I took Jim's comments from the context that the variables in espresso are so intertwined that its impossible to change just one variable without affecting another.

kschendel wrote:And what is one to do if one doesn't know that it's too low dosed? just that it doesn't taste good? ...

If I make a bad shot now, I come up with a hypothesis, and test it by making another shot with the required changes. When I was newbie and I made a bad shot, I came up with a hypothesis, and tested it by making another shot with the required changes. The only difference is that I did a lot more theorizing and testing back then.

On a bad shot my first question seems like the completely natural one: "How did this awful taste get into the cup, why didn't that good one make it in?" This in turn leads to natural hypotheses: maybe the shot is over or under extracted, burnt, etc. Some of these diagnoses can lead to correcting one variable, but most will lead to correcting several at once. But the point is not to correct the variables, it is to correct the taste.

Now, if someone could only tell me how to correct the taste one variable at a time, for instance, add a bit more chocolate, take out a pinch of the walnut, etc, etc, I'd definitely stop being so stubbornly multivariate.

I hate to encourage this not very useful offshoot of my original thread, but I do want to hop in with one small observation. From my research on espresso making, I can think of a situation where I would probably change two variables at the same time: the pour is too slow (or too fast) and there's clear evidence of channeling. If I saw that, I'd most likely adjust the grind to get the pour time in the right range and would also try to figure out what went wrong with the puck and try to fix it. I wouldn't just fix one or the other.

That's a simple case, and I'm sure there are more complex cases that may not have sufficient visual cues. I'm not yet in a position to to know what went wrong just by tasting the result. There are some general guidelines on that in the How-Tos, but there's a vast difference between reading and tasting. I'm looking forward to learning how to correlate what I see and what I taste with the coffee, grind, dose, distribution, tamping, temperature, pressure, etc. If I get good at this, I hope I can effectively deal with more than one variable at a time.

I fall somewhere in the middle on this debate. For the most part, I would try to hold as many variables constant while changing just one. But if I was confident that more than one variable needed to be adjusted, I'd go ahead and do it. The assertion that newbies should stick to one variable assumes that all newbies are incapable of diagnosing multiple causality. I think that's just wrong. Depends on the newbie.

another_jim wrote:If I make a bad shot now, I come up with a hypothesis, and test it

When I was newbie and I made a bad shot, I came up with a hypothesis, and tested it by making another shot with the required changes.

On a bad shot my first question seems like the completely natural one: "How did this awful taste get into the cup, why didn't that good one make it in?" This in turn leads to natural hypotheses:

Jim

Sounds like you're doing experiments to me
It's the method of "testing it" that may or may not be following accepted scientific methods

There are five main variables - water temperature, grind, tamp, dose, and beans. In my opinion, changing one of these at a time is good advice for those who have achieved at least a basic level of competence. Its probably true that there are interactions between the variables, but that doesn't mean we throw the baby out with the bath water.

A simple example is temperature. If I make two consecutive shots, keeping G,D,T,B the same, and I chance the temp. a few degrees, I'll taste the difference. If I change grind and temp, how will I know which change caused the taste difference?

I give up; I'm completely flabbergasted that this is so hard to understand.

Here's my last attempt:

Case 1: I make a change, one variable, two variables whatever. The taste changes. If I cannot describe how the taste changes, just that I either liked it or disliked it. Now it does not matter what I changed, I am completely in the dark. I will need to search over the entire extraction space and find a spot I like. It does not matter how I do that search. Any exhaustive procedure will work equally well.

Case 2: I make a change, one variable, two variables whatever. The taste changes. But now I can describe how the taste changes. I may decide that I want something more extracted but equally concentrated. So I lower the dose and tighten the grind to get a slower flowing shot. Or I decide I want it less acidic and sweeter, so I go hotter and more ristretto. These are the most reliable ways I know for getting these effects. YMMV.

Case X: You tell me ... Give me an example of a poor shot, and a partial state of knowledge or ignorance, either about the equipment or about how to characterize the shot, which will make the one variable at a time procedure more optimal than any other method for correcting a shot.

Case X: You're actually trying to learn about the impact of changes - not simply "correct the shot."

I give up; I'm completely flabbergasted that this is so hard to understand.

Me too, from the other direction.

Case 1: I make a change, one variable, two variables whatever. The taste changes. If I cannot describe how the taste changes, just that I either liked it or disliked it. Now it does not matter what I changed, I am completely in the dark. I will need to search over the entire extraction space and find a spot I like. It does not matter how I do that search. Any exhaustive procedure will work equally well.

No arguments there.

Case 2: I make a change, one variable, two variables whatever. The taste changes. But now I can describe how the taste changes. I may decide that I want something more extracted but equally concentrated. So I lower the dose and tighten the grind

Now you're presupposing that I already know that "lower dose AND tighter grind" means "more extracted". How did I learn that?? (other than by reading what other people claim, which is valid enough as long as they are correct) Take it the other direction; I have a taste, I now lower the dose and tighten the grind. The taste changes and I can describe it in some way, "more concentrated". Which variable caused the taste change? How the devil am I supposed to know whether it was dose, grind, or both, without trying them individually?

If there is a universally applicable, universally accepted set of rules that say "change X, taste changes like Y" for espresso and any single or multiple combination of X's, then we could skip the whole one-variable-at-a-time thing and send the newbie to The Book. I get the strong impression that no such Book exists.

another_jim wrote:I give up; I'm completely flabbergasted that this is so hard to understand.

I think you are confusing "hard to understand" with "don't agree with you."

malachi wrote:Case X: You're actually trying to learn about the impact of changes - not simply "correct the shot."

Exactly! It's only by learning about and understanding the impact, the role, of each variable that will permit the person pulling the shot to (eventually) become like Jim Schulman -- and know that the shot is BOTH low dosed AND too slow, as opposed to Uh, gee, I dunno, there's something not quite right about it . . . (or worse, that they perceive nothing is off about it at all; think that's as good as they will ever make at home; and continue to go to Starbucks to have something "better").