Updated October 2026
Computational Coffee means choosing every brewing setting (ratio, dose, temperature, time and grind) from the measured density of your roasted coffee, instead of using one fixed recipe for every bag. Roasted coffee spans a density range of 0.330 to 0.450 g/ml, and a fixed recipe only suits a small slice of it.
Over-extracted coffee is bitter and under-extracted coffee is sour. Density tells you which way a coffee will lean before you start, so you begin much closer to a sweet cup. This page explains the idea, shows the data behind it, and walks through every tool you can adjust for espresso and filter coffee.
Measure the density of your roasted coffee in a 100 ml cylinder, enter it in the coffee density calculator, and brew the recipe it gives you. If the cup is bitter, extract less by subtracting 0.02 g/ml. If it is sour, extract more by adding 0.02 g/ml.
Jump straight to the part you need.
The idea
This could be the most important picture in the history of coffee 😉, because it shows the range of roasted coffee seen across the planet, the frequency of each roast level encountered, and consequently what you need to do to deal with it.
Figure 1: the actual range of roasted coffee density and how often each roast level turns up.
1
Over-extracted coffee is bitter, and under-extracted coffee is sour. So no one boils their coffee endlessly in a percolator anymore, because it always makes very bitter coffee.
2
How easy or difficult it is to extract the soluble compounds depends on the roast level, and can be measured by the density.
Though beans are always seeds, seeds are not always beans. A bean is just one kind of seed. Specifically, it is a name for seeds of the family Fabaceae (also known as Leguminosae), of which the coffee plant is not a member. So coffee “beans” are not actually beans, and I will often call them seeds.
After a couple of years of running, recipeseries.coffee is approaching 10,000 visitors from all over the globe. Statistically analysing the log of density entries reveals some amazing facts.
Don’t know how to measure coffee density yet? See How to Measure Roasted Coffee Density.
Density range
The first thing the graph shows us is that about 90% of coffee densities fall between 0.330 and 0.450 g/ml. That is 120 steps of 0.1 g per 100 ml, from the darkest roasts to the lightest roasts, and it reveals the density range.
Numbers this small are hard to picture, so multiply by 100: 100 ml of roasted coffee seeds in a volumetric cylinder weighs between 33 g and 45 g. Now look at the two ends of the range and how each affects extraction.
Darkest roast · 0.330 g/ml
At the very darkly roasted end, the seeds don’t weigh much at all, because they have boiled off most of their water. They are so porous and brittle you don’t need a grinder, you can crush them with your fingertips. The slightest bit of lukewarm water and the solubles come flooding out. You have to dial all your tools right back, or you are bound to over-extract and make bitter coffee.
Lightest roast · 0.450 g/ml
At the very lightly roasted end, the seeds weigh a lot more. They are rock hard, and grinding them in a hand grinder is nigh on impossible. You need heaps of very hot water, and every tool and trick used to the max to extract the good sugars. It is quite likely you will under-extract and make very sour coffee.
Fixed recipe
Most baristas use a fixed recipe to extract coffee. The SCA’s research suggests the average barista uses a 1:2 brew ratio for espresso and weighs the output: an 18–20 g dose, a 36.5 g output, extracted in 25–30 seconds at 9 bar and 200°F (93°C), using pre-infusion, through an 18 g basket (source). I will call this the Fixed recipe from here on.
My mission here is to show you why that Fixed recipe can’t work, quite often.
I’m going to use a golfing analogy (quite a lot). The density range is like a 120-yard golf fairway. In golf you have a bag of clubs of different sizes and capabilities, from big powerful drivers right down to your little putter. The first step is to examine the fairway and see how far away the hole is. It might be right in front of you, at the far end 120 yards away, or at any of the 120 yards in between. Then you select the club that gets your ball on the green, so you can finally use your little putter to chip it into the hole.
Density frequency
The second thing the graph in Figure 1 shows is the density frequency: how often each roast level (or density) occurs. You can view it as the probability that your coffee is roasted where you expect it to be, or not, as the case may be.
