Coffee Roast Levels Chart

How you can make better-tasting coffee quickly and easily.

It can be very difficult for anyone, being faced with a new bag of beans and figuring out how to make a nice-tasting cup of coffee. There are so many parameters that could be changed, and yet most Baristas are using one Fixed Recipe to extract coffee:

How most people make espresso, using one, single, Fixed recipe.
(SCA) Our research suggests that the average barista uses a 1:2 brew ratio when extracting espresso and uses weight for output measurement. The average shot of espresso starts with an 18–20 gram dose, has an output of 36.5 grams, and is extracted in 25–30 seconds, at 9 bars of pressure and 200°F, using pre-infusion, through an 18-gram basket. (source). 
The grind size is adjusted to achieve the 25-30 second time.

How most people make filter coffee is also one single, Fixed Recipe :
Coffee-to-Water Ratio: To achieve the Golden Cup Standard, the recommended coffee-to-water ratio is 55 g/L ± 10%. (a ratio of 1:18)
An alternative common recipe is  60 g/L, which has a ratio of 1:16.
Coffee Preparation Temperature: To achieve the Golden Cup Standard, water temperature at the point of contact with coffee is recommended to fall between 200°F ± 5° (93.0°C ± 3°).
Few people are varying the temperature or the other parameter ranges provided.

Over-extracted coffee is bitter, and under-extracted coffee is sour.

The most influential and widely cited reference is “The Extraction of Flavor Components from Roasted Coffee Beans” by E. Lockhart and H. G. Pickles, published in the Journal of Food Science in 1964.
In this paper, the authors conducted sensory evaluations of coffee samples brewed with varying degrees of extraction. They found that under-extracted coffee tended to be sour or sharp in taste, while over-extracted coffee tended to be bitter or astringent.
This research has been cited and built upon by numerous subsequent studies on coffee brewing and extraction and has become a foundational principle in coffee science.

The once popular Coffee Percolator has fallen out of favour because endlessly boiling coffee always makes bitter coffee.

Most people adjust their recipes only with grind size adjustment and then only to achieve an arbitrary extraction time, irrespective of taste.
Most people blame the beans or roaster when this one fixed recipe does not produce a pleasant coffee.
Milk and sugar often compensate for overly bitter or sour coffee.

When adjustments for taste are made, limiting adjustments to grind size changes makes the process difficult. Many people report taking most of a bag of coffee to ‘dial it in’.

The Problem is one fixed recipe can’t work.

The recipe required to make a good-tasting cup of coffee needs to change with the coffee bean roast levels.

There needs to be a higher understanding of changing the coffee extraction recipe to suit the Roast level.

Imagine the same situation for cooking meat. A fine fillet steak is perfectly cooked when seared on either side at a high temperature for a few minutes. A beautifully marbled piece of Brisket is perfect, cooked for 6 hours at low temperature. Recipes are changed to suit the cut of meat. Imagine if the advice was to cook all cuts of meat for 1 hour at mid-temperature. Most meats would be ruined.

You can make every coffee much sweeter, less bitter, and less sour if you match the recipe to the Roast level.

What is the Coffee Ratio

The term ratio refers to how much ground coffee you use, also known as the ‘dose’, to how much the weight of the coffee you make, known as the ‘yield’. For example, in the Fixed Espresso recipe using an 18g dose, stopping the shot at 36g yield produces a dose:yield ratio of 1:2.

The ratio concept is also used in the production of filter coffee. The 55g dose making 1000g yield is a 1:18 ratio.

Water is a powerful solvent. Because of its polarity and ability to form hydrogen bonds, water makes an excellent solvent, meaning it can dissolve many different molecules.

That is exactly what you are doing when making coffee, dissolving flavor compounds out of the coffee grounds.
The more solvent you use, the more compounds you extract.

Ratio is THE most important parameter to change when making coffee. It would be like playing golf using only a putter if you never vary your ratio. That might be OK for the kids playing mini golf, but you want to be out there with the pros like Tiger Woods, using all your tools to their full abilities to make the best-tasting coffee possible, don’t you?

Coffee roast levels explained.

Dark roast coffee beans stay in the roasting machine longer and at a higher temperature than light-roasted beans. The beans lose more moisture in this process, resulting in a less dense coffee. Steam weakens the coffee bean walls as it escapes, making them brittle and easy to extract coffee from. You need a recipe to extract less with dark roasts, or the coffee will be bitter.

Since lightly roasted beans aren’t left in the roasting machine for as long as dark roasted beans, more moisture remains inside the bean, resulting in a denser coffee. The beans are heavy. Light roast coffees are characterized by their light brown color and lack of oil on the surface of the beans. The coffee bean walls remain very hard, and coffee extraction is difficult. With light roasts, you need a recipe that will extract more, or the coffee will be sour.

