Why assigning "HARD" values to "ALL" grains and adjuncts is merely blowing smoke

[Silver_Is_Money]

I will however state that (right or wrong) due to my very own personal biases and data cherry picking choices my regressions differ from others that I have observed.

One such regression I've seen includes Caramel/Crystal malts on the "main sequence", and I outright reject this.

 

[hout17]

Maybe move this thread to the rant section lol.

 

[Silver_Is_Money]

As I recall, my regression derived equations whereby to assign pHDI and BC valuations via malt color are now in the public domain, and to be found on Brewers Friend.

But I somewhat recently made changes to the regressions, leading to somewhat (or slightly) different equations, thereby antiquating what is to be found from me in this regard on Brewers Friend.

My changes were in the general direction of yielding somewhat higher malt acidities, particularly to the 'main sequence' malts/grains. But Caramel/Crystal regression and thereby equations were also modified.

 

[Silver_Is_Money]

Another philosophical blurb, if I may:

"Everyone is free to evade reality, but no one is free to avoid the consequences of evading reality."

 

[Silver_Is_Money]

Let's take the case for 5 Kg of Briess 2-Row Brewers malt, and the extant data from researcher #1 and researcher #2:

For researcher #1:

(6-5.4) = mEq's_acid_base/(57.2 x 5)
0.6 = mEq's_acid_base/(286)
mEq's_acid_base = 171.6

For a mash pH target of 5.40, and given that at that pH the mEq's of acid contained within 1 mL of 88% Lactic Acid = 11.451, we get:

171.6/11.451 = 14.896 mL's of added 88% Lactic Acid to move this mash to pH 5.40


And now for researcher #2:

(5.55-5.4) = mEq's_acid_base/(46.2 x 5)
0.15 = mEq's_acid_base/(231)
mEq's_acid_base = 34.65

For a mash pH target of 5.40, and given that at that pH the mEq's of acid contained within 1 mL of 88% Lactic Acid = 11.451, we get:

34.65/11.451 = 3.03 mL's of added 88% Lactic Acid to move this mash to pH 5.40

14.9 mL's of 88% LA required for researcher #1's data vs. 3.03 mL's for researcher #2's data.

This is how bad the data gathered by D. Mark Riffe is for this malt. If you plot the computed mEq's for both vs. the nominal color of this malt, that is clearly a whopper load of chart scatter.

So if we are mashing 5 Kg. of this Malt in DI water, do we add 15 mL of Lactic Acid, or do we add 3 mL of Lactic Acid? Or should we split the difference, and add 9 mL of Lactic Acid? Or should we apply bias and cherry-pick the data?

As an aside, titrations like this are generally carried out using pulverized malt whereby to liberate all of the malts inherent mEq's of acid or base (with basic meaning with respect to the targeted mash pH). But the typical LHBS gives you malt crushed through a mill gap of about 0.039". Should one's software presume that 100% of the malts inherent mEq's of acid or base (with basic meaning with respect to the targeted mash pH) will be liberated for the case of the 0.039" mill gap? Or should the software compensate for less saturation during the mash and thereby conclude that a bit less acid or base will likely be required whereby to move the poorly crushed malt to the desired target pH during the mash? We already know that the yield of sugars from the poorly crushed malt(s) will result in a low OG. This seems to imply that the mash water is not reaching some portion of the malt. And thereby failing to liberate the acid/base as well as the sugars.

 

[Silver_Is_Money]

Here I must inject that Kai Troester (aKa, Braukaiser) carried out a series of tests to determine a malts liberation of acid (or base with respect to target pH) and concluded that there is measurably less liberation (release into the mash water) as mill gap is widened. From his data points an equation can be regressed whereby to utilize mill gap and directly correlate it to liberation. And this regression has a respectfully high degree of statistical confidence. And (to my knowledge) only one software playing in the mash pH adjustment assistance arena offers the user to voluntarily apply mill-gap compensation.

 

[bwible]

This has turned into a rather scientific thread with high math equations. To cut to the chase, you are saying that we can’t really predict ph of our mash accurately using the data provided by maltsters. OK, we get it that everything is not accurate to 7 decimal digits. What in life is?