I fully expected the frequency distribution of coffee densities to be bimodal (a double-peaked normal distribution), looking something like this:
Figure 2: the expected coffee density frequency.
Bags of coffee are often labelled “Espresso Roast” or “Filter Roast”. That led me to believe:
•
there would be more espresso roast and slightly less filter roast;
•
most espresso would be a mid-roast, probably peaking around 0.390 g/ml. Roasters know most people use the Fixed recipe (1:2 ratio, 18 g dose, 36 g yield, 93°C, 25 second shot), so everything would kind of work most of the time;
•
filter roast would peak a little higher, at about 0.410 g/ml, so the Fixed filter recipe (1:16 ratio, 14 g dose, 224 g yield, 93°C, 4:00 minute V60) would kind of work most of the time.
That is not what the graph in Figure 1 reveals. It is not a bimodal distribution, and it is much flatter than expected. You are going to get coffees roasted right across the range, and well outside the roast level you pre-determined to expect. If you think you are getting a mid-roast that will play nice with your fixed 1:2 routine, the reality is 65% of the time you will be way off. That is more than half. The odds aren’t good, you are on a losing hand.
If you make espresso with the Fixed recipe (1:2 ratio, 18 g dose, 36 g yield, 93°C, 25 second shot), you will be in the rough at 0.390 g/ml 65% of the time. You will be miles off: super bitter or super sour. Worse, by fixing all your available tools you limit yourself to your little putter (grind adjustments) to move toward the hole, which is an impossible thing to do.
Do you think Tiger Woods plays golf like this? That he never looks at the fairway, puts a blindfold on, and smacks the ball halfway down the range, every time, no matter what, and then only uses his putter to finish the job? No.
Using density
You’ve got a great set of tools, but you are not using them, because most of them are rusted solid into pre-fixed positions. I’m going to try to convince you to change that, and to use all your tools to get on the green, in the ballpark, within putting distance of the hole.
| Setting | Darkest roast, 0.330 g/ml | Lightest roast, 0.450 g/ml |
|---|---|---|
| Ratio | 1:1 | 1:4 |
| Dose | 25 g | 14 g |
| Basket | 25 g basket | 15 g basket |
| Yield | 25 g | 56 g |
| Pressure | 6 bar | 9 bar |
| Temperature | 85°C | 95°C |
| Shot time | 20 seconds | 40 seconds |
| Milk | Lots of milk | Hardly any milk |
| Cup size | 240 ml | 150 ml |
| Bypass water | None | Replace some of the milk |
| Grind | About 400 microns | About 180 microns |
Approximate bounds only. Everything in between is calculated by recipeseries.coffee, relative to density and the degree of extraction.
| Setting | Darkest roast | Lightest roast, 0.450 g/ml |
|---|---|---|
| Ratio | 1:13 | 1:20 |
| Dose | 16 g (at 0.350 g/ml) | 10 g |
| Temperature | 90°C | 95°C |
| Brew time | 2:30 | 7:00 |
| Grind | About 1200 microns (at 0.350 g/ml) | About 700 microns |
Dose and grind bounds were measured at 0.350 g/ml for the dark end.
Espresso tools
Each tool below has two bounds: where to set it for the darkest roast (0.330 g/ml) and for the lightest roast (0.450 g/ml). Everything in between is calculated by recipeseries.coffee, relative to density and the degree of extraction you need.
In golfing terms, ratio is your no. 1 wood, your big driver. Because of water’s polarity and its ability to form hydrogen bonds, water makes an excellent solvent, meaning it can dissolve many different kinds of molecules. The more solvent you use, the more you extract.
It is also the most misused tool. Some people randomly choose to make a ristretto or a lungo with no relation to what they need to achieve. A light roast pulled as a ristretto is going to be very sour. A dark roast pulled as a lungo is going to be very bitter.
Darkest roast · 0.330 g/ml
You need hardly any water at 0.330 g/ml: just a 1:1 ratio, or you will over-extract.
