Fingerstrike Dosing = Overdosing ???

[Jim Schulman]

Andy and I have been weighing pucks before and after shots, trying to measure the amount of coffee extracted in an espresso shot. This is still a work in progress, but an important result is firming up that the Barista community should know about:

Fingerstrike dosing almost certainly leads to underextracted shots.

Illy cites extraction rates of 20% to 25% (that is, 20% to 25% of the puck's dry weight ends up in the cup). With fingerstrike dosing Andy and I usually got around 15% to 18% extractions. When we started varying, in particluar when we started using 12 to 14 grams in those LM, Cimbali, and Faema doubles, or 6 to7 in the singles, and used the much finer grind that required, the extraction rates went up above 20%. Not surprising, since these are the doses the baskets are acutally designed to do. We still have more work to do to totally nail it; but the result seems fairly solid.

Does this make any difference to the taste of the shots?

I would give an emphatic yes! Caramels and sugars don;t dissolve as readily as bright and some bitter compounds. Other bitter compounds like clove or tobacco also dissolve slowly but assualt the palate at even low concentrations. Use 13 grams in that LM double, grind it so it makes the same size and time shot, and it will taste a lot sweeter. If you have a blend that's coming out real sour or bitter, this may tame it.

Abe Carmeli got some roasted samples of Sandalj's preblended green espresso blends (Sandalj is a Trieste coffee importer). Their top of the line Vivaldi blend was roasted to cinnamon, barely out of the first crack. At first I thought they had sent it for cupping, but the other samples were all for esresso. Yep, doing it as 6-7 gram singles or 12 to 14 doubles made it perfectly lovely. One could taste the brightness, but it was like very sweet lemonade.

I think fingerstrike dosing locks a cafe into medium roasted blends with subdued acidity and bitterness, since the extraction isn't sweet enough to handle anything else.

Give it a try, and tell me I'm full of it!

[Robert Goble]

Can you elaborate on your methodology for weighing and recording the pre-brewed & after-brewed coffee?

[mandy]

This is really interesting, however I'm not clear on what "fingerstrike" dosing is? What method of dosing are you using to achieve 13-14 grams in a double basket? Is the dosing technique or simply the weight of the ground espresso that is problematic? Are there variations in your dosing techniques or is the focus only on the proper weight and proper extraction times.

[Jim Schulman]

Can you elaborate on your methodology for weighing and recording the pre-brewed & after-brewed coffee?

I have the basket outside the PF for dosing, tare it on a 0.1 gram scale, dose, and weigh. I grind a weighed dose of beans. For instance, when I want to dose 14 grams, I'll grind around 14.1 to 14.2, and end up with 13.8 to 14 in the basket.

After the shot, I collect the basket, clean off the group into a cup, clean out the basket into the same cup, and bake at 275 for a few hours to dry the coffee. We also bake a sample of the fresh coffee, since it usually loses around 2 to 3 percent of the weight in this process.

I am not sure if we have the absolute extraction ratios correct, but, since we use the same procedure each time, we are sure of the relative effect -- finer grind of a smaller dose, same coffee, same basket, same shot time and volume, will extract a higher percentage of the puck.

[Jim Schulman]

I'm not clear on what "fingerstrike" dosing is? What method of dosing are you using to achieve 13-14 grams in a double basket?

By "Fingerstrike Dosing," I mean the method taught in Schomer's book: filling the basket loosely with ground coffee, and levelling by striking off the excess with ones finger. In an LM double, this gets around 17 to 18 grams. If one taps the basket before the levelling, one gets around 19 to 20 grams.

For this experiment, I'm simply weighing the basket empty and filled to get the doses I want. That's obviously not going to fly in a commercial setting. Luca has reported that some Oz baristas are using curved scrappers to level the basket, thereby scooping out ground coffee below the rim of the basket and reducing the dose.

For them the issue was not having the puck come up against the shower screen during the shot. As far as I can tell, one can go around 16 grams on an LM basket before coming against the screen, so dosing down to 12 to 14 grams will require a deeper scooping device.

For a simple experiment in the shop try this: Fill the basket with coffee, and use the rim of another 58mm basket to scrape. By holding the "scrape-basket" sideways, the rim goes inside and scoops coffee out of the basket you are dosing. My quick measure indicates that this will get you around 13 grams in an LM double, 12 in a Faema or Cimbali.

Be prepared to grind **a lot** finer, roughly halfway from the current setting to the grinder zero.

[Philip Search]

I could be seriously wrong, but my understanding is that simple structured sugars disolve much faster than bitter/bright components. Is there a reference for this?