The question is how far off does the data really leave us? When we have a target ph range 5.4-5.5 say - are you saying we can’t ever accurately hit within that range? Or am I missing the point?

 

[bwible]

And when it comes to predicting color, we can’t really do that either. Your software all has 5 different color models built in that all give a different number based on which one you pick.

[edit] I don’t mean your software Larry. I mean the software we all use like Brewers Friend or BeerTools or whatever.

 

[Silver_Is_Money]

My 'USA' (or North America) version asks specifically for 'nominal' Lovibond color input from the user, and my 'Metric' version asks for specifically nominal 'EBC' color input. Each version is asking for only one color input, so I'm confused as to where you derive five.

 

[Silver_Is_Money]

The only point I can emphasize is that it is difficult (to perhaps impossible) to derive software intended for providing "precise" mash pH adjustment assistance given the scant "HARD" data at hand.

Another consideration is that other of such softwares may not be using BC and pHDI whereby to derive malt acidity valuations at all. How they model without utilizing such critical factors remains a mystery, as they seem to not wish to publicly divulge the inner workings of their software.

 

[bwible]

My 'USA' (or North America) version asks specifically for 'nominal' Lovibond color input from the user, and my 'Metric' version asks for specifically nominal 'EBC' color input. Each version is asking for only one color input, so I'm confused as to where you derive five.

Daniels, Tinseth, Garetz, Mosher, “standard” - these are the options in BeerTools and every model predicts a different color off the exact same recipe. I’m sure its the same no matter what brewing software you use

 

[Silver_Is_Money]

I presume that you are referring to "finished beer" color prediction, which has nothing to do with this discussion.

 

[bwible]

I presume that you are referring to "finished beer" color prediction, which has nothing to do with this discussion.

Yeah the models all give different ibu numbers too off the same recipe. It seems like everybody uses Tinseth as the standard but without sending our finished beer out to a lab who knows which model has the right numbers? Its just a continuation of the “all over the place” theme

 

[Silver_Is_Money]

This has turned into a rather scientific thread with high math equations.

I would call the equation rather rudimentary as opposed to high. AJ deLange would apply high math equations.

But as AJ deLange once stated (quoting here from memory): "Well, after all this is the Brew Science forum."

 

[Silver_Is_Money]

Yeah the models all give different ibu numbers too off the same recipe. It seems like everybody uses Tinseth as the standard but without sending our finished beer out to a lab who knows which model has the right numbers? Its just a continuation of the “all over the place” theme

This is also not part of the discussion for this thread, but the simple answer is that none of the extant IBU math models are "right". They are all merely ballpark at best. Sometimes they nail it (in a fuzzy logic or luck sort of way), and sometimes they don't. I'd say they will 'generally' (though not always) get you within roughly +/- 35% of what a lab analysis might derive for IBU. But ones taste buds may not be capable of deriving IBU's to within +/- 35% either.

 

[BrewnWKopperKat]

“All models are wrong, but [many] are useful”.
-- George E. P. Box (?)

edits: typohz

 

[madscientist451]

The real intent of such software is to sell the sizzle as opposed to the steak (or meat) of the programming, and to dazzle the unaware end user.

You are no longer unaware.

The analysis at the end of the post seems a bit harsh, I kind of wish I was still unaware of all of this.
I'll go back to daydreaming about 5pm and my first pint of the weekend....

 

[Silver_Is_Money]

“All models are wrong, but [may] are useful”.
-- George E. P. Box (?)

All models are approximations. Essentially, all models are wrong, but some are useful. However, the approximate nature of the model must always be borne in mind.

George E. P. Box

 

[bwible]

I would call the equation rather rudimentary as opposed to high.

But as AJ deLange once stated (quoting here from memory): "Well, after all this is the Brew Science forum."

I became a computer programmer because I hated math and computers are good at it. The first thing I had to do was learn number systems. Funny how stuff goes sometimes

 

[bwible]

I would call the equation rather rudimentary as opposed to high.

But as AJ deLange once stated (quoting here from memory): "Well, after all this is the Brew Science forum."

Right - Brewing encompasses History, Geography, Math, Chemistry, Biology, Social Studies - and a few other classes I’m probably forgetting.