Lightest roast · 0.450 g/ml
At 0.450 g/ml you need masses of solvent water: a 1:4 ratio to have any hope of avoiding under-extracted sour.
Dose is sadly misused too. Changing the dose helps you achieve the required ratio. It is not a random strength setting.
Darkest roast · 0.330 g/ml
At 0.330 g/ml with the recommended 1:1 ratio you have a tiny amount of water, and so a tiny yield. Maximising the dose to a huge 25 g gets you more of a much better-tasting yield.
Lightest roast · 0.450 g/ml
At 0.450 g/ml with the recommended 1:4 ratio you will struggle to reach the ratio unless you minimise the dose. At 14 g you can make a sweet 56 g yield, and that is about the max.
Your basket should be within ±1 g of your dose. VST precision baskets come in 15 g, 18 g, 20 g, 22 g and 25 g, and you need a full set. The number, diameter and distribution of the holes differ for each size, to maintain a consistent flow rate at different doses.
Now that you are using ratios and doses correctly, you need the full set of baskets. Compared with the cost of your machine they are a cheap, good investment. Get the ridgeless versions. For more detail see How to Choose an Espresso Basket.
Darkest roast · 0.330 g/ml
At 0.330 g/ml you use the 25 g basket with your 25 g dose.
Lightest roast · 0.450 g/ml
At 0.450 g/ml you put your 14 g dose in the 15 g basket. In a bigger basket the holes are larger, so it won’t generate enough pressure or get enough contact time.
Mr Hoffmann et al might hate ratios, but they are useful. Your yield is entirely set by your ratio and dose.
Darkest roast · 0.330 g/ml
At 0.330 g/ml, with the recommended 1:1 ratio and a 25 g dose, you stop at a yield of 25 g.
Lightest roast · 0.450 g/ml
At 0.450 g/ml, with the recommended 1:4 ratio and a 14 g dose, you stop at a yield of 56 g.
Espresso tools
Pressure is used to speed up extraction. It is the whole point of espresso taking 30 seconds versus 4 minutes for a French press.
Darkest roast · 0.330 g/ml
At 0.330 g/ml the seeds are soft and porous, so ease off the pressure to 6 bar.
Lightest roast · 0.450 g/ml
At 0.450 g/ml the seeds are hard, so use more force to extract more: 9 bar.
Higher temperatures tend to flatter lighter-roasted or more acidic coffees by bringing out the sweetness, while lower temperatures can reduce the roasted, ashy flavours that might otherwise be found in darker roasts (source).
Darkest roast · 0.330 g/ml
At 0.330 g/ml, lower the temperature to 85°C.
Lightest roast · 0.450 g/ml
At 0.450 g/ml, increase the temperature to 95°C.
Contact time is the other half of extraction: the longer the water is in contact with the coffee, the more you extract.
Darkest roast · 0.330 g/ml
At 0.330 g/ml, reducing the shot time to as low as 20 seconds helps avoid over-extraction.
Lightest roast · 0.450 g/ml
At 0.450 g/ml, a longer shot extracts more, so for a very light roast going up to 40 seconds helps.
You don’t have to use milk, it is your choice, but milk and coffee are a marriage made in heaven. When milk is steamed to 60°C it reaches its optimum sweetness. Aerate it just enough to produce a thick paint texture.
Darkest roast · 0.330 g/ml
At 0.330 g/ml, lots of milk helps sweeten a very dark roast.
Lightest roast · 0.450 g/ml
At 0.450 g/ml the coffee is at a high ratio and becomes lighter and tea-like. Reducing the volume of milk helps prevent washing it out.
Espresso tools
Cup size is another secret weapon in your toolkit, much like a wedge to help you out of a hole.
150 ml
170 ml
190 ml
220 ml
240 ml
Darkest roast · 0.330 g/ml
At 0.330 g/ml, a lot of sweet milk helps make that dark roast, and a big cup helps: 240 ml.
