is small. It happens more so with subtle differences (smaller effects). 

 

Also, expert judges were no better or worse at identifying the odd beer than non-expert judges. http://brulosophy.com/2016/01/21/investigating-the-bad-palates-argument-a-graphical-look-at-xbmt-performance-based-on-experience-level/

 

ps 15000 posts! Well done.