Hmm... but the fine grind, I think this is the key to the increased sweetness. I just pulled a couple of pounds of light roasted espresso coffee, weighing and tweeking my dose and grind and at 20-21 ish grams in the double basket the grind begins to get too coarse when one adjust to get a reasonably accepted extraction time (around 25-30 seconds). More surface area exposed = more sweetness?

[Jason Haeger]

I could be seriously wrong, but my understanding is that simple structured sugars disolve much faster than bitter/bright components. Is there a reference for this?

Hmm... but the fine grind, I think this is the key to the increased sweetness. I just pulled a couple of pounds of light roasted espresso coffee, weighing and tweeking my dose and grind and at 20-21 ish grams in the double basket the grind begins to get too coarse when one adjust to get a reasonably accepted extraction time (around 25-30 seconds). More surface area exposed = more sweetness?

If the sugars are first to go, wouldn't increasing the surface area of extraction also increase the amount of extracted sugars/flavors in the cup?

Common sense points to "yes".

I agree with this line of thinking, Jim.

The only problem is that not every bean has the same density. Consider density characteristics of origin, and then origin percentages in a blend. Then factor in roast profile and percentage of total weight loss. Doesn't a lighter roasted coffee retain more weight than a medium roasted coffee?

What about roast time. Wouldn't a longer, yet still light, roasted coffee lose more weight than a quickly roasted coffee of the same agtron shade?

Let's assume that most espresso blends are, on average, a relatively slow roast to FC-FC+. Let's also assume that good number of these have a good amount of Brasilian coffee as a base(known to be of lesser density). What is the average volume for a 14 gram dose?

In that case, doesn't the "finger strike" seem semi-accurate? I should like to add, that Schomer's method seems to updose a bit more than a non-compacted finger strike technique. From what I understand, he "pushes the shit out of it".

On the other hand, let's look at coffees like Terroir's Ethiopia Addis Ketema SOE. To achieve the ~14 (or ~16 on a second go) gram dose, I had to dose the basket only about 2/3 full. Distribution was sketchy, to say the least, but the resulting espresso was brilliant.

Of course, the particular coffee was quite dense, and small in size.

How much have things changed in regards to average bean density since the introduction of the "finger strike" to the scene?

[Jim Schulman]

I could be seriously wrong, but my understanding is that simple structured sugars disolve much faster than bitter/bright components. Is there a reference for this?

My undersrtanding is that even at a cinnamon roast, most of the sugars have caramelized. In the drinkable roast range, one is dealing with light to dark caramels, not sugars. These dissolve more slowly (the flavor wheel is also the order of solubility).

Bitter distillates like spice and peaty flavors dissolve slowest, but if I understand the chemistry correctly, they have the lowest sensory thresholds -- a little goes a long way in taste. Bottom line, acidity is certainly tamed with added sweetness by down dosing and grinding finer on light roasts; but I have no idea how it affects the balance of sweetness and bitterness in dark roasts. The idea of running a long series of tests on taming a Starbucks level roasted coffee is unappealing, to say the least.

Let's assume that most espresso blends are, on average, a relatively slow roast to FC-FC+. Let's also assume that good number of these have a good amount of Brasilian coffee as a base(known to be of lesser density). What is the average volume for a 14 gram dose?

I wasn't getting that precise.

I'm doing my extraction tests by weighing the beans, since this is more convenient, going in stages from 12 to 20 grams. My impression is that a finger strike dosed LM double will hold 17 to 18 grams if the coffee is uncompacted, and around 20 if it is. The deep Cimbali, Faema style, and Synesso baskets will be very similar. I'm sure it can vary by a gram or two given different roast levels and bean densities.

My main point is that these baskets are not designed to be dosed in this way, but to be dosed with 12 to 14 grams from a doser. It is at this dose, and the grind this requires that one sees 20% to 22.5% extraction rates. The rates ones sees for finger strike dosing are around 17.5%.

Bottom line: in avoiding stale coffee from the doser, 3rd wave baristas are inadvertantly underextracting their shots. This limits how light a roast level they can use, and also limits them to coffees with a lot of sweetness, since they aren't getting all the sugars out using the current styles of dosing.

[Jason Haeger]

I hear "The Bottom Line".

I've been anti-updosing for awhile now, but I never considered a straight level, with gravity-fill distribution (aka - "focus on the landing") to be updosing.

I think I'll start weighing my doses. I need to buy my own gram scale first, I guess. It really is far overdue.