Lightest roast · 0.450 g/ml
At 0.450 g/ml the coffee has less dose and a high ratio. Use hardly any milk, in fact replace some of the milk with bypass hot water. A small 150 ml cup helps strength.
At 0.450 g/ml, a very lightly roasted coffee needs a low dose and a high ratio of water (as above). If you add a lot of milk, as most people do in flat whites, lattes and cappuccinos, the light delicate flavour notes, fruits, acidity and aromas of a lightly roasted coffee are easily overwhelmed. By replacing some of the milk with hot water you bring out the full range of flavours from your coffee seeds.
This hot water has bypassed the extraction bed (to avoid over-extraction bitterness), hence the term bypass. Bypass is commonly used in Americanos and long blacks, but with the trend to lighter roasts it should start being used in milk drinks too. It is a secret tool to get you out of a sand trap, like a wedge.
Dark roast · 0.350 g/ml
At 0.350 g/ml, the other end of the range, a very darkly roasted seed can be ashy and burnt-tasting. A large volume of milk makes a more palatable drink, and no bypass water is required.
Light roast · 0.450 g/ml
At 0.450 g/ml, replace a proportion of the milk with bypass hot water to keep the delicate flavours of a light roast.
Grind adjustments should be step-less and measured in microns. You should not expect to make big adjustments with your grinder, because you are constrained by physics. Grind too fine and your basket holes plug up and the shot chokes. Grind too coarse and there is not enough puck resistance, so water runs through too fast.
If you change all the other settings appropriately relative to density, you will need to change your grinder much less wildly, and extractions will work much better. You can measure your grind setting with a Brewler. Measurements are good.
Darkest roast · 0.330 g/ml
At 0.330 g/ml you will be grinding coarser to reduce extraction: expect about 400 microns.
Lightest roast · 0.450 g/ml
At 0.450 g/ml you will be grinding finer to increase extraction: 180 microns is ridiculously fine.
recipeseries.coffee does not give you the answer, it gives you a better starting point. You still need to taste the coffee and judge whether you have avoided sour under-extraction and bitter over-extraction. The coffee may not have inherent sweetness, but avoiding these extremes is achievable. Taste the coffee before you add any milk, because milk masks flaws. Try dipping a teaspoon in, just coat it (you don’t need to fill it), then suck it. Practise by making intentionally massively under-extracted and over-extracted coffee to learn the tastes. Note that it is really hard to taste hot coffee, and the taste alters as it cools.
Filter tools
The same principles apply to all coffee extraction methods; only the toolset and the bounds differ. Here are the filter tools for V60, Hario Switch and similar brewers.
Again, ratio is your most powerful tool, your no. 1 wood, so use it carefully. Water is an excellent solvent, and the more solvent you use, the more you extract.
Darkest roast · 0.330 g/ml
At 0.330 g/ml use less water: just a 1:13 ratio, or you will over-extract.
Lightest roast · 0.450 g/ml
At 0.450 g/ml you need masses of solvent water: a 1:20 ratio to avoid under-extracted sour.
Dose helps you reach the ratio you need for the density of your coffee.
Dark roast · 0.350 g/ml
At 0.350 g/ml with the recommended 1:13 ratio you have a smaller amount of water, so increasing the dose to 16 g helps.
Lightest roast · 0.450 g/ml
At 0.450 g/ml with the recommended 1:20 ratio, reducing the dose to 10 g helps.
Higher temperatures tend to flatter lighter-roasted or more acidic coffees by bringing out the sweetness, while lower temperatures can reduce the roasted, ashy flavours found in darker roasts (source).
Filter coffee has a much longer contact time than espresso, and the temperature drops over that time, so the range is smaller: 95°C for lightly roasted, very dense seeds, down to 90°C for darkly roasted, porous seeds.
Darkest roast · 0.330 g/ml
At 0.330 g/ml, lower the temperature to 90°C.
Lightest roast · 0.450 g/ml
At 0.450 g/ml, increase the temperature to 95°C.
