Coffee Data Science

Staccato Espresso is Fundamentally Better than Regular Espresso

Water Contact Time

Published in

Towards Data Science

7 min readNov 2, 2021

The history of the staccato shot has been interesting because the proof that it worked was initially only in taste. Then it was proved in extraction, but still, there was skepticism in part because of the amount of work it takes to sift coffee. However, I believe I’ve figured out why staccato shots should fundamentally outperform regular shots in extraction, and the answer is water contact time.

A staccato shot is a layered shot after sifting the grinds using at least two layers. The variant that is my preference has a fine layer on the bottom (<400 um), a coarse layer in the middle (> 500um), and a mid layer on top (between 400um and 500um).

Grind Size and Extraction

Generally, it is known that finer coffee extracts faster in water than coarser coffee when all parameters like temperature, pressure, and water flow are kept the same. So when you pull an espresso shot, by the end of the shot, the finer particles are tend towards being over-extracted while the coarser particles tend towards being more under-extracted. The goal is to pull a shot just long enough that most of the coffee is properly extracted.

I did a study where I looked at extraction of individual grind sizes using a controlled scenerio so that channeling was not an issue. This experiment also had the same water contact time, water temperature, and pressure for all grind sizes. In this case, I cut the shot into three cups: First, Second, and Third. This graph shows the cumulative EY for each, and this shows the trend of higher extraction particularly in the First part for smaller grind.

Regular Shot Contact Time

For a regular shot with an even distribution throughout the shot, there will be fines towards the top of the puck that have had all the water in the cup pass through them.

Let’s say we’re working with 20g coffee grounds in, and 40g liquid coffee in. For the sake of simplicity, assume the amount of water in the puck is 20g because usually the puck will retain that much. So particles at the bottom have little water contact time by the first drop into the cup, while particles at the top of the puck have been exposed to 20g of water flow.

This means with respect to extraction, fines at the top of the puck will have a higher extraction than the fines at the bottom of the puck, and they will be the first ones to over-extract.

Additionally, for coarse particles, the coarse particles at the bottom of the puck will be more likely to be under-extracted while coarse particles at the top of the puck will be more likely to be properly extracted.

When looking at temperature, let’s assume you do a low-pressure pre-infusion. During pre-infusion, the water at the bottom of the puck will be at a lower temperature than at the top of the puck. This temperature will eventually equalize, but while there is a differential, the higher temperature will cause the top will extract faster than the bottom with all other parameters being equal. This point may be mute if you run a shot long enough, but then again, the longer you run the shot, the more likely you are to over-extract from the coffee starting from the top of the puck.

Staccato Contact Time

For a staccato shot, the fines are at the bottom. The fines are the biggest reason for the over-extracted taste because they over extract the fastest.

By the time the shot finishes, the fine layer will have less contact time than the other layers because it is at the bottom. The amount of over or under extraction for the fines will be more even. This is also true for the others layers.

Additionally, the other two layers are coarser and exposed to more contact time with water than the fine layer, causing them to extract a bit more and have a closer extraction rate to the fine layer. Each layer has a similar contact time throughout the layer so the extraction for those particle sizes should be more homogeneous than in a regular shot.

Extraction by Layer Data

I have collected some previous extraction data by layer. The main data I have is for sifted staccato and staccato tamped shots using dried spent puck weights. For sifted staccato shots, the fine layer extracted much more than the mid/coarse layers. While for the staccato tamped, the mid/coarse layer was the top layer, which had a higher extraction than the bottom layer. The staccato tamped shot is a regular shot with respect to particle distribution.

Additionally, I have collected some other data using Grounds Total Dissolved Solids (gTDS). I have multiple staccato tamped shots where I measured the gTDS for the top and the bottom halves of the shot. A higher number means there are still more extractable solids.

In most of this data, the Top gTDS was much lower than the Bottom gTDS meaning that the top half was much more extracted than the bottom. This gives some evidence to the idea that the top half of the puck is more extracted than the bottom half or rather extracts faster than the bottom.

Additional Wins

A staccato shot is not prone to channeling because by being sifted, water can flow through each layer more evenly. Each particle layer gets more homogeneous extraction per that size.

Typically staccato shots have less crema. Crema is caused by CO2 released during extraction, and I suspect it is connected to extraction rate. So if the particles with higher extraction rates are at the bottom, release gases cause less interference to larger particle sizes during the shot.

High End Grinders

One argument originally against staccato was that my grinder wasn’t great. I was using a ROK grinder, but it turns out, the ROK performed better than the Niche. Still, people argued the even higher end grinders should outperform, but fundamentally, they can not.

No matter what the grinder, a regular shot has this water contact time issue in the puck. So if you took the grounds for a shot from any grinder and sifted them, the sifted shot should always have a greater or equal extraction rate, and that extraction should be more even across the particle grinds therefore, it should taste the same or better.

Reviewing Previous Data

I pulled some data I had to put together paired shots for Extraction Yield (EY) only. For these shots, I had data at the 1:1 output to input ratio as well as the 3:1 because I used a lever machine and put another cup underneath after my shot. There were 9 pairs of shots across 5 roasts and 2 machines (Kim and Flair) where all other parameters were the same.

The staccato shot won out for EY across most of the pairs. This is the most objective measure we have to compare coffee extraction techniques. If a staccato shot is properly sifted using grounds from a dialed in shot, it should have a higher EY than if you used those same grounds unsifted.

Additionally, when looking at control charts for Total Dissolved Solids and EY across hundreds of shots, staccato shots trend very high in TDS vs EY.

All of this understanding for the staccato shot is based on actually sifting correctly. If one expects 40% of their grounds to be less than 400um based on particle distribution but only sifts 14%, then that means the sifted shot could taste completely different. This requires improvements to the sifting process like putting an agitator in the sifter which almost halves sifting time and effort. Weighted paper agitators can bring that time down even m

I firmly believe we are at the boundary of 4th wave coffee. While it takes time to sift, I’m positive new machines will be built to be faster and more automatic.

Staccato shots hold the promise of higher extraction.

If you like, follow me on Twitter, YouTube, and Instagram where I post videos of espresso shots on different machines and espresso related stuff. You can also find me on LinkedIn. You can also follow me on Medium and Subscribe.