[Andy Schecter]

avoiding stale coffee from the doser, 3rd wave baristas are inadvertantly underextracting their shots. This limits how light a roast level they can use, and also limits them to coffees with a lot of sweetness, since they aren't getting all the sugars out using the current styles of dosing.

Perhaps supporting your idea is SL28ave's advice to keep doses below 15 grams in a double basket. Terroir's roasts appear to be one of the lightest '3W' roasts around.

[gscace]

Luca has reported that some Oz baristas are using curved scrappers to level the basket, thereby scooping out ground coffee below the rim of the basket and reducing the dose.

For them the issue was not having the puck come up against the shower screen during the shot. As far as I can tell, one can go around 16 grams on an LM basket before coming against the screen, so dosing down to 12 to 14 grams will require a deeper scooping device.
[/quote]


Yowdee:


My practice over the last couple of years has been to used a curved scraper to fill baskets by some more reproducible volumetric method than the finger swiping thing. I hadn't taken the step of making tools in a variety of shapes so that I could vary the dose, but I think I will now.

Your idea is very interesting and I think I'll give it a go and see what I learn. WRT Illy's numbers, I often ask myself "On what premise are these numbers based?" Is the optimum extraction fraction according to Illy based on using the coffees that Illy uses, or is it constant for all coffees. It's kind of like the canonical pressure thing - 9 bars is right, but where is that 9 bars measured, and under what flow conditions?

-Greg

[Jim Schulman]

WRT Illy's numbers, I often ask myself "On what premise are these numbers based?" Is the optimum extraction fraction according to Illy based on using the coffees that Illy uses, or is it constant for all coffees.

I'm not clear on this. The SCAA in their brewing experiments came up with 20% as optimal. Petracci in the Illy book cites 22.5% to 25%. Neither gave details on the type of coffee.

My initial take after playing with Terroir's and Sandalj's ultralite coffees, both for cupping and shots, was that the best balance was if one underdosed and overextracted light roasts, and consequently overdosed and underextracted dark ones.

I'm convinced relatively overextracting and underdosing very light roasts is the way to go; but I'm getting results that don't add up with very dark roasts. The dark roast results are indicating that the way we handle the bitter-sweet balance (is it ashy or XO?) is much more non-linear than the sour-sweet balance (is it lemons or Eiswein?).

The dry distillates extract more slowly than even the caramels, which speaks for underextracting dark roasts. But if our 'buds saturate on them at low concentrations, the added caramels of a low dose/higher extraction can make the taste less bitter, even though there's a higher proportion of distillates in the cup.

In this case, optimum dosing would be a curve: the highest dose and lowest extraction for medium roasts, and lower doses with higher extractions at both darker and lighter roasts.

[Philip Search]

Where I'm at right now with this is:

1. Measurements and roast styles. The roast parameters, including drum speed, airflow, roast curve, etc. seem to change the way sweetness is extracted and percieved at various roast levels. Here is what I know to be true: Moisture, oil, protein, total carbohydrate, total sugar and starch content of a coffee are all determined as a function of roasting time. Moisture, and total sugar content are perhaps the biggest value shifters due to roast time, with amino acids being next. This value shift will change what is available to extract, perhaps making moot the debate over "correct" dose, and shifting it to certain coffees should be dosed certian ways to give X results. It would be great if we could use some very cool modern chemestry tools and create a detailed chart for various desity of beans, sort of a generic reference form. We would need the tools to do gas chromatography, and thin layer chromatography.

2. The grind issue is being under analized. The finer/coarser you go the more you change the shape and particle size ratio of the grind.

3. The effect of various types of equiptment on taste/extraction/brewing. This would include pressure/pre-infusion, etc.

In no way am I trying to dog or dispute this great research being done. I think it is great, just trying to think it through.

[Shannon Hudgens]

For what it's worth the doser lid of a Super Jolly will scoop out a fifteen gram Faema basket if you twist it instead of just swiping it across (so it looks like a convex tamper has been pushed into the puck slightly).

That was using an espresso with the density of Terroir's Addis Ketema Yirg espresso (which is what I was using).

[gscace]

For what it's worth the doser lid of a Super Jolly will scoop out a fifteen gram Faema basket if you twist it instead of just swiping it across (so it looks like a convex tamper has been pushed into the puck slightly).

That was using an espresso with the density of Terroir's Addis Ketema Yirg espresso (which is what I was using).

That's very useful to know. Now I can be lazy.

-Greg

[Jim Schulman]

Tell me when you're sick of this ... I have graphs.