Light Roast Vs Dark Roast Taste

Dark roasts burn off fruit notes, and you taste less of the individual character of the different varieties of coffee, leaving just deep roasted flavors, even as far as ashy char. Some believe this is “Strong Coffee”, but caffeine is burnt off in the process, so light roasts contain more caffeine than dark. There is simply a consistent, overwhelming charred taste like burnt toast.
Some people do develop a preference for this flavor profile, but this has become less popular these days.

Originally, dark roasts were used to cope with lower-quality blends with a high proportion of Robusta, which was common for a period when Arabica was in short supply. Robusta can taste rubbery; dark roasts mask the problem.

Around this period, the Espresso ‘Normale’ 1:2 ratio became a standard.

Very dark roasts are easy to over-extract and produce bitter coffee, so many people add cream and sugar to counter the bitterness.

Very dark roasts will taste less bitter and sweeter, with lower extraction ratios, which the Espresso Ristretto 1:1 ratio is designed for.

The same principle of lowering the ratio, using less solvent water, and extracting less also applies to filter coffees. Dropping the 1:18 golden ratio to 1:16 makes a better-tasting coffee for darker roasts.

Lighter roasts preserve complex flavor profiles of differing coffee varietals and processing methods. This has led to the modern-day shift towards mid-roasts. The median roast level is now more suited to a 1:2.5 ratio for espresso.
Very light roasts can taste quite sour with insufficient extraction. The Espresso Lungo 1:3 ratio,  and the Espresso Lungo Allongé 1:4 ratio are designed for this.

The same principle of increasing the ratio, using more solvent water, and extracting more applies to filter coffees. Increasing the 1:18 golden ratio to 1:20 helps address sourness in light roasts.

Coffee roasting is less precise and reproducible than you imagine. Changes in the bean happen very fast at the development stage of roasting. Changes of a few seconds or degrees temperature can have a marked difference in the result.

Coffee Roast Names Can Be Deceiving

There is very little naming standardization, making it difficult to know what you have from the packet.

Naming scheme: Espresso Roast / Filter Roast

Many coffee bags these days are labelled either Espresso Roast or Filter Roast. Espresso, as a method, has a short contact time of 20-30 seconds with the coffee, so IF you are only using one fixed recipe of a 1:2 ratio, it needs to be a darker roast. Lighter roasts will taste under-extracted and sour unless you know how to change the recipe appropriately.
Filter coffee as a method has a longer contact time of 4 minutes. IF you use one fixed recipe of a 1:18 ratio, it must be a lighter roast, or it will taste over-extracted and bitter.

The problem is twofold. First, it is a very limiting labelling scheme. The truth is you can make all roast levels with all methods of coffee as long as you use different recipes for different Roast Levels. You can safely ignore the label provided you are not trying to use one fixed recipe. The second problem is that roasting is very imprecise; it does not produce just two roast levels but a much bigger, diverse spectrum. Coffee roasting is non-binary.

Naming scheme: Dark roast / Medium roast / Light roast.

Another labelling scheme is 3 roast levels: dark roast, medium roast, and light roast.
This at least allows the purchaser to avoid dark roasts if they don’t like them or buy them if they do.
But again, roasting produces a big spectrum, and there is no standardized definition of a mid-roast, for example. You will still have a hard time if you try to use one fixed recipe.

There have been attempts to give roast-level names more granularity; they are usually named after cities and are generally achieved by using what is known as the drop temperature. The coffee is ‘dropped’ from the roaster when it reaches the desired temperature.

Naming Scheme: Cities

LIGHT ROASTS 180 – 205°C

   ☕ Light City Roast

         Very lightly roasted coffees between the browning phase and the Cinnamon Roast area.

  ☕  Half City Roast

 Lightly roasted coffees between the browning phase and the Cinnamon Roast area.

    ☕ Cinnamon Roast

They are also known as Light Roast. Officially, it is coffee roasted to 385°F (196°C). Lighter in color than the traditional American roast, it has a toasted grain-like taste with sharp acidity and little sweetness. Some describe it as “grassy.” Roasters achieve this by taking the coffee just before the “first crack.”

   ☕  New England Roast

Coffee roasted to 401°F (205°C), lighter than the traditional American Roast. It will be a moderate light brown with some differences in the colors of the beans. Like the Cinnamon Roast, it is grain-like in taste, with a complex, sharp, almost sour acidity. Its strength is that it allows the origin characteristics of the beans to shine through. This is an easy temperature to know because it is also the point of “first crack,” where rapidly expanding steam and carbon dioxide causes the beans to “pop” and expand.


   ☕ American Roast

This is the official name for coffee roasted to 410°F (210°C). It is also the traditional taste you can expect in any East Coast American coffee shop or restaurant. In short, medium brown in color, with slightly muted acidity, and a decent display of the origin characteristics.

