FWIW... my results from the last few days of tasting.

Coffees: Ecco Caffe Reserve, Ecco Caffe experimental 60/40 blend, 4 Barrel espresso

Protocol: 3 different tasters. Espresso prepared in 3 manners ("traditional", with crema swirled to combine into cup, skimmed). Initial tests were blind (in this case using blindfold in fact). Secondary tests were not.

Preferences: All 3 tasters expressed a slight preference for the "swirled" espresso. This was true both blind and not. There was no agreement on which was "second best" - with 1 taster preferring "traditional" when blind and 2 preferring "skimmed" (oddly this flip-flopped when not blind). 1 taster expressed a strong preference for traditional over skimmed - while the other 2 felt only a very slight preference. All agreed that the difference in experience between the two was very large ("like they are different drinks and different coffees").

Weakness in testing: The primary weaknesses in the testing were: too few tasters, limited coffee selection and (most of all) sequencing. It became clear that which shots were served in which order had a huge impact on the results. If a "traditional" shot was served after a "skimmed" shot, the responses were more negative than it if was served after a "swirled" shot or as the first shot in a sequence. Finally, shots were prepared using a naked portafilter - producing significantly more crema and a different mouthfeel and profile. As a result of these issues - I feel like my results are very preliminary.

My own opinion: While my tasting was not blind, of course, I have some thoughts on the topic. First - given the coffees I used, and given that I was using a naked portafilter, I actually found the "quick swirl" a valuable tool. This removed some of the "fuzziness" that I don't like in a shot from a naked portafilter and produced a shot that was more consistent in flavour from first sip to last. I don't know if this would be true if using a spouted portafilter. I found the "skimmed" shots to be very interesting - but also very coffee specific. With the Ecco experimental blend, the resulting shot was IMHO enhanced as compared to the traditional shot. With the 4 Barrel, however, the result was unbalanced and bland. The biggest negative I found in all three coffees when preparing a shot in this manner was that sour notes started to emerge that could be off-putting. In the end, I think that the skimming approach is really valuable - but primarily as a tasting and evaluation tool. The (simplistic) analogy I would make is the obvious one - a comparison to brewed coffee and cupping. Skimmed shots, to me, allow you to clearly evaluate the coffee(s) in the espresso you're preparing and make good decisions on the result. I don't, however, feel like it's a particularly great way to simply prepare and consume the beverage (if you know what I mean). As a result, I (personally) will start using this technique whenever I evaluate a new espresso - but will then move to a more traditional prep once I start simply drinking shots. Now... when I say "more traditional" I have to admit that, going forward, I will do the "swirl" whenever working with a naked portafilter (and might well do the same with spouted ones - depending on the results of testing in the next few days).

By "crema swirled to combine into cup" do you mean swirling the cup after the pour, but not stirring? Or are you swirling the cup as the shot pours? How is swirling different than stirring?

Chris, Thanks for your detailed, if preliminary, tests.

I'll jump into the bucket by way of my own little test from this afternoon. I'm currently working through the last of my COE Rwanda Mariba, which you may recognize as Mike Phillips' WBC Blend. It's a pretty bright coffee for me, I would say the nicest shots I've had are characterized for me as sweet cherry-lemon juice with a hint of some funky stuff--the description says tamarind and ginger snaps. I sense the ginger dimly, dunno about the tamarind. Anyway, it's a great example for me of an extremely acidic coffee that's pleasurable for me inasmuch as it's really, really crisp.

So I pulled a pair of shots today, each shot being a double split with a spouted portafilter into two demitasses. The first shot, I skimmed the crema off the first sample and sampled it. Unexpected--bittersweet and actually quite pleasant in its own way. Then I sampled the skimmed shot and it was extremely acidic with a minimum of sweetness balancing it. The texture was like dense, juicy presspot--actually pretty nice, but... I swirled the second sample to mix the crema in a little and tried it. Definitely a balanced shot, though my palate enjoyed the bittersweet more than the comparatively more balanced swirled shot. So for me, the first sample got a nod to the swirled espresso, even after letting that shot sit for a few moments while I "processed" the skimmed shot.

For the second shot I sampled the unskimmed shot first. I stirred it rather than swirling. Juicy and heavy, really nice. Better than the previous shot by a bit. I skimmed the crema off sample 2, and sipped the crema. It tasted dull and slightly bitter by itself. Not as nice as the previous shot's crema, but nothing awful. By contrast, the skimmed shot was pretty nice--as expected from comments above, enhanced clarity plenty of sweetness to balance the acidity, and like before a really, really juicy texture.

