The double hump

This post has come out of an email conversation between Andy Schecter, Scott Rao and myself. Scott had originally considered posting this as a comment on the Brewing Outside of Gold Cup post, but I thought it was too interesting to take the chance of being missed by people who might find it very interesting.

The timing seemed excellent after my last post on Cupping and French Press – for reasons that I hope will make sense once you read it.


Yes, there are three situations that come to mind where I have preferred coffee extractions that were “outside the box” (my usual box is 19-20%)

Situations in which different areas of the coffee bed were extracted to very different degrees (i.e. the strange ways most companies seem to make v60’s, some clover extractions (also usually absurd technique), and, in fact, espresso (it is impossible to get a uniform extraction from espresso.) I know you said you weren’t referring to these, but I would like to point them out. These extractions generally have tasted best below 19%.

There is something I call the “double hump”. These comments assume extraction (not water) temperature between 198-201F (92-94C) and the use of a commercial flat burr grinder with relatively sharp burrs. I would not generalize about extractions at different temperatures or made with different-quality grinders):

<15%: I have very little experience, or at least have had very few enjoyable experiences below 15% extraction. 15-16%: This is the “little hump”. I’ve often found these extractions to be cirtusy and juicy, with a more delicate acidity than that of higher extractions. To me they’re the coffee equivalent of Beaujolais Nouveau. They’re also more aromatic (perhaps simply due to the higher grounds: water ratio… aromatics extract almost instantly, and the aromatic profile of a coffee correlates far more with extraction temperature than it does with solubles extraction level.). Using a higher grounds:water ratio will inevitably result in more concentrated aroma. 16-18%: 
If a relatively uniform extraction, 16-18% is a bit of a dead zone: less fruit than either hump, and acidity tends to be flatter. The caramel content of the cup may be higher than at the little hump, but the overall cup is rarely great. Woody and harder flavors often come to the fore at this range. 18-19%: As you creep towards 19%, the presence of bigger, riper fruit and deeper caramel tastes increases. This range can be great or sometimes boring. 19-20%: The “big hump”: This range most often produces the ripest, most pleasing fruit (the type Robert Parket champions in wine) and the most caramels without the presence of appreciable astringency or bitterness. >20%: As the extraction increases beyond 20% the risk of bitterness and astringency taking over the cup increases exponentially. Some of the best cups of my life were between 20-20.5% but in a café setting I would usually err on the side of caution and aim for closer to 19.5% in order to avoid any creeping astringency.


I’ve had some recent experiences with brand new, very sharp, very large grinder burrs, as well as experiences with roller mills, that have convinced me there can be beautiful flavors in the 21.5-22.5% range without appreciable bitterness or astringency. I would speculate that such high-quality grinds, with minimal very-small and very-large particles, can extract to a higher degree more successfully than can the typical grind quality and particle size distribution most of us use in our cafes everyday. Please note: I do not think that using a Ro-Tap or a fines-separator with a “normal grinder” to produce a narrower particle size distribution can reliably replicate the grind quality of a truly superior roller mill or a machine with very large, flat burrs.

Part of me wants to leave it here and just let people discuss stuff in the comments but I can’t resist a little commentary:

1 – Uneven extractions: This is interesting, and I think a lot of people will jump on this, but it isn’t the most interesting thing for me here. I’d like to avoid the “ExtractMojo can’t gauge the evenness of extraction, and is therefore redundant” argument here please.

2 – The small hump: This is really interesting to me. It is something similar to what I have experienced with brew temp in espresso. If you go a little low it easily turns sour. Drop lower again and its sweetens up but gets quite boring. I’d really like to understand what is causing the dip in quality between 17-18% range. Thoughts and comments on this would be especially welcome.

3 – Grinders: First of all – on this point I am quite jealous, I’d love to play with roller mills and the like. It also brings up my frustrations with the lack of grinder research in the last 50 years.

Thoughts and comments on the above welcome!