 ☕ Breakfast Roast

Coffee roasted anywhere from American Roast to City Roast.

   ☕ City Roast

Aka Medium Roast

This represents coffee that has been roasted to 426°F (219°C). It is darker than the traditional American Roast and tends to be smooth in taste, with less acidity.  This is just before the roast characteristics start to kick in, so you are primarily getting the flavor of the coffee variety, not of the roast. It is also sometimes known as Medium Roast and is the degree to which most specialty coffees are roasted.

  ☕  Dark Roast

Any roasted coffee that is darker than the traditional American norm. That means City, Vienna, French, Italian, or Turkish Roast all fall into this same, broad category.


 ☕Full City Roast

By definition, this is coffee roasted to 437°F (225°C) and is right at the “second crack.” It is a bit darker than the traditional American Roast and may have tiny droplets of oil on the beans. This roast is often smooth in taste, with less acidity. Roast characteristics start to become apparent, so you might be losing some of the distinctive taste characteristics of the coffee variety. This roast is also sometimes known as “medium+.”

 ☕ Vienna Roast.

Coffee roasted to 446°F (230°C). It is darker than the traditional American Roast and lighter than French Roast. Oils on the beans. Less acidity. Some call this a Medium-Dark Roast.

☕After Dinner Roast

Coffee roasted anywhere between Vienna Roast and Italian Roast. Being low in acid, it is designed to go well with dessert.


 ☕   French Roast temp 240

 Coffee roasted to 464°F (240°C). It is darker than the traditional American Roast and will likely be shiny with oils. The taste is almost entirely due to the quality of the roast and may range in taste anywhere from rich and bittersweet to thin-bodied and burned.

  ☕  Italian Roast 245°C

Coffee roasted to 473°F (245°C). The resulting coffee is much darker than the traditional American Roast and is, in fact, darker than French Roast. It may even be burnt. The resulting cup is usually bittersweet in flavor, with an overlay of burned and charcoal-like tones. Additionally, it is often thin-bodied, which blenders try to mediate by mixing in robusta beans to add extra body. Because Italians invented espresso, it is also sometimes known as Espresso Roast.

    ☕Darker that Dark Roast

By reaching 465ºF (241ºC), you are in the Vienna Roast range. The beans are getting darker. Beyond this, the roast takes over completely: the sugars are heavily caramelized and burnt, the internal cellulose of the bean starts to carbonize, and the bean itself expands. The second crack is now quite rapid and nearing its end. The temperature is at 474ºF (246ºC), and you have attained a French Roast. Keep going to 486ºF (252ºC), and you are at what is known as Italian Roast. Go much farther, and you have a roaster full of charcoal and a fire if you are not careful.


These names sound great, but you hardly see them used these days. This is because they are very restrictive for the roasters. Roasters use more parameters than just the drop temperature, such as rate of rise, long slow roasts, short sharp roasts, and different Roast profiles to suit different bean sizes, densities, moisture levels, and varieties.

Naming scheme: Roast levels 1-7

Another labelling scheme uses Roast Levels, such as 1-7, to indicate a degree of roast. Still, even with this, it is very hard for Roasters to maintain consistency when the roasting process is so variable.

Naming scheme: Agtron

Yet another labelling method is to use Agtron.

The Agtron® is a spectrophotometer that uses infrared light to determine roast degree. The higher the Agtron number, the lighter the roast. Agtron parameters range from about 55 on our lightest roasts to 25 on the darkest. This digital device validates the roaster’s decisions concerning the degree of roast.

The problem with using Agtron at labelling for the consumer is it doesn’t account for how the Roast was done. A long, slow roast may be lighter in color than a shorter, high-temperature roast that is darker. It’s useful information for a roaster because they know how the Roast was done, so it can tell them when to stop. But Agtron can be misleading for consumers who don’t know how the Roast was done.

The problem is that you need different recipes for different Roast Levels, but the labelling is non-standardized and inadequate to inform you how to make the coffee taste best.

The Solution: Using Density to Assist with Extraction (DAE).

There is a very strong correlation between density and roast level.
With a very darkly roasted bean, you can easily break it with your fingers; the bean is not very dense.
With a very lightly roasted bean, it is impossible to break with your fingers; even a hammer has difficulty, and the bean is very dense.

Density can be measured by dividing the weight in grams by the volume in mls (g/ml).

Settled density is the non-destructive method of putting dry roasted beans in a 100ml volumetric measuring cylinder, tamping them level to measure the volume in mls, and weighing them, preferably with a 0.1g accuracy set of digital scales.

There are other methods of measuring density, such as water displacement, but this is destructive because the beans get wet.

What is the real-world density range for roasted coffee? was started in 2020 with research on 20 bags of coffee intentionally roasted to different levels.