I can't conclude anything yet, but I'd say my preliminary playing around with this suggests to me that this is a lot more complex than an assertion that crema is 'rubbish.' It surprises me that the coffee I used produced the tastes noted above, too--I had assumed it'd be a posterchild for what James described. I'm guessing that what Chris said is right on: skimming off crema will be a good way to try an espresso when evaluating it/approaching it initially. I don't expect it'll be something that can or should ultimately supplant enjoying espresso untouched or stirred.

i think tests should be done on split shots using the spouts to give to halves of one shot. then you either skim or stir or whatever to one of the 2 halfs. this would help remove the shot to shot variations that occur. you could even leave one half un touched and modify the other half and see what happens when you compare them.

jon

Jon, this is how I prepared my single experiment to date: pulled one double, using a pf with two spouts, and pulling the double shot into two separate cups.

* * * * *

Chris, I want to thank you also for your detailed report. I hope to add to my own tasting notes with the arrival of additional coffees this week. Your test results seem quite "predictable," not in their specific outcomes but rather in that it is a) too soon to dray any solid conclusion, and b) that the results were split -- at least between "skimmed" and "traditional." The fact that the blind and non-blind results "flipped" does not come as a surprise (at least to me). But you re-affirm the desire to stir my espresso, something I have now forgotten to do twice since this morning's post.

Cheers,
Jason

I did one double shot each of three different coffees at 14.5 gram doses yesterday, and one each at 16.5 gram doses today.

The coffees were
-- Yemen Anesi, a dry processed, fairly clean coffee with strawberry. chocolate and a small hint of leather when the shot is good.
-- Fraijanes Agua Tibia, a wet processed coffee with cherry, chocolate and peaty notes when the shot is good
-- Cappa blend consisting of 50% dark roasted Kenya with some medium roasted Bourbon or Typica, this time a Yellow Bourbon Brazil. The flavor is nuts, caramel, and a big clove kick (I'm not really recommending this, the flavor is probably more nostalgic than good)

Each shot was divided into two singles. I swirled the first one and drank some. Then I skimmed the second and drank some of that. I rated the second skimmed shot against the first on sweetness, body, high, middle and low flavors, and overall. I also took some short notes.

The results are as follows: the plusses and minuses refer to the skimmed shot, so a plus means that it was better.

Effect of Skimming the Crema
-------------------------------------------------------------------
Coffee       Overall Sweet Body  Highs Middles Lows   Notes
-------------------------------------------------------------------
Anesi 14.5     --      -     --    =      -     -   weak, awol
Anesi 16.5     ++      +     =     +      +     =   winey intense
Ag Tib 14.5    =/-     =     -     +      -     +   less complex
Ag Tib 16.5    =/-    =/-    =     +~     =     =   both sucked
Cappa 14.5      -      =     --    NA     =     +   watery, clearer
Cappa 16.5      -      -     =     NA     -     -   mocha pot taste
--------------------------------------------------------------------

The Agua Tibia is clearly a low dose SO, with the high dose shot tasting like peanut lemonade. For the 14.5 shot, I felt that most of the complexity was in the middle flavors, and these were enhanced by the crema; this is why I gave the edge to the crema shot. However, the highs and lows were cleaner and clearer on the skimmed shot; but the body was less. It behaves similar to the way James reports for his coffee, except that the middle flavors on this one may be a lot cleaner than on James's

The Cappa blend has no highs, and the crema shot performed similarly at both doses. The skimmed shot was watery at the low dose and had that over concentrated mocha pot taste at the higher dose. Here, I think the crema acts like a balance wheel, and the brew seems unstable without it.

The Anesi is a head scratcher. At 14.5 it behaved conventionally, the crema shot was very good, and the no crema shot was a watery joke. At 16.5, the regular shot was cartoonish and not well integrated. The skimmed shot tasted brilliant, like a complex wine. This coffee brews beautifully at a lighter roast, and I'm guessing it would do real well as Turkish coffee too. I have no clue why the skimmed tasted better. If a shot is already cartoonish and not well integrated with the crema, one would have expected worse without its buffering effect.

The shots don't show any consistent pattern except for body. At low doses, the crema adds to the body and mouthfeel, at higher doses it does not. Since taste clarity increases with lower doses; it could be that skimming the crema is something that is more likely to work well at higher doses: there the body is less affected, while an increase in clarity is more telling.

