Espresso Masterclass #5 — Puck Prep: WDT, Distribution & Low-Pressure Tamping

Everything that happens inside a 58 mm basket before the pump starts decides what the pump can do.

Part five of Espresso Masterclass, the ten-part series that builds the scientific foundation behind our Espresso Alchemy masterclass. In the first four posts we established espresso as a physical process, replaced taste-guessing with TDS and extraction yield, dismantled the 25-second and 9-bar myths, and looked hard at what the grinder actually produces. This post takes the next logical step: what you do with those grounds once they land in the basket.

Puck preparation is where the most extraction variance is won or lost, and it is also where coffee culture carries the most unexamined folklore. Below, seven chapters separate what the evidence supports from what has simply been repeated for twenty years.

Chapter 1

The Puck Is a Filter Bed, Not a Coffee Cake

Before any tool touches the grounds, it helps to be precise about what the puck actually is. It is not a compressed tablet of coffee that water dissolves from the outside in. It is a packed bed of porous particles — a filter medium — through which pressurised water must travel. Darcy's law tells us that flow through such a bed is governed by permeability, and permeability is set by particle size, packing density and, critically, uniformity. Water is lazy: it takes the path of least resistance. Any region of the bed that is less dense than its neighbours becomes a highway, and everything around it is starved. Puck preparation, properly understood, is not decoration or ritual. It is the deliberate engineering of a uniform permeability field so that every gram of coffee sees roughly the same water at roughly the same time.

What you'll take away

• Extraction evenness is a property of the bed, not the machine.
• Permeability differences of a few percent are enough to redirect flow.
• Every prep step should be judged by whether it makes permeability more uniform.

Chapter 2

Clumping — What Your Grinder Actually Delivers

Grinders do not deposit a tidy pile of independent particles. Burrs generate triboelectric charge, fines cling to coarser particles, and the exit chute compacts everything into clumps that can be several millimetres across. A clump is a locally dense lump surrounded by locally loose grounds — precisely the permeability defect described in Chapter 1. Worse, clumps are invisible once the surface has been levelled. Single-dosing with an RDT spritz of water before grinding measurably reduces static and clumping, and knock-boxes, bellows and blind shakers all attack the same problem from different angles. But none of them fully declump; they reduce severity. The realistic goal is not zero clumps, it is clumps small enough that stirring can break them apart.

What you'll take away

• Static and clumping are grinder-side problems, not barista error.
• RDT (a light water spritz on beans) reliably cuts static and retention.
• Levelling the surface hides clumps rather than removing them.

Chapter 3

WDT — Needle Gauge, Depth and Stir Pattern

The Weiss Distribution Technique, devised around 2004 by John Weiss, is the most evidence-backed prep intervention available. Fine needles are stirred through the dose to break clumps and randomise particle placement. The details matter more than most people assume. Needle gauge sits in a narrow band: needles much thicker than about 0.4 mm start creating their own voids as they withdraw, while 0.3–0.35 mm needles pass through the bed without leaving structure behind. Depth matters just as much. Raking only the top few millimetres leaves the lower third of the bed exactly as the grinder dropped it, and that is the layer under the most compression during the shot. Stir with intent into the bottom third, use a consistent pattern rather than a random scribble, and finish with a gentle levelling pass. Reported refractometer comparisons put the gain from proper WDT at up to roughly 1.5 percentage points of extraction yield over tapping alone.

What you'll take away

• Use 0.3–0.35 mm needles; thicker wires create the defects they should remove.
• Reach the lower third of the bed — surface raking is largely cosmetic.
• Adopt one repeatable stir pattern so the variable stops moving.

Chapter 4

Distribution vs. Levelling — What Surface Tools Can and Cannot Do

Spinning levellers, wedge distributors and palm-tap routines all operate on the top of the bed. They are excellent at producing a flat, horizontal surface and at moving grounds from a mound into the corners of the basket. What they cannot do is reach into the bed and break clumps at depth. Aggressive spinners can even be counterproductive: a leveller set too deep smears and compacts the upper layer, creating a dense crust over a looser interior — a permeability inversion. The sensible hierarchy is declump first (WDT), level second (spinner or a light tap), tamp third. Treating a leveller as a substitute for needle work is the single most common prep mistake among otherwise careful baristas.