The website is a self-learning system; the more it is used, the more accurate it gets.

The 2023 model is based on measurements taken from over 4,000 bags of roasted coffee worldwide. Giving a 98% confidence that the real-world density range accurately represents worldwide coffee roast levels.

The mean is 0.400 g/ml

The standard deviation is 0.039g/ml

95% of coffees will fall between 0.321g/ml and 0.478g/ml.

Making the density range 158 0.1g steps, from darkest to lightest roast.  

Density generated Recipes

The SCA Standards are not, in fact, single recipes; each parameter has ranges. For example, the espresso ratio range of 1:1 to 1:4, or the filter temperature range of 90-96c. The recipe ranges need to be applied to the Roast Level ranges.

The ‘model’ is the math that has generated the recipes.

You are not expected to do the math; this is here to explain the recipes.

Relatively simple linear proportional equations apply the SCA ranges for each parameter to the real-world density range, generating the 158 recipes for each coffee-making method. Contrary to some critics of the system, these are not my recipes to my tastes. They are simply the SCA recipes an experienced barista would make to suit a roast level.  They are the adjustment ranges you should already be using.

There are 4 forms of equations in the model. Inclining, declining, derived and stepped.

Inclining parameters: (ratio, temperature, time, pressure, bypass)


This is like the E=MC2 of coffee formula for using measuring and math to solve coffee extraction. It looks a bit terrifying, but it is a simple linear equation to spread each SCA-supplied parameter range over the real-world density range.

For example:
There is the real-world density range:
minDensity = 0.321
maxDensity = 0.478

There is the SCA Espresso Ratio parameter range:
minRatioEspresso = 1
maxRatioEspresso = 4

Just to show the bounds work as expected:

The minimum ratio that we already know:
If ThisDensity = 0.321, then ThisRatio =1, from ((((0.321-0.321)/(0.478-0.321)))*(4-1))+1

The maximum ratio that we already know:
If ThisDensity = 0.478, then ThisRatio = 4, from ((((0.478-0.321)/(0.478-0.321)))*(4-1))+1

An example calculation for a density in between the bounds:
If ThisDensity = 0.400, then ThisRatio =2.5 from ((((0.4-0.321)/(0.478-0.321)))*(4-1))+1


Stepping through each of the 158 real-world density range from 0.321 to 0.478, calculating the espresso ratios, the results can be displayed as a graph.

The same proportional formula works for each parameter within its known ranges.
For another example Filter method, temperature:

Coffee Roast Levels to Espresso Ratio
Filter Coffee Temperature by Roast Level

Declining parameters: (dose, grind)

The declining dose may seem counterintuitive. Why do you maximize the dose when you need to minimize the extraction? It is because the ratio is low, so to increase the yield, you need a bigger dose. Conversely, when you need to extract the most, you need a big ratio. It’s much easier to achieve a big ratio with a smaller dose.

minEspressoDose = 14g
maxEspressoDose = 25g

Declining formulae are of the form 1-.
ThisEspressoDose = (1-((ThisDensity-minDensity)/(maxDensity-minDensity)))*(maxEspressoDose-minEspressoDose)+minEspressoDose

Derived parameters: (yield, VST baskets)

Yield = dose * ratio

VST baskets are +/- 1g of the dose.

Stepped parameters: cup size

When all the extraction parameters are put together, the result is the model generates 158 sequenced, numbered recipes that use the full range of extraction parameters to achieve recipes that will go from extracting the least to extracting the most. Matching the real-world density range. A set of 158 recipes for each coffee method.

The recipes are effective because they use each parameter in concert to change the level of extraction.

The huge benefit of the 158 pre-prepared recipes is that they use all your extraction parameter tools to their full extent. Whereas with the fixed recipe, people are only adjusting grind size.

Grind adjustments are constrained by physics. Grind too fine, and your puck chokes. The grind is too coarse, and there is insufficient puck pressure.
It is not possible by grind adjustment alone to cover extraction requirements for the entire density range from the fixed recipe.

Adjustments by Taste

The next huge benefit is having 158 sequenced recipes; now, you can adjust by taste. If the coffee tastes sour, you have under-extracted. To extract more, use a recipe further up the list.

If the coffee tastes bitter, you have over-extracted. To extract less, move down the list.

Moving in big steps of 20 recipes at a time has two advantages. It means you make a significant change that will likely fix the problem. It also means you can swiftly sweep through the list of recipes to zero in on the best-tasting area quickly.

There are two ways to use the recipes.

  1. Suppose you don’t measure your coffee density, the site provides the middle recipe by default as a starting recipe.

There is less than 20% chance this will be correct. There are 4 stepped recipes up or down to cover the range.

68% (one standard deviation) will be within 2 adjustments. 30% of the coffees may take up to 4 tries. Because it is based on your taste, the outcome is always great coffee.

