Dialing In Espresso

Phenomenon

You finally get to use an espresso machine. The recipe online says: eighteen grams of coffee, a yield of thirty-six grams, twenty-five to thirty seconds.

The first shot runs out in fourteen seconds and tastes sharp and sour. You decide to fix it in one go: grind two steps finer, add one more gram of coffee, and tamp a bit harder.

The second shot takes almost fifty seconds, dripping out drop by drop, bitter and drying.

Which step went wrong? Too much grind change, too much coffee, or tamping too hard? You cannot tell — because all three changed together.

Variables

When you dial in an espresso, only a few things are really dials you turn; a few other numbers are results you read:

Distribution and tamping change the result too, but mainly by changing how evenly the water flows — Chapter 33 is devoted to that. For this chapter, treat them as "done the same way every time".

Experiment

A thought experiment: keep dose and grind the same, and only change when you stop the machine, collecting three shots at a yield of one and a half, two and two and a half times the dose.

Write down your prediction before reading on

How will the three rank in time, strength and taste? If you could turn only one dial, which do you think affects taste most?

Your prediction

Save prediction

Stored only in this browser on this device. Never uploaded.

A second thought experiment: keep dose and target yield the same, and grind in three steps from coarse to fine.

Write down your prediction before reading on

How will the time change across the three? Which way will the flavor move? If the third shot is both sour and bitter, how would you explain it?

Your prediction

Save prediction

Stored only in this browser on this device. Never uploaded.

A third, back to the opening: the first shot ran in fourteen seconds, sour and thin. This time you change only one thing.

Write down your prediction before reading on

Which one would you change? Afterwards, how do you expect the time and taste to move?

Your prediction

Save prediction

Stored only in this browser on this device. Never uploaded.

Observation

In the first experiment, the larger the yield, the longer it runs, the weaker the cup and the higher the extraction: a small yield is stronger and leans sour; a large yield is weaker and more likely to turn bitter and drying Confidence: [Rule of Thumb]. Yield often matters more than you would think: a systematic study varied grind, flow rate and water temperature, and found that their effects on what ends up in the cup were all smaller than the effect of how much liquid was collected Confidence: [Strong Evidence].

In the second experiment, the finer the grind, the more the puck resists and the longer it takes to reach the same yield, and extraction usually rises, moving the flavor toward less sour and more bitter Confidence: [Rule of Thumb]. But if the finest shot turns out both sour and bitter, it is probably not as simple as "over-extracted": past a certain fineness, water flows unevenly through the puck and overall extraction actually drops Confidence: [Strong Evidence] (Chapter 31). Grinding even finer at that point only makes things worse.

In the third experiment, as long as you change only one thing, the answer means something: one step finer, so many seconds longer, so much less sour — all of it belongs to that one step. The person at the start changed three things at once; nobody can say which of them the fifty seconds belongs to.

Model

Set the frame first, then let grind find the way

There is a common order for dialing in. The usual practice is to fix the dose first, then the target yield, and finally use grind to find the flow and time Confidence: [Common Practice]:

1. Dose: choose one to suit the basket, then leave it alone. Baskets are designed for a range of doses Confidence: [Common Practice]; too little or too much changes the space above the puck and its depth.
2. Yield: choose a ratio for the strength you want. A yield of about twice the dose is a common starting point Confidence: [Common Practice], but stronger or weaker is fine too — as Chapter 31 said, this is largely preference.
3. Grind: with dose and yield fixed, grind is your main dial. Too fast and sharply sour, grind a little finer; too slow and bitter-dry, grind a little coarser.
4. Taste: once you are close, use yield for small corrections: a few grams more for a weaker, slightly more extracted cup; a few grams less for a stronger one.

The advantage of this order is that each step moves only one dial, and everything fixed earlier is treated as a constant.

Change one variable at a time — and what "one" really means

Changing one variable at a time makes every result attributable. Change two things and you cannot tell which one did it; their effects might even cancel out, making you think neither mattered.

But in espresso, even turning one dial pulls on other numbers:

• One more gram of coffee: a deeper puck, more resistance, a longer time; with the yield unchanged, the ratio is also stronger.
• One step finer: more resistance, a longer time, and extraction moves with it.
• A few more grams of yield: a longer time, a weaker cup, more extraction.

So "change one thing" means: turn one dial, then predict what it will pull along with it. Time and strength moving as a result are readings, not a second thing you changed.

Navigate by taste

Translating taste into direction, this table is a starting point that works most of the time Confidence: [Rule of Thumb]:

Move only a small step in one direction each time. Espresso is very sensitive to grind; one step on the grinder often shifts the time by several seconds Confidence: [Rule of Thumb].