I spent the last four days doing shots at various doses. I weighed the pucks and worked out the extraction percentage; I timed and weighed the shots; and I rated the taste for acid versus bitter and dry to sweet. I used a 50/50 blend of Idido and Cenaproc roasted to 435, aboiut 5F below the first pops of the second -- a light roast, but not sunglasses light. I did 4 shots each at 12, 14, 17 and 20 grams in an LM triple basket (no ridge and perfectly cylindrical for easy low dose tamping; plenty deep, so no shower screen collisions even at 20 grams).


The relation between extraction percentage and shot time, volume and dose is pretty tight:


Note that the X axis is logarithmic.



The relation between the sour/bitter balance and dry/sweet to extraction percentage is not nearly as tight, buit still present:




The conclusion: you can certainly manipulate the extraction percentages by varying the dose and keeping the shot time and volume relatively fixed. Whether this affects the taste as systematically as I think is more doubtful; you'll have to decide for yourself.

[malachi]

Interesting and impressive.
But the concern is that this may not apply in general.
Much as Schomer's statements about temp and dose etc were dependent upon his coffee (and desired flavour profile), or the same with Illy... the challenge is that what works for one coffee (and one desired flavour) may not be applicable for others. The risk is that people treat their unique results as possessing universality and that others then apply them as a "golden rule."
Universality is, IMHO, a trap -- as are derived rules. Understanding the relationships... not just the static variables but also the connections and interconnections... and then creating generalized concepts and structures for experimentation within (on a coffee/target profile basis) is far more likely to prove transportable.

[Jim Schulman]

Interesting and impressive.
But the concern is that this may not apply in general.
Much as Schomer's statements ...

I think there are very few rules like Schomer's, which set absolute levels of dose, temperature, pressure, roast, etc. that apply outside their specific context. To be fair to him, he doesn't pretend they do: he's advocating the complete Vivace package of doing things. Trouble is people do apply his rules piecemeal, and then wonder why they don't work out very well.

I try for more generality. The shot making rules that derive from graphs like this are either diagnostic -- if the shot is too sour/bitter/dry/sweet; do this -- or conditional -- if the coffee is roasted light, do one thing, if it's dark, do something else. I think conditional "rules" like this are far more generally applicable, although there's always the cool exceptions that get me doing the next round of research.

There are certain things that seem quite specific, and not subject to any sort of rule. For instance, should the puck touch the shower screen or not? On my new Elektra Semi, which has the same group bell as the A3, it definitely should not, the shot instantly goes to hell in an astonishingly clear fashion (1/2 gram dose increase to go from nectar to dreck). After I found this, I dragged out the E61 Tea, and there it seemingly didn't make the slightest bit of difference. On LMs, I keep hearing arguments one way or the other; so it'll be interesting to figure out what happens in them.

[Jaime van Schyndel]

jim,
I have to say I am struck by the irony of this paper being written by you of all people. it's a good thing but funny to see.

It seems that you were one of the bigger critics of 'underdosing' and 'bright' coffee (or at least the biggest critic of one eccentric proponent) and now you wrote a paper where you are in essence going to give validity to it.

I still don't really understand how you tie everything to sweetness and I don't follow where you are quantifying sweetness other than the caramelization bit. If you could point it out I would appreciate that. It's the one part of the paper that sort of sticks out to me.

I am afraid the message many will take away is that lower dosing is the new triple ristretto. A fatiguing debate. I would ask that you push hard to make everyone understand there is a sweet spot for every bean and avoid the monolithic approaches where someone says 'that's not espresso' or 'if it isn't this volume, it's not espresso'.

I'm still not with you on the french curve any more than I would suggest using two sizes of tampers to tamp, I want better designed baskets at different volumes and with adjusted hole patterns that come like ina range sizes. 10g through 24g in increments. If I lose my feel of the coffee, I lose a lot. Old topic though.

[Jim Schulman]

I'm mainly saying that dosing changes, and quite radical ones from 12.5 to 20 grams in a single basket (or hole area), should be part of every barista's repertoire. My finding is to expect that acidic and bright-bitter flavors become more prominent at high doses and more balanced by sweetness (I'm thinking slow dissolving caramels) at lowered doses.

This means one can do lighter roasts and more acidic coffees for espresso than if one just stuck to the usual 17 to 18 grams customary in most of the US. Obviously George Howell and Peter Lynagh are way ahead of me on this.

My quarrel with highly acidic coffees is that they smelt like dessert and tasted like aperitifs. Controlling the extraction may bring smell and taste into line.

I have no quarrel at all with very dark roasts; I don't drink them.