  1. You can make the process considerably faster by measuring your coffee density.

Instead of using the default starting position, your measured resultant recipe will be bang on 70% of the time.

The margin of error is still large. 30% of the time, adjustment will still be required.

Settled density is affected by many factors, especially seed size and packing.

It’s not expected that one density measurement could give you a 100% accurate result every time, no matter how accurately you measured it, even if you used the more accurate displacement density method.

With a closer starting position, 95% of the time, you will achieve your goal of great-tasting coffee with only one step.

Because the coffee selected is based on your taste, the outcome is always great coffee.


Density Assisted Extraction (DAE) is a Probabilistic System.

Probabilistic data can be used to add more value to deterministic datasets and to scale deterministic data models. If something is unknown in a deterministic dataset, enriching the data with probabilistic data can offer more accurate insights. (source)

Settled density is not deterministic, and there is no exact match to roast level or extraction. But it is probabilistic.  The connection between density and extraction required is high. You are much more likely to be close by using density.

Probalistic systems are commonly used. If you want to know what the weather will likely be like tomorrow, you are probably better off checking the weather forecast. Is the forecast guaranteed to be 100% correct every time (deterministic), no. But statically, it is a very good system and is widely used.

Density Assisted Coffee Extraction (DAE) will probably result in a better-tasting cup of coffee most of the time.


Roast level frequency distribution.

The 4000 measurements also reveal the Roast Level frequency distribution or the probability the coffee will be roasted to where you think it is. The main takeaway is that Roast Levels are diverse and spread across the spectrum.

If you expect an espresso roast level that will work well with a 1:2 ratio, you will be off 90% of the time.

In the chart below, you can see the density range, the probability of getting a coffee Roasted to that level, the corresponding espresso ratio, the filter ratio, and the recipe numbers.

Have a look at the row with an espresso ratio of 2.0. Of 4000 bags of coffee, 6.3% were ideally suited to the 1:2 ratio. Changing your grind size can’t achieve a sufficient change in extraction to avoid bitter or sour coffee most of the time.

Have a look at the midpoint, recipe number 80. It’s only marginally more probable at 7.2%, which is insignificant and likely the wrong extraction recipe.

However, you can use the Recipe numbers in steps of 20 recipe numbers. If coffee recipe 80 is sour, trying the next step up at 100 will extract significantly more. Still sour, keep going 120, 140, 158. That covers the entire range in 4 steps, and you are 95% likely to have found a recipe that tastes significantly better.

If coffee recipe 80 is too bitter, trying the next step down at 60 will extract significantly less. Still bitter, keep going 40, 20, 1. That covers the entire range in 4 steps, and you are 95% likely to have found a recipe that tastes significantly better.

Recipe 80 is unlikely (90% of the time) to be the best starting position.

Getting yourself a relatively cheap 100ml volumetric cylinder, typically under USD 15, and using it with your 0.1g accuracy digital scales is a process that takes less than 1 minute for each bag of coffee you buy. This gives you a much better starting position.

The recipe indicated by your measurement is likely to be correct 68% of the time. Because it gets you much closer to where you need to be, only one step will be needed to achieve the desired outcome 95% of the time, rarely 2 steps.

Recipe number steps Density Range % grand total (probability) Espresso Ratio Filter Ratio Agtron Roast Names Roast levels Roast degree Roast intention
<321 or (blank) 1.70%
>482.2 2.20%
1 321-326.2 1.00% 1 16.5 25 Darker than dark 7 Dark Espresso
326.2-331.4 0.90% 1.1 16.6 27
331.4-336.6 1.30% 1.2 16.7 30
20 336.6-341.8 1.00% 1.3 16.9 32 Vienna
341.8-347 3.40% 1.4 17 34 6
347-352.2 2.20% 1.5 17.1 37
352.2-357.4 3.20% 1.6 17.2 39 Full City
40 357.4-362.6 1.70% 1.7 17.4 41
362.6-367.8 4.40% 1.8 17.5 43 5
367.8-373 3.50% 1.9 17.6 46 Dark
373-378.2 6.30% 2 17.7 48
60 378.2-383.4 3.30% 2.1 17.9 50 Medium
383.4-388.6 7.20% 2.2 18 53 Breakfast Filter
388.6-393.8 3.80% 2.3 18.1 55 4
393.8-399 3.80% 2.4 18.2 57
80 399-404.2 7.20% 2.5 18.3 60 American
404.2-409.4 4.90% 2.6 18.5 62
409.4-414.6 5.10% 2.7 18.6 64
414.6-419.8 5.60% 2.8 18.7 66 City 3
100 419.8-425 3.50% 2.9 18.8 69
425-430.2 4.10% 3 19 71 Light
430.2-435.4 3.40% 3.1 19.1 73 Medium
120 435.4-440.6 1.80% 3.2 19.2 76
440.6-445.8 3.70% 3.3 19.3 78 2
445.8-451 2.40% 3.4 19.5 80 Cinnamon
451-456.2 3.10% 3.5 19.6 82
140 456.2-461.4 1.10% 3.6 19.7 85
461.4-466.6 1.60% 3.7 19.8 87 Half City
466.6-471.8 0.70% 3.8 19.9 89 1
471.8-477 0.90% 3.9 20.1 92
158 477-482.2 0.30% 4 20.2 94 Light City
Grand Total 100%

Using Density to Assist with Extractions (DAE) gives you 4 huge benefits.

