Here's an updated plot where the horizontal spacing of the Agtron readings is scaled proportionally. Still only 3 data points, but now appears more linear. To the extent one can rely on the Onyx Commercial Agtron values, an Espresso Vision colorimeter allows you to get a reading that is sufficiently accurate for guiding espresso prep and extractions in both an absolute and relative sense.

^^^ I mean, yes, BUT I'd feel a lot better if you had n > 3. Cool nonetheless.

Batch numbers:

Eclipse = 52382 -> listed as Peru Celestino (not a current offering) instead of Eclipse; roast = "Super Dark" (for Onyx, that is), but listed tasting notes would be for a medium light to light. Brewed as 16.5g with an LR profile at 195degF, BR = 1.5 tasted of slightly tart-sour, slightly sweet intense red wine with some dark choc (interestingly like Monarch is supposed to be).

Monarch = 52438 -> is listed as Monarch; roast = "Expressive Dark" (dark choc, red wine, molasses, dried berries). Bordering on medium light roast, and brewed with a longer BR= 2.2, it does have some red wine with dark choc. This is another instance of the current trend by roasters to obtain dark roast flavors from non-dark roast- at least in this case Onyx has qualified dark with "expressive".

Columbia Apnte Village = 52386 -> this one comes up as a Decaf version.

Hey Bizmark, I'm just curious why you made your EV to Agtron table just slightly different than the scale in the image from the EV website? For example, EV 25 lines up with Agtron 93 but you list it as 94 in your table. Was that intentional?

Onyx says they go as dark as they can without oil and a RV15 is right there in my experience. A 13 is where I start getting an oily sheen over the beans and a 14-15 gets spotting after a few days. I'm actually a bit surprised they go that dark, so that is interesting. I've had their Eclipse before and it was pleasant for that roast level.

From what I can tell, Onyx is listing whole bean Agtron scores, not ground. I've measured Wendelboe's coffees at around 70 whole bean, 110 ground. Ground is always higher than whole and the delta decreases as the roast darkens. 70/110 is my typical target for Nordic roasts. Morten Munchow's work usually only discusses ground scores, since that is what you taste. I like to measure both whole and ground and track the delta, but so far the two scores seem to be highly correlated, at least in standard roasts. I'm sure you could do a roast where you scorch the outside and leave the inside unroasted.

A 40 point spread is pretty spooky. Hard to believe it is not underdeveloped.

It all depends on what you consider as "underdeveloped". Onyx coffees, for me, are generally too darkly roasted for my tastes. Tim Wendelboe's roasts are, in my opinion, reliably well developed. There are a handful of roasters that roast lighter than TW and are reliable enough for me to order regularly. I guess it is all perspective and then skill and experience in roasting.

Well according the the Agtron gourmet scale, 100 is underdeveloped. 110 is extra-underdeveloped. I'm not sure how Todd measured that TW coffee he mentioned. Grind can certainly affect ground coffee readings, but a 40 point spread is crazy town.

I just did a few sample roasts this afternoon. While the Cormorant was hot I did another 230g in a quick (6:20 total), hot roast, dropping at 10* above 1C in 45s and got a 72 whole bean score and an 84 for a ground score. I'm at a loss to see how I can go 2 points darker on the WB and 26 point lighter on the ground.

Maybe we're getting into the same territory when trying to compare roast profiles. Different equipment, different data. Makes communication about coffee difficult. Hence my ongoing frustration with the coffee industry.