A Note on Calculating Espresso Brew Ratios

Beginner and pro baristas share tips and tricks for making espresso.
User avatar
another_jim
Team HB
Posts: 13871
Joined: 19 years ago

#1: Post by another_jim »

Brew ratios since the time of the CBI have been defined as the water to coffee ratio. But there has been a confusion in applying this to espresso brew ratios. The confusion is that we are using water in for brewing, and water out for espresso. Since the puck holds about twice its weight in water, this results in a large discrepancy.

For instance, the traditional Italian espresso is 25 grams of water to 7 grams of coffee, or roughly a 3.5:1 ratio. We have been making fun of this tradition, since we are assuming it talks about liquid in the cup. But this 3.5:1 brew ratio is a (25-14) grams to 7 grams ratio, or a roughly 2:1 cup ratio as we have been using it. A ristretto is classically defined as 7 grams coffee to 15 grams water, and that comes to a (15-7):7 or a 1:1 cup ratio, again exactly as we make it.

Suggestion one, stop making fun of the Italians. Suggestion two, lets discuss how we should measure espresso strength. Automatic flow measurements will work with water going in; so it may be better to go back to the traditional measures, rather than the cup weight measures we are using here.

For the geeky: here is the equation that converts cup weight to brew ratio:
  • (espresso.in.cup.weight + 2* coffee.weight)/coffee.weight
Jim Schulman

MWJB
Posts: 429
Joined: 11 years ago

#2: Post by MWJB »

I see your logic but it's not quite right. We don't really know how much water comes into contact with the puck, so we can't know how much is retained in the puck (without weighing before & after) and the 2g:1g ratio isn't constant even for brewed methods, it's a best average (to make the CBI chart work you have to make assumptions) & applies to non pressurised methods.

The Italian/traditional espresso (though I hate the latter phrase, pull the shot as long/short as you have to, to hit the desired yield) was defined originally as less concentrated drink than is typical today, with more water per g of grinds pushed through the puck.

It isn't uncommon to see Asian pourover recipes that focus on the beverage weight yielded, rather than the water added.

vze26m98
Posts: 264
Joined: 10 years ago

#3: Post by vze26m98 »

MWJB wrote:We don't really know how much water comes into contact with the puck, so we can't know how much is retained in the puck
Just a newbie here, but with a machine like a Pavoni, don't you know the amount of water going in? The lever's action fills up the group head to a fixed amount, shuts off the water and then presses it through the puck.

In a perfect world, the volume of the water from the group head would be either in the cup or in the puck.

Best, Charles

MWJB
Posts: 429
Joined: 11 years ago

#4: Post by MWJB »

Well, I guess you could work from the swept volume of the group, *if* you were sure that all available space was filled with water?... You see we are already into the realms of guesstimating, rather than "measuring" (which is easier if you can actually weigh the water added, as you can with some manual brew methods - though I don't see why you couldn't with a Mypressi/Rossa too?). However, the beverage itself (in percolation methods) only consists of water that has been through the puck/bed & landed in the cup...we can't taste any water/coffee retained in the puck/bed, even if that water contains coffee TDS it's pretty much irrelevant to the taste of the beverage itself and could vary considerably for various reasons (fines, wet puck/dry puck, etc.).

For what it's worth, the only situation where I can guarantee that my grinds bed holds 2g of brew water/1g of grinds is when I use my Aeropress as a pourover. I use 10g of coffee, lightly tamp the bed just to stop it from falling apart & coffee dripping through when the water is added, add 167g of brew water, fit plunger, wait a minute or so, press slowly for 1 minute and stop when I get 147g in the cup. So far (with the same, dialled in grind, differing beans), this has kept me in the ideal box, but extractions are still 20% +/- 2%.

User avatar
Jacob
Posts: 367
Joined: 18 years ago

#5: Post by Jacob »

A while ago I made a little ten shot test to see if it's possible to predict the in cup weight based on the weight of coffee dose and the amount of water used (measured by the build in flow meter).