What you'll take away

• Levelling makes the surface flat; it does not make the bed uniform.
• Set spinners shallow — deep settings compact rather than distribute.
• Correct order: declump, then level, then tamp.

Chapter 5

Tamping — Why Level Beats Hard

Tamping is the most over-mythologised step in espresso. The folklore figure of 30 lb (roughly 13.6 kg) has no experimental basis. Controlled work — including studies from Christopher Hendon's group at the University of Oregon and independent refractometry comparisons at 10, 15 and 20 kg — finds no statistically meaningful difference in shot time, TDS or extraction yield above a modest threshold. The physics is simple: coffee particles rearrange into their tightest practical packing at around 8–10 kg, after which additional force compresses almost nothing and simply loads your wrist. What does matter enormously is levelness. A tamp that is off-axis by even a degree or two produces a wedge-shaped bed, and the thin side becomes a guaranteed channel under 9 bar. Use a self-levelling or palm-guided tamper, press straight down, and stop caring about the number.

What you'll take away

• Anything from roughly 10 kg upward behaves the same — chase repeatability, not force.
• Tamp levelness is a first-order variable; tamp pressure is not.
• A calibrated or self-levelling tamper removes the variable entirely.

Chapter 6

Basket Geometry, Dose and Headspace

The basket is not a neutral container. Wall taper, hole pattern, open area and floor thickness all shape the flow field, and precision baskets exist precisely because inconsistent perforation produces predictable dead zones. Dose has to be matched to basket volume: overfilling leaves no headspace, the shower screen imprints the puck and the bed is disturbed before the shot begins; underfilling leaves the puck free to expand unevenly and can allow water to pool above it. A useful working check is the nickel test — after tamping, a coin placed on the puck should just kiss the screen when the portafilter is locked in. Beyond that, keep the basket scrupulously clean; oil and stale fines in the perforations change open area shot by shot.

What you'll take away

• Match dose to the basket's stated capacity rather than to a favourite number.
• Leave a few millimetres of headspace so the puck can swell without screen contact.
• Clean baskets daily — blocked holes are an invisible permeability variable.

Chapter 7

Verifying Prep With Data

The point of all this is to be measurable. Prep changes should be validated the same way any other variable is: hold everything else constant, change one thing, and read the numbers. Track dose, yield, shot time, TDS and calculated extraction yield across at least three or four shots per condition, because single shots are noise. Then read the physical evidence — the spent puck. A clean, evenly textured puck that releases as one disc suggests a uniform bed; craters, soft wet patches or a puck that crumbles into wet and dry regions point to specific failures upstream. Combine numbers with a bottomless portafilter and you can usually attribute a fault to a stage: clumping, levelling, tamp angle or dose.

What you'll take away

• Change one prep variable at a time and average several shots.
• Falling variance in EY is a better success signal than a single high reading.
• The spent puck is free diagnostic data — read it every session.

Who is this series for?

Baristas who already pull acceptable espresso and want to know why it works; café owners who need prep routines that survive a busy service and multiple staff; competitors who need repeatability under pressure; and serious home enthusiasts who own a scale and a refractometer and are tired of contradictory advice. If you have ever been told to tamp at 30 lb without anyone explaining where the number came from, this series is written for you.

What's next in the series

• #6 — Staccato & Layered Pucks — Engineering the Bed
• #7 — Pre-infusion & TCF — Timing the Start of the Shot
• #8 — Pressure Profiling & the SOUP Method
• #9 — Reading the Shot — Channelling, Tiger Stripes & Spent-Puck Analysis
• #10 — Building a Data-Driven Espresso Practice at Home

Espresso Masterclass is a 10-part educational series from Tasse Coffee Roastery exploring the science behind exceptional espresso, and the foundation for our Espresso Alchemy data-driven masterclass in Takadanobaba, Tokyo. Post 5 of 10.