Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
By using a 3 coefficient quadratic regression of DI_pH data which Briess provided to me upon request I have achieved an R squared confidence factor of nearly 94% for the correlation of Lovibond color to DI_pH across their entire range of caramel malts spanning from 10L to 120L. This proves to me that a commonly seen alternative approach of attempting to group caramel malts into "color groups" and thereby assign "group" related acidity values (as opposed to calculating their acidity directly from L color) is a flawed approach that will not achieve such a level as 94% precision, and going with a calculation of DI_pH via L color is indeed the proper means to apply here.
But the same source also contains DI_pH vs. Lovibond color data for a number of crystal malts (presumably from the testing of competitors "crystal process" malts), and my attempts at linear and various methods of non-linear regression of this crystal malt data have at best achieved R squared correlation of about 56% for crystal. The odd data is such that acidity is high (DI_pH is low) for lower L crystal malts, then acidity goes way down (DI_pH's go noticeably up) across the mid-range of L colors, and then acidity dramatically rises again for the highest L range of crystal malts.
My preliminary conclusion from this is that caramel and crystal malts should not be lumped together into a single category for the assignment of acidity via color (as has been commonly done for many years), and the two processes (caramel and crystal, albeit related) likely need to be handled separately and divided into two distinct malt categories. This is somewhat similar to my acidity related conclusion reached for malted and unmalted deep roasted grains, but whereby the solution to deep roasted acidity was rather easy via categorically separating them, the solution for crystal malts may require grouping or some other method (higher order quadratic regression perhaps). Stay tuned.
But the same source also contains DI_pH vs. Lovibond color data for a number of crystal malts (presumably from the testing of competitors "crystal process" malts), and my attempts at linear and various methods of non-linear regression of this crystal malt data have at best achieved R squared correlation of about 56% for crystal. The odd data is such that acidity is high (DI_pH is low) for lower L crystal malts, then acidity goes way down (DI_pH's go noticeably up) across the mid-range of L colors, and then acidity dramatically rises again for the highest L range of crystal malts.
My preliminary conclusion from this is that caramel and crystal malts should not be lumped together into a single category for the assignment of acidity via color (as has been commonly done for many years), and the two processes (caramel and crystal, albeit related) likely need to be handled separately and divided into two distinct malt categories. This is somewhat similar to my acidity related conclusion reached for malted and unmalted deep roasted grains, but whereby the solution to deep roasted acidity was rather easy via categorically separating them, the solution for crystal malts may require grouping or some other method (higher order quadratic regression perhaps). Stay tuned.
Last edited: