HBT 2014 Big Giveaway - 94 Chances to Win!

Huge Supporting Membership Discounts!
 Home Brew Forums > Olive Oil - Testing

04-05-2013, 12:01 AM   #61
Denny
Feedback Score: 0 reviews
Recipes

Join Date: Jun 2005
Location: Eugene OR
Posts: 4,033
Liked 360 Times on 273 Posts
Likes Given: 353

Ah, forgot you were doing it in the starter. Man, that seems ike it might be harder to quantify. I wish you luck. If I can convert that spreadsheet for ya, I'll let you know.

__________________

Life begins at 60....1.060, that is!

www.dennybrew.com

http://www.experimentalbrew.com - the website for the book "Experimental Homebrewing"...coming Nov. 2014

04-05-2013, 01:29 AM   #62
Wynne-R
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2009
Location: Texas
Posts: 845
Liked 72 Times on 57 Posts
Likes Given: 67

I’m not down with the math, but I get it on a superficial level. I know AJ comes up with some non-intuitive, but probably correct stuff. I might question the assumptions for the model, but I don’t question the analysis. I value what he does for me and the great unwashed.

I plan to run my triangles on Saturday, so I should have some raw data on Sunday. I imagine somebody will whip out some probabilities. If not I will figure it out or find someone to do it.

I just bought 120 cups and the beer is chillin’. Stay tuned.

__________________

04-05-2013, 12:35 PM   #63
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2007
Location: Southern Maine
Posts: 3,073
Liked 240 Times on 166 Posts
Likes Given: 150

Quote:
 Originally Posted by Wynne-R broadbill, I thought that was implied. Is this more clear? I want this to be as simple as possible, for me and everybody. ---------------------------------------------------------------------------- Different how? How is the one you picked different than the other two? Indicate more+ or less- If same, leave blank ----------------------------------------------------------------------------
I will say I'm confused about the 6 samples (A-F) instead of just three...how many different ones will be in there? Are you giving all 6 to everybody or are there two sets of tasting samples?

Also, my understanding of a triangle test is that all they could all be the same, or they could all be different. Your question set only works for one beer being different than the other two. Furthermore, you have indicated as much to the taster by the way the question is stated. This may or may not influence how they judge the beer.

Lastly, if you are going to use AJ's speadsheet you need to run the test identically to how he did his; otherwise his calculations may not work on your setup.

Sorry to throw all of these variables into the mix; I'm beginning to understand how complicated taste tests can be!
__________________

04-05-2013, 01:39 PM   #64
Wynne-R
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2009
Location: Texas
Posts: 845
Liked 72 Times on 57 Posts
Likes Given: 67

Broadbill my test has two sets of samples, abc and def and they are indeed odd man out. ABC is two x’s and a y, and DEF is two y’s and an x. I did it this way to keep from running out of x or y too soon.

It would be fun to have a test with identical samples, and my homebrew club is probably going to do that this year.

I’m not going to worry about AJ’s spreadsheet. It would be good to know how many votes it takes to kill the null hypothesis. I’ve never liked that critter.

__________________

11-20-2013, 10:19 PM   #65
Tall_Yotie
Feedback Score: 0 reviews
Recipes

Join Date: Jul 2010
Location: Santa Cruz
Posts: 951
Liked 9 Times on 9 Posts
Likes Given: 8

Currently working on my own test. I have a question in regards to your test method, apologies for not just posting results.

Have you (or anyone else in the experiment) did a beer that did not use any form of aeration? To see if standard transferring actually gives enough O2 to let the yeast do their thing. As adding oxygen is far harder in high gravity beers, that is where I play to do my real proving grounds (as basic transferring will no where near get the O2 levels required).

__________________

11-20-2013, 11:27 PM   #66
agenthucky
Feedback Score: 0 reviews
Recipes

Join Date: Jan 2010
Location: CT
Posts: 246
Liked 6 Times on 6 Posts
Likes Given: 3

Quote:
 Originally Posted by Hanso http://sciencebrewer.com/2013/11/12/what-to-post-after-a-long-hiatus-an-experiment-yeast-aeration-of-course/ Looking forward to the results here.
" Moreover, olive oil will not dissolve in wort and must first be dissolved in 100% ethanol. For any homebrewers reading this – do not add a drop of olive oil to your beer. I had to weigh 1 gram of oil, dissolve it, and make a serial dilution until I had close to 0.1 ug/ml. Of this solution, I added 100 uls directly to the wort."