[Jason Haeger]

Wouldn't it have made more sense to simply declare the statement and end with a "try it" rather than going through all this mess to arrive at the conclusion that others have already touted as fact, yet not "preached", as it were, on the merits of such practice? (...that's a long sentence)

Not that this project is pointless, but it seems that there are far too many variables influencing the results to make any conclusive statements, and at this stage in our progress, I'm not convinced that anyone COULD come to any conclusive evidence due to our lack of variable control abilities.

I also feel the need to ask what constitutes as "over" and "under", and according to who?

Now, I do agree that underdosing(by volume) is better, all this assuming that the roast profile is light.. really light, but that's not to say that there is absolute truth on the standard dosing paramter in terms of weight and/or volume. I've read a book or three that says otherwise, but the funny thing is that they don't always agree, and neither do baristas.

Not to argue or bring up a debate, but I'm not sure this is as conclusive as it may appear to be.

[SL28ave]

Wouldn't it have made more sense to simply declare the statement and end with a "try it" rather than going through all this mess to arrive at the conclusion that others have already touted as fact, yet not "preached", as it were, on the merits of such practice? (...that's a long sentence)

I've been praying for Jim's conversion for a long time. I'm just happy he's come to the conclusion. Some people, including me, need to work through weighty messes to convince ourselves; I imagine the same may be true with Jim.

Also, while our results are almost the same, his "mess" shows just how wide the chasm between our results is. He's providing a few super-valuable hunks of meat for me to chew on and eat, including insight on how to REALLY convert him.

[Rich Westerfield]

I'm mainly saying that dosing changes, and quite radical ones from 12.5 to 20 grams in a single basket (or hole area), should be part of every barista's repertoire.

Jason,
As I read it, Jim said the above quote is the object. Are you arguing with that?

Naturally nobody has the time to taste every permutation of every coffee or roast, but if you're sticking with just a couple of espresso profiles, dose surfing seems a somewhat useful training exercise, right up there with separating the shot. It's one more way to help explain how a shot got to what it is. If something like this gives a barista more knowledge on how/why to adjust a dose for a specific situation if they ever find that need appropriate, why take issue with it.

Besides, we want to encourage Jim to keep up the experimentation no matter how far afield. Hey, if he lands on some combination of variables in technique and equipment that uses 12.5g and produces as good a shot as our current updosed technique (18.5g), we can cut our COG by 40% or more.

And then we can finally afford to send folks to origin ourselves!

Keep it up Jim.

[Jaime van Schyndel]

We need the baskets. It would be the greatest lab accessory one could have, then I could scrap my collection of different doubles, even the 10g and work from one standard set.

Tacy, you must agree with this?

I'm mainly saying that dosing changes, and quite radical ones from 12.5 to 20 grams in a single basket (or hole area), should be part of every barista's repertoire. My finding is to expect that acidic and bright-bitter flavors become more prominent at high doses and more balanced by sweetness (I'm thinking slow dissolving caramels) at lowered doses.

So, about the sugars...

How do you make a statement, any statement, about sugars or sweetness unless you can better quantify the meaning of sweetness. You must explore understanding where and when sugars are appearing and at what roast levels(if at all).

it's a loaded question. I know, because I came down the same path trying to understand what was the variance between a unique style like George and a classic style like Barnett. I came to some hard conclusions after spending a lot of time and plenty of money out of my own pocket.

[malachi]

Hey, if he lands on some combination of variables in technique and equipment that uses 12.5g and produces as good a shot as our current updosed technique (18.5g), we can cut our COG by 40% or more.

So you're looking for a generalized rule that combine technique and equipment and dose weight and is universal?

Good luck.

More likely will be:

"For the Stumptown Hairbender roasted on 1 March and stored in paper, unsealed, but in the dark, in Portland on 5 March at 10am in a location at 500 feet elevation - with ambient room temp of 71 degrees and ambient humidity of 47 percent - using a Robur with burrs that have had 117 pounds ground through them - on the middle group of the #14 3-grp Mistral set at 9.1BAR and 189.5F with a size 6 gicleur that had been cleaned and scrubbed and then had 5 shots pulled from it - I found that to my personal taste (see attached notes, taste-bud test results and both personal and cultural backgrounds as well as information on age, smoking and alcohol habits) and after drinking flat water (Panna brand) and eating a single soda cracker (but not brushing my teeth in the morning) a combination of a 19gram dose in a stock (ridged) LM double basket (1 month old) with stockfleth distribution and a 50 pound single tamp (with no polish) extracted to 1.72oz in 27 seconds was ideal."

And that leaves out water, building plumbing activity, etc, etc.