  1. Recipes: Instead of one recipe that can’t work for every roast level, you now have 158 recipes for every roast level. The recipes are the SCA standard ranges. They use all the parameters to the fullest extent.
  2. Adjustments: Instead of adjustments being limited to grind sizes that can’t cover roast level range requirements, making dialling in difficult and often impossible, you now have a much more effective adjustment mechanism that covers the entire roast level range, is quick and nearly always works.
  3. Starting recipe: Instead of starting at a recipe that is unlikely to work, you can now start at one that is much more likely to work, reducing dialling in time.
  4. Taste: The biggest benefit is that you are finally using your taste buds to adjust the coffee. You find the recipe that tastes the best.

Compared to using one fixed recipe that is likely wrong and only adjusting grind size to achieve an arbitrary extraction time. Most people are not adjusting to taste and don’t know how to.

Using Density Assisted Extraction (DAE) solves the problem using measurement and math. Now, everyone can make better-tasting coffee quickly and easily.

Total Dissolved Solids, Extraction Yield (TDS/EY)

Now and then I get questions like:

>Is there any scientific backup to the findings?  Is EY better with the ratio you recommend?

The optimal guidelines set by the SCAA for extraction are 18 – 22%. This means that when we take the total weight of our coffee beans, 18-22% of that mass will be dissolved by water and end up in our cup.

TDS can be defined as the ratio of solubles to water in a cup. This is largely a matter of personal preference, but it is generally considered that 1.15% to 1.35% is an optimal ratio.

The solubles released during extraction determine the flavor, and TDS determines the intensity of those flavors. Increasing TDS will produce a stronger flavor, but too high may cause some flavors to overpower others completely.

This led to the creation of the Brewing Control Chart:

The Coffee Brewing Control Chart, adapted from The Coffee Brewing Handbook (page 9), by T.R. Lingle (2011), Specialty Coffee Association of America

Currently, the industry has a tool to measure extraction yield (refractometer), which, if you think about it, cannot indicate flavor qualities because the refractive index is a quantitative measurement, not qualitative.

Many people are misusing this tool and refer to ‘achieving’ high Extraction Yields. You can achieve a high Extraction Yield by boiling your coffee in a percolator for an hour. The coffee is going to taste terrible and bitter.

Measuring your TDS and EY with a refractometer is unlikely to assist you in figuring out what to change to make a better-tasting cup of coffee.  

Frequently Asked Questions:

What about the 5% of coffees that fall beyond the real-world density range for roasted coffee? Roast defects include roasting too dark or too light. If you wish to persevere, there is nothing to stop you from following the trajectory further, e.g. going below 1:1 or above 1:4 ratio for an espresso.

What if I have an automatic machine and can’t change some of the parameters, such as pressure or temperature, easily? No problem, change what you can and keep moving until it tastes right. The system will still work. The other extraction parameters compensate to an extent.

What if I can’t taste which direction to move in?  Coffee is inherently bitter. You are looking for the least bitterness—the calm spot between overwhelming bitterness and sourness. There is an element of learning to taste. You can always try making the coffees in both directions and picking which tastes the best. You don’t have to do this all at once. Keep a note of what you feel, and try something different next time.

Moving 20 x 0.1g density steps in either direction, significantly alters the extraction, because it is changing all the extraction parameters in concert.

8 jumps covers the entire range.

The direction you need to move is determined by taste. If its sour, it is under extracted, use the Extract More button that adds 20 to the entered density, and gives you the new recipe.

If its bitter, it is over extracted, use the Extract Less button, subtracting 20.

Measuring the density 70% of the time gets it right first time, but it also gets you a closer starting position, so 95% of the time a single adjustment is all that is required.

Very rarely, 4%, of the time, two adjustment steps are required.

Even more rarely, 1%, you may just be faced with a coffee defect or roast defect that can’t be fixed by extraction.