Has anyone else come across this? This is the first I've seen of this and no one else is talking about this. Where does this come from? The solubility of oil in water? This seems like an important part of testing this theory and I haven't seen it anywhere.
__________________

11-21-2013, 01:21 AM   #67
ross_daly
Feedback Score: 0 reviews
Recipes

Join Date: Jun 2013
Posts: 29
Liked 1 Times on 1 Posts
Likes Given: 10

Quote:
 Originally Posted by agenthucky Has anyone else come across this? This is the first I've seen of this and no one else is talking about this.
I've read the whole paper and didn't notice any mention of adding alcohol to the oil. I wonder if he was relying on first hand knowledge of oil and water mixing, or if he knows specifically how the oil would be used by the yeast.

I also would like to know more.
__________________

11-21-2013, 09:42 PM   #68
ajdelange
Senior Member
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2010
Location: McLean/Ogden, Virginia/Quebec
Posts: 5,068
Liked 430 Times on 352 Posts
Likes Given: 13

I kind of lost track of this thread but I think I described how a triangle test works earlier. It should be clear that this is not an AJ inventions. It is a standard testing technique and the procedure is spelled out in detail in the ASBC MOA's.

Quote:
 Originally Posted by broadbill I will say I'm confused about the 6 samples (A-F) instead of just three...how many different ones will be in there? Are you giving all 6 to everybody or are there two sets of tasting samples?
Apparently there are 2 sets of sample (6 beers total) presented. This is not a triangle test.

Quote:
 Originally Posted by broadbill Also, my understanding of a triangle test is that all they could all be the same, or they could all be different. Your question set only works for one beer being different than the other two. Furthermore, you have indicated as much to the taster by the way the question is stated. This may or may not influence how they judge the beer.
In a triangle test 2 beers, A and B are compared. The beers are presented in groups of three of which there are 6 possible combinations in which two samples are the same and the third different:
AAB, ABA, ABB, BBA, BAB, and BAA.

For each panelist a die is rolled and one of those six groupings is assigned to him according to which face comes up. Thus not only which beer is the odd one is randomized but also the order in which they are presented. Looking at #64: the expected value of the amount of A and B used in filling the cups is equal. If the volumes are not nearly equal the panel size is too small or you have been very unlucky. For example, if you have n panelists the total number of ways in which groups can be assigned is 6^n. The number of ways in which groupings in which A is the odd beer can be assigned is 3^n. Thus the probability that all n panelists get groups with A odd is (1/2)^n = 0.1% for 10 panelists, .02% for 12 etc. Possible but quite unlikely.

The panelist is asked to identify the odd beer and then answer some second question (e.g. ("Which is better?"). Obviously, if he can't tell them apart his opinion as to which is better is not worth much. That's what makes triangle testing so powerful. The odds of correctly guessing which is the different beer and of choosing it as better by coin flipping by an ensemble of panelists is very low. The spreadsheet calculates those probabilities and it duplicates the numbers found in the MOA so there is nothing magic about the spreadsheet.

Quote:
 Originally Posted by broadbill Lastly, if you are going to use AJ's speadsheet you need to run the test identically to how he did his; otherwise his calculations may not work on your setup.
You don't have to do things identically but you have follow all the details given in the MOA. If you present 4 beers or 6 beers or don't randomize the order of presentation or allow color differences to be perceived when A is a different color than B then it isn't a triangle test and the numbers in the ASBC Table and the numbers computed by my spreadsheet don't apply. Other schemes may have statistical power comparable to the triangle test but the triangle test numbers won't fit.

Quote:
 Originally Posted by broadbill Sorry to throw all of these variables into the mix; I'm beginning to understand how complicated taste tests can be!
Yes, amen to that. You really have to think your experiment through. If you are interested in telling whether A tastes better or not than B while A is darker than B then you have to obscure the color as the panelist will be able to tell A from B by the color if he is able to see it. If, OTOH, you want to know if OO darkens beer noticeably then you would want the panelist to be able to see it.