With filter coffee you are making the same fixed cup size. The only change is that different dose sizes retain slightly different amounts of water.
Lighter roasts need more contact time to extract, darker roasts need less.
Darkest roast · 0.330 g/ml
At 0.330 g/ml, lower the time to 2:30.
Lightest roast · 0.450 g/ml
At 0.450 g/ml, increase the time to 7:00.
Grind adjustments should be step-less and measured in microns. They are your little putter. You should not expect to make big adjustments with your grinder, because you are constrained by physics. Grind too fine and your filter plugs up and the brew stalls. Grind too coarse and there is not enough filter resistance, so water runs through too fast.
If you change all the other settings appropriately relative to density, you will need to change your grinder much less wildly, and extractions will work much better. You can measure your grind setting with a Brewler.
Dark roast · 0.350 g/ml
At 0.350 g/ml you will be grinding coarser to reduce extraction: expect about 1200 microns.
Lightest roast · 0.450 g/ml
At 0.450 g/ml you will be grinding finer to increase extraction: 700 microns is ridiculously fine.
The key rule
Computational Coffee refers to the use of data analysis, mathematical modelling and computational simulations to understand coffee extraction systems and relationships. An intersection of computer science, biology and big data, the field also has foundations in applied mathematics and chemistry.
The website recipeseries.coffee is Computational Coffee, and it does all the calculations for you. It takes your measured density reading and calculates the relative setting for each tool (ratio, dose, temperature, grind, pressure, yield, time, cup size, bypass water and milk) for each extraction method (espresso, Hario Switch or Clever Dripper, Aeropress, filter and more). That gives you a better starting position for coffee extraction, to avoid bitter or sour coffee.
Because recipeseries.coffee changes all the settings in concert relative to the density, you get one tool to rule them all: density. Shifting the density shifts every setting together, which makes a very appreciable step in the right direction.
Coffee too bitter?
Extract less
Subtract 0.02 g/ml from the density, e.g. change 0.375 to 0.355 g/ml.
Coffee too sour?
Extract more
Add 0.02 g/ml to the density, e.g. change 0.375 to 0.395 g/ml.
The most powerful tool of all is ±0.02 g/ml. Sure, if you are in the ballpark and only need a minor adjustment, a grinder nudge may do the job.
Really small seeds (yes, we are looking at you, Ethiopian landraces) are difficult to roast and tend to overcook quickly. With very small seeds you will sometimes need to extract less by subtracting 0.02 g/ml.
Very large seeds, such as Pacamara and Maragogipe, are also difficult to roast. They tend not to cook to the same level all the way through. With very large seeds you will sometimes need to extract more by adding 0.02 g/ml.
Measure first
You have heard the expression “measure twice, cut once”. You don’t see builders cutting wood first, then trying to nail it up and going “oh, that didn’t work”, and blindly trying again.
No, they use a simple tool to take a measurement first. You don’t have to make coffee like this:
You have simple, cheap tools to measure first, which gives you a better starting point from which you can avoid bitter and sour coffee much more successfully. All you need is:
Worked example
Let’s do a couple of real-world examples to compare the recipeseries.coffee recipe with a Fixed recipe. The first coffee is Sumava Ruva Costa Rica Black Honey Gesha roasted by Ozone, with a measured density of 0.421 g/ml. This coffee is roasted slightly on the light side: not massively beyond mid-roast, but certainly in the upper 40% of roast levels you are likely to encounter. Already we can see why the Fixed recipe runs into trouble. It is going to under-extract and undershoot the hole, ending up sour.
Temp
recipeseries.coffee: 90°C
Fixed: 93.5°C
The Fixed recipe is too hot and will extract some off notes.
Ratio
recipeseries.coffee: 1:3.1
Fixed: 1:2
This will cause a real problem. Fixed is not going to extract enough, which will result in a sour coffee.
Dose
recipeseries.coffee: 19 g
Fixed: 18 g
Not a hugely significant difference, but using less dose is not going to help if you are only using a 1:2 ratio.