15 Comments The double hump

  1. Pingback: Tweets that mention The double hump – jimseven --

  2. Matt Flipago

    This idea of brewing to get numerous points of balance isn’t that strange, at least to me. Tea is often brewed at different balance points, depending on skill and dosage of tea. Western tea drinkers are you to a heavily extracted one time brew, while many people who are familiar with the use of gaiwans and yixing teapots is drinking a much more heavily dosed but shorter brew. Even within that, there are different balance points you can stop it at, depending on how it smells, with that general curve of a lighter, more aromatic flavors coming out at first and later points being much darker, with more presence. As the tea gets used more and more it fades into sweet nothingness, however you can still over brew these, and get a terrible astringency.
    Simply by having different solubility curves for different particles, some plateauing earlier than other will lead to these types of “humps”.
    I am always perplexed by the assumption of evenness in extraction, because I don’t see how it’s possible. Cupping or french press aren’t an even extraction because the top of the floating grinds isn’t brewing the like the bottom of the floating grinds. So much air is caught between the grinds and so much super saturation in the bed of floating grinds, that I would think a much “more even” extraction is the constant stirring, getting the grinds more equidistant, and as larger grinds separate from smaller ones in a vortex could we get a more even extraction there?
    That coupled with the idea there are different balance points or humps during brewing, makes me wonder, could a certain unevenness during the brew be good. Could the finer grinds tend to one balance point, while getting coarser grinds to favor another? That said I don’t think the “humps” during brewing are caused by uneven grinds.

  3. Evan Gentry

    Could the solubility of different compounds, as opposed to different particles, within the coffee cause different flavors/aromatics to come out or be masked at different extraction levels? I would imagine that different compounds’ varying solubilities, saturation points, etc. would come into play here.

  4. Jay C.

    While I find the notation of 19-20% (conveniently) equivalent to Robert Parker’s highest scores to be questionable at best, the correlation of the extraction percentages to wine types is an interesting one.

    Without a doubt, Parker has his preferences. His preferences have caused some producers to craft wines in an attempt to garner his high scores. Other producers have relatively ignored Parker’s preferences and continue to create great wine as always.

    The correlation to wine is interesting because it (seemingly) unintentionally underscores that coffee (like wine) can be appreciated across a range of flavors. That to say that 19-20% is equivalent to Parker’s best is also to acknowledge that coffees in these other ranges are just as valid and valued as those wines not meeting Robert Parker’s top preferences.

    If the 2006 Alban Vineyards Syrah Lorraine Thorevilos rates a perfect 100 by Robert Parker, is it the ultimate wine? Or is that to say that the 2007 Coudoulet de Beaucastel Cotes du Rhone at 92 points is not worthy of enjoyment?

    Let’s apply that to coffee and enjoy the drinking.

  5. John Piquet

    While the understanding is great to have, I see many using it as a crutch. The idea that coffee and espresso should taste great is the right objective, but beware the path you take to reach it. As Scott pointed out, his comments are specific to sharp flat-burr grinders. Grinders, specific roasting parameters, age of beans, type of espresso machine, type of filtration used for press, vac pot, or chemex, etc. All of these things will change the various nuances of the cup. It’s a balancing act that one has to adjust for on the fly. Numbers will often guide you to a great cup, but they can also leave many a wonderful cup behind. For those who like to use the Mojo, it’s a great tool to get you in the right ballpark… but tasting will get you there just as fast.

    And this conversation reminded me of a great post by Chris Tacy following his experience after judging at the Good Food Awards.

    “…It was incredible to see a single coffee get scored 90+ points by one judge and 80 or below by another — and to have each judge present a compelling, rational and well-reasoned argument for why their scoring was correct.”

  6. James Hoffmann

    This is where I disagree. I think Parker’s effect on wine, redacted to coffee would see producers trying to make their coffees taste like the Kenyas he loves so much. i.e. repressing terroir and variety in favour of producing a specific taste.

    Parker may consider a Brazil invalid, or poor, because of its lack of sparkling fruit – but the correct thing to do would be to enjoy its genuine expression of its terroir.

    I should add that it kills me a little inside to talk about producing countries in general like this, but I am trying to make a point…

  7. James Hoffmann

    I disagree on a few points:

    Tasting will get you into the ballpark, but it won’t do it as quickly. Of this I am certain. I am not saying rule taste out, but when you are trying to move a certain distance (in brew parameters or extraction) it helps a great deal to understand the exact impact of changing a variable in a way that lets you make an even more educated guess.

    As for the range of scores for a coffee from 90+ to below 80: This is what I haven’t seen when it comes to extraction. I haven’t seen people prefer the sub 16% updosed extraction when comparing it with a evenly extracted 19-20%. I asked, in a previous post, for people to give examples of this but the only responses were examples of when underextraction “didn’t taste awful” – which is perhaps explained by Scott’s post above. I am absolutely open to being proven wrong here, but so far the data has been overwhelming. I will change my mind should the evidence be compelling, and I will do so publicly because it would be incredibly interesting to do see such evidence.

    I’m not saying “As long as it hits 19% on my software and refractometer then it is exceptional and perfect coffee.” No one is saying this, but I am saying that when coffees are relatively evenly extracted, with good equipment and water, then the majority of people would enjoy brews where the extraction reached is approximately 20% of the ground coffee by weight.