Does really work? Some testimonials:

I feel so lucky to find this way by accident. I just do not understand why it is not common knowledge. I feel ease of mind compared to my endless struggle to replicate god shots, now I drink god shot twice a day almost. (Zac, Israel)

 Love your site. Hi Richard, I thought I’d let you know that I really appreciate your efforts to illuminate the relationships of the different variables in preparing espresso. So much of the “instruction” out there in videos and blogs is an oversimplification of the process. It’s been a confusing and frustrating journey until we ultimately settled on one particular roast that we managed to dial in. Now, having the freedom to try different types of coffee (even kinds we previously thought we didn’t like) is really great! Thanks again! (Jason, USA)

I have to say this worked extremely well. Im new to the Flair (just got a Pro 2) and (good) espresso-making in general. Couldn’t get a medium roast to not taste sour for the life of me. And by sour I mean I would almost pucker up when tasting the espresso and just thinking about it I have my mouth watering in disgust as I type this. I tried the site and on my first attempt I nailed it. No sourness at all. I’m sure I can still make improvements but it was night and day! (Zac, USA)

An anecdote, a true story. I was at Ikarus Coffee Farm doing voluntary picking with a group. A young couple wanted to start a coffee shop and were there to learn. The owner Rob, was great. He had a new semi-automatic espresso machine that he let us use. It was a la marzocco  linea mini, an EK43 grinder, and the extraordinary Bourbon Pointu Pt Leroy coffee.

Rob had given this guy (we will call him John) some quick lessons in how to drive everything and had gone off to the market to sell coffee, leaving us to play with the gear.

John was doing everything right – as Rob had shown him. 1:2 ratio, 18g in 36g out in 25 seconds. But after several tries, John was not happy with the result. In fact he was looking stressed. If he couldn’t make a coffee he liked himself, how could he start a business and sell coffee?

I had already mentioned the concept of using density to Rob & John, but they hadn’t tried it.

Now, John agreed to give a go. We measured the density and found they were really small beans and roasted quite a dark roast at about 0.357 g/ml. The recipe indicated a bigger dose and a lower ratio were required.

John made the new recipe and tried it, I asked him to dip a teaspoon in, suck it, and tell me if it was bitter or sour. Bitter, he pronounced. Back to the website, use the Extract Less button. Now we are up at a 24g dose and a 1:1.3 ratio.

John made the coffee and tasted it. It was like watching the sun break out of that furrowed brow. He grabbed the coffee and ran outside. I was a bit surprised but busied myself making my coffee. He came back about 10 minutes later, he had taken it to his partner. He could not believe he had just made the best-tasting coffee he had ever had and that it was with the same beans that tasted terrible before. He was now confident he could make any coffee taste better using He was so happy and excited and said he would use the system for life. I kid you not. This happened.

The Coffee Roast Levels Chart

This provides another view on how Density Assists with Extraction (DAE) works showing all the coffee recipes available on, in an ordered list. The icons are hyperlinks to the recipes.