Yield
recipeseries.coffee: 59 g
Fixed: 36 g
This is the real problem: insufficient water is going to under-extract, making it sour.
Time
recipeseries.coffee: 31 s
Fixed: 25 s
With less contact time, Fixed is also going to under-extract, increasing sourness.
Grind
recipeseries.coffee: 283 microns
Fixed: 150 microns
With a Fixed recipe the only tool you use is grind size. In an attempt to extract more you have to grind incredibly fine, which will choke your basket holes and ultimately fail.
Result
recipeseries.coffee: Beautifully sweet cup
Fixed: Sour
Many would argue 0.421 g/ml is too light a roast for espresso, and that it would work better as a filter coffee. The only reason for this is not using all your espresso tools and limiting yourself to grind adjustments. Espresso is perfectly capable as an extraction method for the complete density range.
Worked example
The second coffee is Los Angeles El Salvador Natural Pacamara roasted by Cypher Roastery, with a measured density of 0.355 g/ml. Not the darkest of roasts, but well into the dark side. This was an expensive COE #5 coffee, so you want to make the best of it. The Fixed recipe is going to over-extract this, overshooting the hole and ending up bitter.
Temp
recipeseries.coffee: 85°C
Fixed: 93.5°C
The Fixed recipe is way too hot and will extract bitterness.
Ratio
recipeseries.coffee: 1:1.3
Fixed: 1:2
This will cause a problem. Fixed uses more solvent water, which will over-extract, resulting in a bitter coffee.
Dose
recipeseries.coffee: 24 g
Fixed: 18 g
With a 1:1.3 ratio you need to increase the dose. The fixed 18 g dose is just too small.
Yield
recipeseries.coffee: 31 g
Fixed: 36 g
This might not look significant, but a 31 g yield from a 24 g dose is very different from a 36 g yield from an 18 g dose. The Fixed recipe is extracting more from a lower dose, increasing bitterness.
Time
recipeseries.coffee: 22 s
Fixed: 25 s
Fixed has more contact time, increasing bitterness.
Grind
recipeseries.coffee: 390 microns
Fixed: 500 microns
With a Fixed recipe the only tool you change is grind size. In an attempt to extract less you have to grind very coarse, which provides insufficient resistance in your basket (particularly with lower doses) and ultimately fails, with water just pouring through.
Result
recipeseries.coffee: Beautifully sweet cup
Fixed: Bitter
By measuring the density before you start, and using recipeseries.coffee, you get a much better starting position from which any adjustment is minor and achievable. If you keep using the Fixed recipe you are doomed to very poor results quite often.
Trying to use a Fixed recipe for all coffee densities is very limiting, and destined for failure a lot of the time.
Analogy
Imagine if you tried to cook meat the same way you are making coffee, using one fixed recipe for all meat.
If you had a beautiful tender eye fillet steak, a suitable recipe might be 2 minutes on each side at high temperature. But if you have a tough old piece of brisket, is that going to be a good result? No! You need a long, slow 4 hours at low heat to achieve a tender, pull-apart result. When cooking meat you accept that you have to use an appropriate recipe for the piece of meat you have.
It is high time people woke up to the same principle with coffee: you need to measure the density of the coffee to determine an appropriate extraction recipe.
Common shortcuts
World Aeropress Championship: Winning Recipes of the Last 5 Years! – JayArr Coffee
This is the weirdest thing: any of these recipes could be hideously sour or terrifically bitter, depending on the density you faced. Recipes on their own, without knowing the density, are useless.
I’m afraid that doesn’t really work. The label can be misleading. Even when bags are simply labelled Filter or Espresso Roast, the coffee might still not be what you expect. Many don’t say anything, which is very sensible really.
Can’t you just tell the roast level by looking at the colour of the bean? I’m afraid not: colour is not a good indicator. If you cook a roast at a really high temperature, it might be black on the outside but still tough on the inside. Coffee cooked high and fast will be dark but still hold a lot of water, be dense and be difficult to extract.
