take a 1 pound bag of M&M's (male or female).. pour these into 18g portions. separate each one and count the colors. will it be the same each time? probably not. IMO a bag of espresso is a similar thing. if you could color each type of bean in a blend similar to the M&M's, the same thing would happen.

so if you try to use a blend to test or compare grinders or machines for taste i think that you will get a flawed result. if you just want to see what grinder can grind the most in the shortest time or the most consistent then you can use any bean you want and you will get a good result. if you want to see what machine can be the most stable temp, etc.. any blend or SO can work.

but it seems like if you used just a single origin (a bag of green M&M's) you would just inherently get more consistency because you will be starting out with a more consistent set of beans. i think that you would still get a variation in the taste. think, different darknesses of green instead of all the same exact green like a M&M. you would be better off with a S.O. than a true blend unless you counted the beans in each 18g dose.

is this why SO's are gaining popularity? is this why they are winning?

jon

Well, the M&M's are all the same size, while espressobeans differ in size and shape depening on the type of bean. If this means that you get even more inconsistency (like if you scoop the beans out of a jar which have been shaken) or maybe more consistency, I don't know

We hashed this out a while ago on alt.coffee.

The statistics is a straight forward application of the binomial distribution. There are roughly 8 beans per gram. The RMS error in the percentage of each blend component is sqrt(p*(1-p)/N), where N is the number of beans in the shot, and p is the proportion of the component in the blend. Twice the RMS error in each direction gives you the 90% percent range. A 10% blend constituent is going to be 5% to 15% in 90% of the shots, which could be rather unacceptable. A 20% blend constituent is going to be about 13% to 27%, which isn't too bad.

Ever since this discussion, I either use blend constituents that are each fairly good and balanced in a straight shot; or I spoon them in from separate jars, to get them exactly right. In both cases, I tend to use relatively simple proportions. The statistics of the binomial distribution makes a claim like "at 12% this blend component is perfect, at 15% it's awful" either false if the blend is actually being used, or a warning not to use it if that is true.

Stuggi wrote:Well, the M&M's are all the same size, while espressobeans differ in size and shape depening on the type of bean. If this means that you get even more inconsistency (like if you scoop the beans out of a jar which have been shaken) or maybe more consistency, I don't know

you will get a variation of flavors regardless. the different size beans will roast differently. giving different mini blends based on the overall blend. one to get a consistent blend in every shot would be to count each variety of bean and make a specific mix each shot.

it doesnt matter if you shake the jar or scoop it will still be a consistently inconsistent mix of the blend shot to shot.

in experiments it seems like eliminating variables will produce the most consistent result. using a single origin should IMO make for a more consistent less varied flavor profile in the cup. its one less variable.

jon

It depends what you're testing.

A more telling question,

"Now how did the machines know what Tasty Wheat tasted like, huh? Maybe they got it wrong. Maybe what I think Tasty Wheat tasted like actually tasted like, uh ... oatmeal or tuna fish. That makes you wonder about a lot of things."... The Matrix

How do I know what I taste as Harrar tastes like what you taste as Harrar, and if we both taste blueberry.... describe blueberry. I think all you can do rationally say is, "That was good!" and "That was ok." "That was baaaad!"

John P wrote:It depends what you're testing.

A more telling question,

"Now how did the machines know what Tasty Wheat tasted like, huh? Maybe they got it wrong. Maybe what I think Tasty Wheat tasted like actually tasted like, uh ... oatmeal or tuna fish. That makes you wonder about a lot of things."... The Matrix

How do I know what I taste as Harrar tastes like what you taste as Harrar, and if we both taste blueberry.... describe blueberry. I think all you can do rationally say is, "That was good!" and "That was ok." "That was baaaad!"

i agree.

i am mainly talking about removing a variable. not a specific flavor. but i agree with you that we perceive flavors differently. everyone has a favorite flavor of ice cream.

can i come over and visit your synesso? i am also in SLC.

maybe we can get everyone in town together for a jam.



jon

RegulatorJohnson wrote:in experiments it seems like eliminating variables will produce the most consistent result. using a single origin should IMO make for a more consistent less varied flavor profile in the cup. its one less variable.

I agree. That's why I used one roast batch of a SO Kenya for the ubergeeky TGP particle sizing and SEM experiments.

OTOH if you're taste testing, and you normally drink a certain blend, it makes sense to use that blend. Then you just have to taste enough samples to overcome the stochastic variation.

RapidCoffee wrote:stochastic variation.

aka "the M&M effect" ?



jon

RegulatorJohnson wrote:you will get a variation of flavors regardless. the different size beans will roast differently. giving different mini blends based on the overall blend. one to get a consistent blend in every shot would be to count each variety of bean and make a specific mix each shot.

it doesnt matter if you shake the jar or scoop it will still be a consistently inconsistent mix of the blend shot to shot.

in experiments it seems like eliminating variables will produce the most consistent result. using a single origin should IMO make for a more consistent less varied flavor profile in the cup. its one less variable.

jon

It does actually matter if you do, since bigger particles will "flow" to the top of the jar if you shake it laterally. Try it with musli sometime. So then when you scoop out the beans from the top layer in the jar, you would get more of the bigger beans than the smaller ones.

Stuggi wrote:It does actually matter if you do, since bigger particles will "flow" to the top of the jar if you shake it laterally. Try it with musli sometime. So then when you scoop out the beans from the top layer in the jar, you would get more of the bigger beans than the smaller ones.

you are correct.. i misread your point. sorry.

i was thinking about but i in my head used tennis balls and basketballs.

jon