I haven't been able to find an algorithm using this little dataset!
Shot	Total	Coffee	Water	Time	MinFlow	W(minF) T(minF) W(100f) T(100f)
   9	-2,40	 0,70	 0,10	-3,14	-69,50	 18,00	 1,72	 4,60	 0,13
  10	-1,10	 0,10	 0,10	-1,26	-28,50	  2,00	 0,20	 0,60	 0,14
   4	-0,30	 0,00	-0,90	 0,20	  4,50	 -4,00	-0,70	-2,40	-0,22
   2	 0,10	-0,30	 0,10	-1,69	-49,50	 -3,00	-1,17	 1,60	 0,08
   6	 0,20	-0,20	 0,10	 1,45	 39,50	 -1,00	 0,23	 0,60	-0,01
   1	 0,30	-0,10	 0,10	-1,00	-41,50	  4,00	 0,08	 2,60	 0,25
   7	 0,50	-0,20	 0,10	 0,97	  6,50	 -1,00	 0,38	-1,40	-0,08
   8	 0,50	 0,00	 0,10	 1,19	 34,50	 -7,00	-1,01	-1,40	-0,08
   3	 0,90	-0,10	 0,10	 0,75	 -7,50	  2,00	 1,10	-1,40	-0,01
   5	 1,30	 0,10	 0,10	 2,54	111,50	-10,00	-0,81	-3,40	-0,18
Values are the difference from the average value

User avatar
endlesscycles
Posts: 921
Joined: 14 years ago

#6: Post by endlesscycles »

another_jim wrote:...

Suggestion one, stop making fun of the Italians. Suggestion two, lets discuss how we should measure espresso strength.
The same as we should discuss coffee. dose(g), yield(g), strength(%).
-Marshall Hance
Asheville, NC

User avatar
another_jim (original poster)
Team HB
Posts: 13871
Joined: 19 years ago

#7: Post by another_jim (original poster) »

Jacob, subtracting the average destroys the dataset for the purpose of ratio calculations; that's why there is no pattern. For instance, raw data: 3/2 = 1.5 and 2/1 = 2. while the deviations are .5/.5=1 and -.5/-.5 = 1. Ratio calculations are log-linear, and for hypothesis testing you need to convert the raw numbers to logs before normalizing.
Jim Schulman

User avatar
another_jim (original poster)
Team HB
Posts: 13871
Joined: 19 years ago

#8: Post by another_jim (original poster) »

Thanks for the comments.

It's true that any water in the head space above the puck probably does not brew at all. But the water in the puck will hold some dissolved solids, although not as much as the water in the cup (since the "early" water is in the cup).

This means that brew ratio, no matter how it's calculated, may not be very repeatable from one machine to another, or even for different basket and dose combinations.
Jim Schulman

User avatar
Jacob
Posts: 367
Joined: 18 years ago

#9: Post by Jacob »

another_jim wrote:Automatic flow measurements will work with water going in; so it may be better to go back to the traditional measures, rather than the cup weight measures we are using here.
As I understand it you would like to use the build in flow meter instead of a scale placed on the drip tray for calculating brew ratios, right?

User avatar
another_jim (original poster)
Team HB
Posts: 13871
Joined: 19 years ago

#10: Post by another_jim (original poster) replying to Jacob »

That would be in line with how it is measured for other brewing methods. Using the 2 grams water retained per 1 gram coffee rule of thumb, it also makes the classical water to coffee ratios for espresso, from 4:1 to 2:1 compatible with the 2:1 to 1:1 ratios we see when taking cup weight.

However, as others have pointed out, we do not know how well the rule of thumb, i.e. two units water stored by each unit of coffee, works for espresso. Your raw data, properly labelled, and completely unmassaged, would be useful in this. I spent several years helping on surveys at NORC; and there are some absolute commandments -- always label every data row and column completely. Never change anything. Then lock the orignal data away. When you do an analysis, use copies of the data and make changes there. That way there's no need for do overs
Jim Schulman

Post Reply