Whether one likes a shot more or less with skimmed crema is not as interesting as the rather non-obvious and irregular effects it can have on the taste. Our assumptions about what crema does for the taste may need to be looked at more closely. Sadly, my results don't not suggest a general rule of thumb on this. Do anyone else's?

Yah see? That's why I like Jim. Not afraid to call a spade a spade. No pretentions, no, holier palate than thou...
When both shots suck, he simply writes, 'both sucked'.
Thanks Jim, that provided a well need chortle this afternoon!

malachi wrote:Have any of you actually tried this (ie pulling a shot "standard" and tasting it, then pulling a shot, skimming it and tasting it -- and then comparing the two)?

Or is everyone just talking out of their a**es?

After watching the video blog, it prompted me to try the skimming method on both espresso and americano. The details are in my blog. I'm using a Quickmill Anita, naked portafilter, double dose, w/ Eric's Thermocouple (the start/end temps are listed). I'm using SO Brazil Moreninha Formosa, Full City +, 4 days old.

My results are that the crema only cup gives off Esmeralda-like tastes, esp. in the americano taste test. It's not truly americano 5:1 ratio, more closer to 2:1 or 1:1. But the crema for Brazil FC+ tastes really good, almost as if the coffee was roasted to a degree before 2nd crack.

One additional data point; no conclusions.

I prepared a double, split into two cups. Having already decided yesterday that, with this coffee (Vivace Dolce), I prefer my shot with crema than skimmed, I tried today "traditional" versus "stirred."

The "stirred" shot was certainly different than the "traditional" one. I'm not sure if "better" or "worse" applies. I know I preferred the traditional shot, but I cannot discount the fact that I am also used to the traditional shot. Thus it was "difference" that stood out, not a distinct "plus" or "Minus" -- just "different."

This morning, I repeated the experiment of skimming off the crema vs. a "traditional" shot, this time with coffee from Klatch Roasting (2008 Competition Blend), as opposed to Vivace (Dolce).

Same methodology; same results. As I said above,

zin1953 wrote:To my palate, Cup A ("skimmed") tasted much more like a great sip or two from a press pot, and not like an espresso. Yes, it does lose some body; yes it is lighter and a bit sweeter on the palate. But it also misses something on the mid-palate and has a more delicate, cleaner finish. Cup B ("traditional"), on the other hand, was a classic, rich, full-bodied espresso -- not overly bitter, yet there is a detectable, balanced note of bitterness to compliment the inherent sweetness (no sugar added; both cups were consumed "straight"), the mouthfeel was fuller, more velvety than silky, and the finish much longer, richer, and more lingering.

Cheers,
Jason

i always use "shot glass and pour". this mixes three layers together and helps me to get the wholeness of the espresso in first sip.

I tried it again at another coffee shop using their house blend. I also tried it with an Americano. The blend this time was, obviously, different from the one at the previous coffee shop that I tried. This blend was also a lighter roast than the other coffee shop from before. I followed the same pattern as before: double shot into two cups, scoop the crema off of one and drink it first, drink the one with crema second, both were mixed before drinking.

1. Crema is bitter.
2. I liked the espresso with the crema better this time.
3. I liked the Americano without the crema better.

I have no clue if it is a blend thing or a roast thing, but the results were opposite from before.

Does anyone else have a better method for doing this experiment?

I wonder if the skimming test might give a preview of how a coffee might age? Obviously oxidation will change other things besides crema and CO2 volume, but just thinking aloud (so to speak) about how one might generalize the use of such a test in practice to determine coffee specific qualities (other than the obvious taste and mouthfeel effects of the crema of a particular bean at a particular age with a particular pour style)... Hmmmm....

You know, reading over some the tasting notes a lot of the descriptors look like the flavours you get from a Micro Casa Leva. A low pressure shot with a thin crema but with a clean and clear flavor profile. No real point other than a general observation. The spring lever pulls shots well below 9 bar and produces a thin crema but the drink itself has a very unique flavor. Less body than your typical espresso but with a flavor that is more along the lines of a very vivid cup of coffee having the middle and high notes accentuated while the deeper notes are muted. Some coffees simply dont work while others pull wonderful shots. I wonder what the comparison between a shot from the Leva (or other spring lever) and a skimmed shot from a traditional pump machine would be like.

I think that's a good point; there definitely is a family resemblence between the two styles of shot. But if one can get both the big and the "cupping" espresso from the same machine and coffee just by skimming the crema, that's a good thing to know.

I wonder if turning the pressure down to 6 bar would yield a similar cup in a pump machine. It would be to much work to do on a whim but would be interesting to try. Skimming the crema off is a lot cheaper than a Micro Casa leva.