Now supposed A and B are the same color or color is obscured but one is more highly carbonated than the other. Again, the panelist will easily be able to tell which is the odd beer with no consideration as to what you are after which is taste.
__________________

01-14-2014, 04:37 PM   #69
Cauchy
Feedback Score: 0 reviews
Recipes

Join Date: Dec 2013
Location: Detroit, Michigan
Posts: 39
Liked 4 Times on 3 Posts

Were there any results form the triangle test?

__________________

01-22-2014, 07:24 AM   #70
Schmutzli
Feedback Score: 0 reviews
Recipes

Join Date: Jan 2012
Location: Chicago, IL
Posts: 20
Liked 3 Times on 3 Posts
Likes Given: 2

 Musings/calculations

Quote:
 Originally Posted by agenthucky " Moreover, olive oil will not dissolve in wort and must first be dissolved in 100% ethanol. For any homebrewers reading this – do not add a drop of olive oil to your beer. I had to weigh 1 gram of oil, dissolve it, and make a serial dilution until I had close to 0.1 ug/ml. Of this solution, I added 100 uls directly to the wort." Has anyone else come across this? This is the first I've seen of this and no one else is talking about this. Where does this come from? The solubility of oil in water? This seems like an important part of testing this theory and I haven't seen it anywhere.
You can think of it this way. Imagine the olive oil is corn starch and the wort is hot soup. Maybe you've encountered this during cooking: if you plop a tablespoon of corn starch in the soup, you are going to get a ball of cornstarch floating in soup. If you break it up it will still be dry in the center. That's because the corn starch repels the water enough that it can't wet the inside.

Now, if you swirl the corn starch in cold water until it is suspended before adding it to the hot soup, the starch will dissolve into the soup and work as intended.

The olive oil works the same way. Added to water it remains a 'blob' like the corn starch, with most of the oil unavailable to the solution. Even if it breaks up into tiny bubbles, these are still huge compared to the yeast and the liquid between these bubbles contains zero molecules of oil. Since the olive oil is essentially being used as a vitamin to replace a product of yeast's aerobic metabolism, this is not ideal.

However, if you dissolve the oil in ethanol, it is like suspending the corn starch, only even more so. When dissolved, all molecules are separate and enter the wort separately. They spread throughout the wort and are much more available to the yeast. Over time they might reassemble, but it would take a very, very long time and hopefully the yeasties would chomp them up before then.

I looked up some studies using my academic access. Under fairly aerobic (i.e. healthy) conditions, it looks like Sac make close to 3.5mg/kg palmitoleic acid, dry cell weight. Palmitoleic acid would be the limiting constituent in just about any plant oil you could add, so we'll take it as a basis. I haven't dehydrated and weighed my yeast after a fermentation, but if someone were to do so it would be easy to find out the right amount of oil to add:

(weight of cake)* (3.5mg/kg) / (0.3-3.5% palmitoleic acid in olive oil)

or, since the numbers work out so nicely,
1-10mg/100g of anticipated dry yeast cake per batch

Taking a conservative guess at 100g of dry yeast cake for a 5 gallon batch, I get 10mg (or 0.011mL) of olive oil per 5 gallon batch (in line with other estimates), or appx 1.1mL of a 1% v/v olive oil solution in ethanol.

I don't know if its been mentioned earlier in the thread, but extra virgin olive oil is not the best choice. You want the thinnest, yellowest, most tasteless olive oil you can find to reduce phenolic content and ester generation. Macadamia oil might be a better choice because of its lower phenolic content (~48mg/kg vs 220mg/kg for the lightest olive oil)* and higher concentration of palmitoleic acid, which yeast need the most. What flavor is there is nutty rather than fruity or bitter, which should blend into the malt profile much nicer, although it would take a prince's nose to detect any flavor in 5 gallons: you'd be adding only 0.22mL of a 1% solution of macadamia oil in ethanol to match the above calculations.

Forgive me if any of this was explored earlier in the thread as I only heard of the idea today, and haven't yet read the whole thing.

Sources:
*Influence of Oxygen Addition during Growth Phase on the Biosynthesis of Lipids in Saccharomyces cerevisiae (M330-9) in Enological Fermentations, Valero et al.

*Antioxidant properties of phenolic compounds in macadamia nuts, Quinn & Tang

*http://www.agbiolab.com/files/agbiolab_Polyphenols.pdf

__________________

ajdelange Likes This