  8. Benjamin Schellack

    Infrequent use of the Mojo/refractometers, would account (in part) for:

    “I asked, in a previous post, for people to give examples of this but the only responses were examples of when underextraction “didn’t taste awful”.”

    That said, using scales (and a 60-70% extraction ratio) for espresso has worked well in my case! Good coffee = science+art.

  9. Jay C.

    “This is where I disagree. I think Parker’s effect on wine, redacted to coffee would see producers trying to make their coffees taste like the Kenyas he loves so much. i.e. repressing terroir and variety in favour of producing a specific taste. “

    Hmmm, perhaps I might not have been clear in my writing. I think you and I are of similar thought on this one. Part of my point is that it is an error to rely solely on the score of Parker – that is, unless you actually share his tastes in wine.

    By extension, it is an error to rely solely on the percentages being pumped out of an iPhone.

    “I haven’t seen people prefer the sub 16% updosed extraction when comparing it with a evenly extracted 19-20%.”

    I don’t think I’m understanding something here. Is there a difference between an evenly extracted 16% coffee and an updated 16% extraction coffee? I mean, isn’t a 16% coffee a 16% coffee – regardless of how you arrived at the number?

    I am absolutely open to being proven wrong here, but so far the data has been overwhelming.

    Forgive me if I end up being the stalwart again, but what overwhelming data has been presented??? I think I read rather frequently but I don’t recall an overwhelming amount of data presented. Ever.

    Must of what I’ve read has been anecdotal. People essentially agreeing with each other. People seeking to hit that “magic number.” That without the magic number (or at least that piece of equipment to measure that magic number) one must be off the mark.

    “Tasting will get you into the ballpark,…”

    This seems to imply that while tasting coffee will get you somewhere near ideal, it won’t arrive you there. With the implication that one must have this whiz-bang piece of coffee wizardly equipment otherwise one will be incapable and unable to hit the mark.

    While I don’t think it was your intention, this implies that taste alone is not enough to tell whether a coffee is “good”, “great” or simply just “bad.”

    To bring it back to the odd correlation to Parker – it would be as though Parker crafted his scores without tasting the wines themselves and merely fed them into a chromatograph (or some such other device).

    Since Parker lives not too far from me and frequents some of the same restaurants, I’ll ask him about that the next time I bump into him!

  10. sanford bledsoe iii

    The only people saying that taste and MoJo are opposed, are the people that are opposed to using MoJo. I love our refractometer because we can use it simultaneously to our taste buds, while we taste, and it helps us better understand why we like what we like. Everytime we do an experiment with brew methods, we use the device simultaneously, and we pick what cups we like best based on taste, THEN we look at the data. Everytime, we’ve found that the cups we like best are the ones in that 19-20% range. All the experiments that James has posted here on his blog reflect this exact same idea. Blind tastings, THEN reveal the TDS and EXT%.

    Rereading this, it seems a little abrasive, which isn’t my intention, John. I just don’t appreciate the dichotomous idea that tasting and measuring are opposed; that we can only take one path to arrive at our goal. What we are actually doing is gathering as much data as possible, from as many different instruments as we can. By being scientific in our approach, we can take every path!

  11. Lorenzo

    I know this post is a little dated, but during a training session today, the new barista asked a question that sparked an idea while looking at the SCAA flavor wheel. I’m not very knowledgeable in the field of coffee science, but from what I understand, the aroma and taste aspect of the wheel was created by grouping the volitile compounds that contribute to those flavor perceptions and then ordered by molecular weight. One then postulates that the lighter the molecular weight of the compound, the quicker it is to show up in the cup. Am I right so far?

    So taking this idea, along with the double hump theory, I’ve created a VERY rough jpg grouping these compounds by extraction percentage – kind of a rough idea that still needs to be ironed out, and I don’t have the gift of simplifying ideas into easily understandable language in writing, but I hope the image explains itself.

    I do think it is important to note that this would be a simple generalization of extraction, varying by coffee and roast and method. Also, am I correct in assuming that everything extracted in a 16% brew would still be present in a %20 brew, only in a different balance? As such, thinking about a blend of coffees, the flavors work in a synergistic manner to create new perceptions of flavor and aroma. In a similar way, different extracted compounds “blend” within a coffee, varying by extraction percentage, creating synergistic flavors at each level.

    I do think it’s crazy that this “double hump” was unintentionally, maybe intentionally as Ted Lingle is a sharp fellow, built into the flavor wheel.

  12. Pingback: The Test Roast Solubility Quality Check | Prufrock Coffee

  13. Pingback: Joe Olujic

  14. Pingback: French Press Recipes – Not Your Parents' Coffee

  15. Pingback: European Extraction Preferences in Brewed Coffee – The Coffee Hot Spot

Leave A Comment