I have found the thread very interesting but simply have been so busy I have not had time to try it out in a controlled environment myself. I need to set some time aside and try this out.

I think you guys are skirting an area I've often wondered about especially since this discussion began:

Should espresso be homogeneous? If brewing at lower pressures provides this is that in of itself "better"?

Clearly some blends pull best at higher pressures where the drink can be enjoyed in a series of levels as one descends the cup (each level or part being high quality and desirable). Is this a case of the coffee being tailored to the machine? Should we be tailoring our machines to the coffees? I would think the blends that throw bad crema are ones that need lower pressures?

In short is bad crema an indication that the shot pressure is too high (assuming other variables are on target)? Perhaps the era of standard 10-bar ristrettos and typical 9 bar shots should be reevaluated. Myself personally I don't like shots off my Anita anymore than 7-bar (per gauge blind basket). And less than 6.25-bar the crema is non-existent and the coffee under extracted (not to mention the shot volume tanking).

Perhaps pressure or what we consider "ideal pressure" isn't so ideal after all certainly not in every instance, that is if crema is any indication. Clearly the advice of "set to 9-bar and forget it" should be reconsidered. I did and my espresso is tenfold better (to me). In my experience Anita at 7-bar creates a shot of good volume (1.5 to 2-ounce) espresso where the crema is more incorporated into the shot rather than a stiff foam on top. Remarkable body and still full flavor balance and presence. I've done this with many blends and they all seem better, again to me. One thing I noticed is that the shot at lower pressures in a 58-mm basket takes longer, but the extra time doesn't create over extraction, it just takes longer to push the water through. My normales run at least 30-35 seconds, but they are sweet delicate normales, not a ristretto like shot some would suspect from a 35 second shot.

Lower your pressures and see what you find. Perhaps you won't have to skim or stir, I don't. And the espresso is even more amazing.

TimEggers wrote:Lower your pressures and see what you find.

Sounds interesting and simple enough to try

A quick note to anyone who does try a lower pressure, in a 58mm basket I noticed 2 things, first if using a bottomless a normale looks like a ristretto in appearance (tight cone, slow flow) but in taste nothing like a ristretto and second the shot will take longer as well. My normales now are closer to the 35-second mark. That seems to work really well for me judging by taste, YMMV.

Should espresso be homogeneous? If brewing at lower pressures provides this is that in of itself "better"?

Clearly some blends pull best at higher pressures where the drink can be enjoyed in a series of levels as one descends the cup (each level or part being high quality and desirable). Is this a case of the coffee being tailored to the machine? Should we be tailoring our machines to the coffees? I would think the blends that throw bad crema are ones that need lower pressures?

Isn't this what these conversations almost always come down to though? Different coffees respond differently to different things, some do better young, some with more rest, some at higher temps, some lower, some at finer grinds, some coarser, some at higher extraction volumes, some at less complete extraction, and now, some with more pressure, some with less?

Not to say that it isn't notable that most machines come at a fixed pressure and some coffees do better with less (or more) - it just doesn't seem all that surprising, and perhaps an advance in forthcoming machines will be pump pressure adjustable without opening up the box (not to speak of programmable pressure profiles that adjust during the shot).

EDIT/ADDITION #1: One thing that jumps out at me is that the skimming test might be a reasonable part of testing for the best extraction volume. Consider: the taste of the isolated crema on a ristretto extraction might be very different for a normale extraction, in a way that you might not get the full effect of if drunk with the rest of the espresso. So this could be a useful tool in figuring out quickly how much to pull out of a given bean. (The crema for the first half of the shot might taste great, and much less so for the second half, or vice versa, in a way that might be exaggerated as compared to the shot as a whole, helping to zero in on the preferred overall extraction.)

EDIT/ADDITION #2: Another question that comes to mind is whether it's really the crema that's the problem in coffees that are tasting better skimmed. I know it's sacrilege, but anyone tried letting them sit for half an hour after grinding to reduce crema and compared the flavor with the skimmed coffee? It could be that some unfavorable elements (to some people, for some extractions of some coffees) are in the crema, but could it also be that the way they are suspended in the crema for those shots is itself the unfavorable part, and that this could be reduced by simply letting the beans wait more than a minute from grind to pull? Or letting the beans rest a few more days before use and comparing that? Or {gasp} letting the shot sit for a minute, allowing the crema to dissipate before drinking? I'm certain I've had a shot that tasted much better after sitting for a moment, so it seems to beg the question...