A piece of meat cooked long and low is not going to be black on the outside, but it is going to be pull-apart tender. Coffee roasted long and low won’t be dark, but it will have lost most of its moisture, be brittle and porous, and extract really easily. Colour is not the best tool to use. Measuring density is much better.
Conclusion
It blows my mind how everyone tries to extract coffee without first measuring how hard it is going to be, and then doesn’t use the full range of their tools. Measure the density, set all your tools relative to it, then taste and adjust by ±0.02 g/ml.
PS: don’t try to use the chart under the graph to get your settings. It is a crude approximation to illustrate the concept. Use the website recipeseries.coffee, which is much more accurate.
FAQ
Computational Coffee uses data analysis, mathematical modelling and simulation to understand coffee extraction. recipeseries.coffee applies it by taking the measured density of your roasted coffee and calculating the ratio, dose, temperature, time, grind and other settings for each brew method.
About 90% of roasted coffee falls between 0.330 and 0.450 g/ml, from the darkest roasts to the lightest. That is 33 g to 45 g per 100 ml of roasted seeds.
A recipe only suits the density it was developed for. Over-extracted coffee is bitter and under-extracted coffee is sour. A light, dense coffee needs more extraction than a fixed recipe gives it, and a dark, porous coffee needs less.
Fill a 100 ml volumetric cylinder with whole roasted coffee, weigh it on 0.1 g scales, and divide the weight by 100 to get g/ml. For example, 39 g in 100 ml is 0.390 g/ml. The full method is in How to Measure Roasted Coffee Density.
If it is bitter, extract less by subtracting 0.02 g/ml. If it is sour, extract more by adding 0.02 g/ml. This shifts every setting in concert, which is far more effective than changing only the grind.
Not reliably. Labels such as Espresso Roast or Filter Roast can be misleading, and colour is a poor indicator, because a coffee roasted hot and fast can be dark outside and still dense inside. Measuring density is more reliable.
Yes. Espresso can extract the whole density range if you use all your tools (ratio, dose, basket, temperature, time and grind) instead of adjusting only the grind.
It depends on the density: roughly 1:1 at 0.330 g/ml, rising to about 1:4 at 0.450 g/ml, rather than a fixed 1:2.
Keep reading
Measure your density, get a starting recipe, and use the guides below to dial in the rest.
Fill a cylinder, weigh it and get your density in g/ml.
Read the guideAppendix
A single coffee seed contains about 4,500,000 cells, containing around 2,000 soluble compounds (substances that can be dissolved by water).
Only 30% of a coffee seed is soluble. The other 70% is made of insoluble fibres and carbohydrates that create the seed’s structure. About 20% of the seed contains good solubles and the other 10% are bad and taste awful. To brew the best cup of coffee we have to release the good solubles while leaving the bad ones locked up in the cells.
Luckily the bad solubles move slower and take longer to dissolve than the good solubles, so we can use all our tools to control the extraction:
•
The more solvent water you use, the more you extract.
•
The longer water is in a cell, the more solubles are extracted.
•
The higher the temperature, the more you extract.
•
The more contact time, the more you extract.
•
The finer you grind, the more you extract.
The soluble compounds are of different shapes and sizes but fit into four main categories. Being different sizes, they extract at different rates.
The easiest to dissolve.
Lipids are the natural fats and oils found in coffee seeds. They aren’t technically soluble in water, but water can still release them from the coffee cell as an emulsion. Brewing methods that use metal filters, like French press and espresso, allow lipids to pass through into the cup, producing the mouthfeel those methods are known for. The pores in paper filters are so small that they prevent most lipids passing through. Drip brewing methods like pour-over contain only about one tenth of the lipid content of methods that use a metal filter.
When coffee is roasted, the Maillard reaction produces melanoidins, which are responsible for the browning colour of coffee, in both seed and liquid form.