Recipe# Density Probability Espresso Ratio Espresso recipe Filter Ratio V60 or Switch recipe Aeropress recipe Filter recipe
1 0.321 0.18% 1 13
2 0.322 0.09% 1 13
3 0.323 0.09% 1 13
4 0.324 0.23% 1.1 13
5 0.325 0.27% 1.1 13
6 0.326 0.09% 1.1 13
7 0.327 0.05% 1.1 13
8 0.328 0.14% 1.1 13
9 0.329 0.27% 1.2 13
10 0.33 0.32% 1.2 13
11 0.331 0.36% 1.2 13
12 0.332 0.23% 1.2 13
13 0.333 0.41% 1.2 14
14 0.334 0.14% 1.2 14
15 0.335 0.05% 1.3 14
16 0.336 0.41% 1.3 14
17 0.337 0.18% 1.3 14
18 0.338 0.27% 1.3 14
19 0.339 0.27% 1.3 14
20 0.34 0.27% 1.4 14
21 0.341 0.23% 1.4 14
22 0.342 0.23% 1.4 14
23 0.343 0.36% 1.4 14
24 0.344 0.91% 1.4 14
25 0.345 1.54% 1.5 14
26 0.346 0.18% 1.5 14
27 0.347 0.55% 1.5 14
28 0.348 0.36% 1.5 14
29 0.349 0.23% 1.5 14
30 0.35 0.36% 1.6 14
31 0.351 0.45% 1.6 14
32 0.352 0.45% 1.6 14
33 0.353 0.50% 1.6 14
34 0.354 0.64% 1.6 14
35 0.355 0.82% 1.6 15
36 0.356 0.36% 1.7 15
37 0.357 0.68% 1.7 15
38 0.358 0.36% 1.7 15
39 0.359 0.32% 1.7 15
40 0.36 0.36% 1.7 15
41 0.361 0.36% 1.8 15
42 0.362 0.59% 1.8 15
43 0.363 0.41% 1.8 15
44 0.364 0.68% 1.8 15
45 0.365 1.54% 1.8 15
46 0.366 1.23% 1.9 15
47 0.367 0.32% 1.9 15
48 0.368 1.00% 1.9 15
49 0.369 0.77% 1.9 15
50 0.37 0.77% 1.9 15
51 0.371 0.73% 2 15
52 0.372 0.86% 2 15
53 0.373 0.68% 2 15
54 0.374 0.73% 2 15
55 0.375 1.68% 2 15
56 0.376 1.00% 2.1 15
57 0.377 0.68% 2.1 15
58 0.378 1.18% 2.1 16
59 0.379 0.77% 2.1 16
60 0.38 0.77% 2.1 16
61 0.381 0.45% 2.1 16
62 0.382 0.82% 2.2 16
63 0.383 1.04% 2.2 16
64 0.384 1.18% 2.2 16
65 0.385 1.32% 2.2 16
66 0.386 1.64% 2.2 16
67 0.387 1.41% 2.3 16
68 0.388 1.32% 2.3 16
69 0.389 1.09% 2.3 16
70 0.39 1.04% 2.3 16
71 0.391 1.00% 2.3 16
72 0.392 0.64% 2.4 16
73 0.393 0.91% 2.4 16
74 0.394 0.45% 2.4 16
75 0.395 0.95% 2.4 16
76 0.396 1.14% 2.4 16
77 0.397 1.09% 2.5 16
78 0.398 1.23% 2.5 16
79 0.399 1.09% 2.5 16
80 0.4 1.14% 2.5 17
81 0.401 1.36% 2.5 17
82 0.402 1.45% 2.5 17
83 0.403 1.68% 2.6 17
84 0.404 1.04% 2.6 17
85 0.405 0.77% 2.6 17
86 0.406 0.73% 2.6 17
87 0.407 1.18% 2.6 17
88 0.408 0.95% 2.7 17
89 0.409 0.91% 2.7 17
90 0.41 0.91% 2.7 17
91 0.411 1.18% 2.7 17
92 0.412 1.45% 2.7 17
93 0.413 1.27% 2.8 17
94 0.414 1.54% 2.8 17
95 0.415 0.77% 2.8 17
96 0.416 1.14% 2.8 17
97 0.417 1.23% 2.8 17
98 0.418 0.64% 2.9 17
99 0.419 0.73% 2.9 17
100 0.42 0.77% 2.9 17
101 0.421 0.82% 2.9 17
102 0.422 0.73% 2.9 18
103 0.423 1.00% 2.9 18
104 0.424 1.18% 3 18
105 0.425 0.64% 3 18
106 0.426 0.73% 3 18
107 0.427 0.82% 3 18
108 0.428 0.50% 3 18
109 0.429 0.73% 3.1 18
110 0.43 1.09% 3.1 18
111 0.431 0.45% 3.1 18
112 0.432 0.59% 3.1 18
113 0.433 0.41% 3.1 18
114 0.434 0.73% 3.2 18
115 0.435 0.45% 3.2 18
116 0.436 0.50% 3.2 18
117 0.437 0.50% 3.2 18
118 0.438 0.23% 3.2 18
119 0.439 0.50% 3.3 18
120 0.44 0.45% 3.3 18
121 0.441 0.91% 3.3 18
122 0.442 0.41% 3.3 18
123 0.443 0.68% 3.3 18
124 0.444 1.04% 3.4 18
125 0.445 0.77% 3.4 19
126 0.446 0.50% 3.4 19
127 0.447 0.27% 3.4 19
128 0.448 0.73% 3.4 19
129 0.449 0.45% 3.4 19
130 0.45 0.50% 3.5 19
131 0.451 0.23% 3.5 19
132 0.452 0.50% 3.5 19
133 0.453 0.27% 3.5 19
134 0.454 1.14% 3.5 19
135 0.455 0.32% 3.6 19
136 0.456 0.32% 3.6 19
137 0.457 0.45% 3.6 19
138 0.458 0.18% 3.6 19
139 0.459 0.14% 3.6 19
140 0.46 0.05% 3.7 19
141 0.461 0.09% 3.7 19
142 0.462 0.27% 3.7 19
143 0.463 0.14% 3.7 19
144 0.464 0.86% 3.7 19
145 0.465 0.14% 3.8 19
146 0.466 0.14% 3.8 19
147 0.467 0.27% 3.8 20
148 0.468 0.18% 3.8 20
149 0.469 0.14% 3.8 20
150 0.47 0.09% 3.8 20
151 0.471 0.27% 3.9 20
152 0.472 0.23% 3.9 20
153 0.473 0.14% 3.9 20
154 0.474 0.18% 3.9 20
155 0.475 0.05% 3.9 20
156 0.476 0.05% 4 20
157 0.477 0.23% 4 20
158 0.478 0.00% 4 20