Time is a clue, not a target

"Twenty-five to thirty seconds" is a common starting point Confidence: [Common Practice] that tells you whether the puck's resistance is roughly in a sensible range. It is not the goal itself.

One example: a study found that using less coffee and grinding coarser can reach the same extraction yield as a conventional recipe even though the shot runs much faster, and the shots are more consistent from one to the next Confidence: [Strong Evidence]. Such a shot is weaker and tastes different, and whether you like it is another question Confidence: [Preference]. But it shows that a time outside the "standard window" does not make a shot wrong.

The real target is taste. Time is a signpost along the way: when it suddenly changes, the puck changed.

The beans change every day

With the same bag, a recipe dialed in today may drift in a few days. Cafés often dial in again every morning before opening Confidence: [Common Practice]: in the weeks after roasting, beans keep releasing carbon dioxide and slowly changing (Chapter 17), and room temperature, humidity and the grinder's own temperature change too. So dialing in is not a one-off job but a small loop you repeat: taste, pick one dial, change it a small step, taste again.

Transfer

This logic is not only for espresso machines. Pour-over works the same way: fix the dose and the water first (Chapter 22), use grind to find the drawdown time (Chapter 27), then fine-tune by taste. The five variables of Chapter 24 are the same map in every brewing method.

In the Atlas Espresso Machine Cutaway, you can drag just one of grind, dose or yield and watch how flow and time follow — practicing "one change at a time" in the simulation first.

At a café, you can use the same language with the barista: "this one is a bit sharply sour" is much easier to dial in than "this one doesn't taste good".

Misconceptions

"Twenty-five to thirty seconds is a law of espresso." It is a common starting point for checking the puck's resistance. A faster or slower shot is not wrong as long as it tastes the way you want; and a shot with a perfectly "standard" time can still be sour and astringent because the flow was uneven.

"The thicker the crema, the better the espresso." The brown foam on top of an espresso is called crema. It forms mainly from carbon dioxide left in the beans by roasting, which dissolves into the liquid under high pressure and comes out again as foam once it leaves the machine Confidence: [Strong Evidence]. So recently roasted beans, and blends containing canephora (robusta), often produce more crema Confidence: [Rule of Thumb]. How much crema there is says more about freshness and the beans than about whether the shot tastes good Confidence: [Rule of Thumb]: a shot with thick crema can still be bitter and astringent.

"Changing several things at once gets you there faster." Maybe, if you are lucky. But as soon as the result gets worse, you do not know which change to undo, and you end up starting over. Changing one thing at a time looks slow, but every step builds knowledge you can use.

"The finer the grind, the more it extracts." Only within a range. Past a certain fineness, flow becomes uneven, extraction actually drops, and the taste is harder to repeat.

Real Lab

With an espresso machine. Build a "yield ladder", changing only the yield:

1. Keep the dose, grind, distribution and tamp as usual, with a scale under the cup.
2. Pull three shots, stopping when the yield reaches one and a half, two and two and a half times the dose, and note each time.
3. Let the three cool to about the same temperature and taste them side by side: which is the most sour, which the most bitter, which do you like best?
4. Take the ratio of the one you liked best, fix it the next day, and then dial in with grind only.

Without an espresso machine. The logic of dialing in is the same in every method. Use pour-over or a French press for the grind experiment from Chapter 20, practicing "hold everything else, change only the grind":

If you have a moka pot, you can also change only the grind: fill the basket the same way without tamping, with the same water and the same heat, once at your usual grind and once a little finer. Note how long it takes from the first coffee appearing until it finishes, then taste side by side. A moka pot works at far lower pressure than an espresso machine (Chapter 30), but the chain "grind changes resistance, resistance changes time and taste" is the same. Be careful: grinding too fine for a moka pot raises the pressure in the pot, so do not go anywhere near espresso fineness.

Whatever you use, write it down: what you changed, how much the time moved, and which way the taste shifted.

Check Your Model

Predict. Your espresso is eighteen grams in, thirty-six grams out, running in twenty seconds, and tastes sharply sour. You grind one step finer and change nothing else. Before looking at any result, write down: how will the time change? Which way will the taste move? And if, after grinding finer, the time reaches thirty seconds but the shot is sour and astringent at once, what would you do next?

Diagnose. A friend says: "I switched to a new bag and kept my old recipe. It ran really fast and really sour. So I added two grams of coffee and went three steps finer — now it's too slow and too bitter." What went wrong? How would you suggest they start again?

Explain. In your own words, explain to a friend: why is espresso time a "reading" rather than a "target"? Why can a shot outside twenty-five to thirty seconds still be a good espresso?