Carbohydrates make up 50% of a dry coffee seed’s total mass, yet only some of the carbohydrates are soluble. Their main role is to add sweetness and earthy flavours. Espresso also contains 1.5 grams of soluble fibre per cup, so 3 coffees a day could provide 20% of your recommended daily intake of soluble fibre. The sugars are big molecules: if you under-extract you get all the little acids and not enough of the big sugars, but you don’t want to over-extract and go all bitter.
Water can only extract solubles from cells it can touch. Grinding the coffee seed increases the number of cells water can access.
For context, the thickness of one of the hairs on your head is 50 to 70 microns.
200 microns
Espresso: about 70,000 particles per seed, 64 cells per particle.
900 microns
Filter: about 695 particles per seed, 6500 cells per particle.
1500 microns
French press: about 62 particles per seed, 72,000 cells per particle.
Grind too fine and the basket holes clog; too coarse and the puck resistance is insufficient. Grind changes are your worst tool and should be your last resort: change everything else first.
Appendix
A green coffee seed is small, dense and very hard. It is one of the hardest substances in the plant kingdom. When coffee is roasted, the increased temperature and the transformation of water into gas create high levels of pressure inside the seeds. These conditions change the structure of the cell walls from rigid to rubbery, creating a physical change in the coffee. This happens because of the presence of polysaccharides (bonded sugar molecules).
The internal matter pushes out towards the cell walls, leaving a gas-filled void in the centre. This means the seeds expand in volume as they decline in mass. Much of the gas build-up is carbon dioxide that is released after the roast.
Water makes up around 10 to 12% of processed and dried green seeds, but roasting reduces this to around 2.5%. As well as the water already present in the green seeds, additional water is created by chemical reactions, but this is vaporised during roasting, physically changing the coffee. The loss of moisture and the transformation of some dry matter into gases is why seeds have a reduced overall mass after roasting. On average, seeds lose 12 to 20% of their weight.
Roasting increases porosity, making the seeds less dense and much more soluble. The longer the roast, the more pronounced the structural transformations. Seed density decreases continuously, as more gases are developed over time.
When you see the coffee seeds start to turn brown, the Maillard reaction has started. This happens at around 150°C (302°F). During the Maillard reaction, gases including carbon dioxide, water vapour and some volatile compounds are created. The internal pressure increases enough to break the cell walls of the seeds, making a pop. This event is known as first crack.
After the first crack, the roast changes from an endothermic reaction (the seeds absorb heat from the drum) to an exothermic one (the seeds release heat). During this stage the physical transformations continue: seeds increase in porosity, oils migrate to the walls of the cells, and the colour darkens.
The time after first crack is sometimes referred to as the development time. Things happen quickly! The coffee changes in appearance and flavour most quickly at this point. Literally 30 seconds can produce a different-tasting cup of coffee. How much development time to apply to a roast becomes quite personal. Everyone likes a different style of coffee, and it is usually after first crack that most people make their adjustments to how a coffee tastes: longer roasts, shorter roasts, darker roasts, lighter roasts, and combinations of them all.
If the roast continues long enough, coffee will go through a second crack. This is slightly softer sounding than first crack, and it is usually when oils begin to migrate from the inside of the seed to the outside. The cellular matrix actually fractures here, allowing oils to migrate outward.
A very lightly roasted coffee seed remains very hard, dense and comparatively heavy. If you have ever tried to hand-grind lightly roasted coffee you will know how hard it is: you can barely turn the crank, it’s like grinding rocks. It is very difficult to extract enough without the coffee being sour and under-extracted, and all your tools and parameters need to be dialled up to the max to get to those sugars.
On the other hand, a very darkly roasted coffee seed has become very brittle, and you can easily crush it with your fingertips. It is very easy to over-extract and end up with bitter coffee, so you need all your tools dialled right down.
An intuitive guide to coffee solubles, extraction and TDS – Hand Ground
Physical changes coffee beans experience during roasting – Perfect Daily Grind
The Stages of Coffee Roasting – IKAWA Coffee
Craiyon, formerly DALL-E mini
