Taste first, reveal the recipe later

Set up a fair comparison

What you need

Two 20 g portions of the same coffee, the same brewer and grinder, a scale, timer, two similar cups, a separate mapping note and ideally a helper who prepares or codes the cups.

The factor you change

Recipe-water ratio, with concealed labels. A: Preparation A: 20 g / 300 g. B: Preparation B: 20 g / 340 g.

What to keep consistent

Use the same coffee, 20 g dose, grinder, 93°C kettle setting and pouring convention. A uses 300 g input water and B uses 340 g. Tasting volumes, cup appearance and serving temperature should be comparable; do not let the taster inspect the A/B preparation mapping.

These are planned settings. Template durations are editable placeholders, not targets to force or measured results. For espresso the water/yield field means collected espresso; otherwise it means input water.

Setup saves two recipes only, without changing existing recipes, creating brew records or deducting stock. Choose a bean bag and confirm usage separately when recording each brew. Local storage details.

Run the comparison

1. The preparer creates recipes A and B and keeps the preparation identities private. The taster should not view the journal recipe names or the mapping during the tasting.
2. Prepare the two recipes consistently and note actual yield and timing. Use equal small tasting portions in similar cups to avoid revealing the recipe from serving volume.
3. Have the helper use the Blind Tasting Order Generator with two samples, agreeing privately that Sample 1 means recipe A and Sample 2 means recipe B. Position codes belong to the taster; sample numbers stay hidden.
4. Write descriptors, confidence and enjoyment for each coded position before revealing anything. Keep descriptions such as stronger or drier separate from a preference.
5. Reveal the mapping and transfer the coded observations to the matching A/B brew records. Repeat with a fresh concealed mapping and varied preparation order; do not assume a random draw must reverse the order.

Record what actually happened

Preparation identities, concealed mapping held by the helper, presentation order, serving conditions, pre-reveal observations, confidence, reveal and matched recipe records.

Use the unfilled table below as a prompt. To retain your own observations, download the blank CSV sheet or save A and B as separate brew records in My brews. Text typed into a downloaded file stays in that file; it is not synchronized to the journal.

This download contains labels and blank observation cells, not fabricated sample outcomes.

Interpret without overclaiming

Compare whether the preference is stable after reveal and across repeats. One correct guess or one preferred cup is not statistical evidence. A changed preference may reflect serving conditions, expectation or normal variation; keep the uncertainty in your notes.

What this cannot establish

When you prepare, label and taste alone, you may remember the mapping: call that comparison unblinded. This two-sample exercise is not a triangle test, a balanced Latin square or professional certification. The templates use A/B preparation names and must stay hidden from tasters.

Choose the next test

Repeat the same pair on another occasion before testing another variable. Save the mapping in notes only after tasting if you want to preserve the blind period.

In the journal, select the actual A and B records and check dates before drawing a conclusion. The paired comparison link uses the latest saved record for each experiment recipe; it is not a statistical test and may select different sessions. A repeated pair with reversed preparation order is more informative than a single cup, but it still describes your coffee and conditions.