Why do professional tasting panels dilute whisky to 20% ABV before they nose it?

Key takeaways

• 20% ABV is the documented industry standard for nosing, not a preference. “The standard industry practice for ‘nosing’ is to dilute samples of spirit to 20% ABV to prevent sensory fatigue” (Baxter, Réveillé & Conner, Journal of the Institute of Brewing, 2024).
• Ethanol suppresses the very aromas a panel is paid to find. Dilution to 20% ABV “reduces alcohol burn, reduces sensory fatigue and improves the perception of volatile flavours” (Jack, J. Inst. Brew., 2003).
• The evidence backs the number. For the musty taint TCA, “dilution to 20% ABV is optimal for the detection,” with headspace concentration falling sharply between 30% and 40% ABV (Baxter et al., 2024).
• The panel outperforms the instrument. SWRI panellists detected TCA at an average of 4.40 ng/L in 20% ABV spirit, while conventional SPME-GC-MS limits of detection sit at “ca.10 ng/L” (Baxter et al., 2024).
• At 20% ABV the dilution water is half the sample. Published protocols specify distilled water (Baxter et al., 2024) or ultrapure 18.2 MΩ·cm water (Ashmore et al., Foods, 2023) — because chlorophenols are perceptible at around 0.1 µg/L (WHO), and a tainted diluent reads as a tainted cask.

Professional sensory panels dilute whisky to 20% ABV because ethanol masks the aromas they are trying to assess. At bottling strength, alcohol burn dominates the nose, fatigues the panellist after a handful of samples, and suppresses the volatile compounds that reveal both quality and fault. Cutting the sample to 20% ABV with purified water removes that interference, and the published evidence shows faults invisible at 40% ABV become detectable at 20%. This is analytical dilution, done to see clearly — a different exercise from the drop of water you add to a dram for pleasure.

What exactly is the 20% ABV nosing standard?

It is the working convention of the Scotch whisky industry, stated in the literature as established practice rather than proposed as a novelty. The 2024 guidelines paper from the Scotch Whisky Research Institute (SWRI) describes it as settled: “The standard industry practice for ‘nosing’ is to dilute samples of spirit to 20% ABV to prevent sensory fatigue” (Baxter, Réveillé & Conner, J. Inst. Brew. 130:112–123). It traces to guidelines published two decades earlier by Frances Jack, also at SWRI (J. Inst. Brew. 109:114–119, 2003), and it is not confined to one institute: a 2021 study of nine unmatured Scotch spirits records the same step — “In accordance with standard industry practice, the spirits were diluted to a uniform alcohol strength of 20% ABV using water” (Daute et al., Applied Sciences 11:1410).

Two words there carry the weight. Uniform, because a panel comparing a 63% ABV cask sample against a 40% ABV bottling is measuring strength, not character. Water, because at 20% ABV the diluent is the majority ingredient in the glass.

Why does ethanol hide the flavours a panel is looking for?

Through two separate mechanisms, one physical and one physiological.

The physical mechanism is partitioning. Aroma is perceived from the headspace above the liquid, and how readily a compound leaves the liquid depends on the ethanol around it. Molecular simulation of guaiacol — the smoky marker compound of peated whisky — found it “primarily found at the liquid-air interface in mixtures that contain up to 45 vol-% of ethanol,” while at “59 vol-% or higher, guaiacol is increasingly surrounded by ethanol molecules and is driven to the bulk” (Karlsson & Friedman, Scientific Reports 7:6489, 2017). Above a certain strength, the aroma compound is in the wrong place to be smelled.

The physiological mechanism is suppression at the receptor. A 2024 study of odorant detection in spirits found thresholds for all tested odorants rose with ethanol concentration, driven primarily by ethanol’s direct effect on olfactory perception rather than by volatility. Its authors state the consequence directly: the findings “support the practice of diluting ethanol content in distilled spirits by half (from 40% to 20% ABV) … during nosing” (Wang & Cadwallader, Beverages 10:116). A third, cruder factor sits on top: in the SWRI trials, “the perception of alcohol burn increased with the concentration of ethanol.” A panellist working through twenty samples at cask strength is nosing pain by the end of the flight.

Crucially, dilution does not simply thin the flavour. Work on rum tested this directly: samples cut with water to 20% ABV kept flavour profiles “very similar” to the originals, whereas samples cut with 40% ethanol solution — same dilution factor, no drop in strength — lost intensity across nearly all attributes (Ickes & Cadwallader, Food Science & Nutrition, 2018). Removing the ethanol, not the dilution itself, is what preserved the perception.

Does the evidence actually support 20%, rather than 30% or 40%?

For at least one commercially critical fault, yes — and the test was run against exactly those alternatives. The SWRI team spiked grain neutral spirit and whisky with 2,4,6-trichloroanisole (TCA), the compound behind musty, damp-cardboard off-notes, and presented it at four strengths chosen to mirror real checkpoints: 50% ABV (cask strength), 40% (standard bottling strength), 30% and 20%.

The sensory result was unambiguous: “Dilution of samples to 20% ABV made it easier for panellists to detect TCA,” and in single malt “the off-note being easier to detect at 20% ABV.” The instrumental result explained why: “Headspace TCA concentration diminishes as the ethanol concentration of the whisky increases with a sharp decrease observed between 30-40% ABV,” leading the authors to conclude that the GC-MS peak areas “clearly show that dilution to 20% ABV is optimal for the detection of TCA” (Baxter et al., 2024).

A fault a panel can find at 20% ABV can be sitting undetected in the same liquid at 40%. That is the commercial case for the protocol: the sample presented at bottling strength is the sample most likely to pass.

What does the panel catch that the instrument does not?

Earlier SWRI work with 217 panellists “found an average detection threshold of 4.40 ng/L TCA in 20% ABV neutral ethanol,” with individual thresholds spanning “0.25 to >10 ng/L.” Against whisky matrices, the SWRI expert panel found thresholds of “2 to 33 ng/L.” Meanwhile, using conventional solid phase microextraction with GC-MS, “instrumental limits of detection are ca.10 ng/L” (Baxter et al., 2024).

Read those figures together: the average trained nose is roughly twice as sensitive to this taint as the routine laboratory method, and the most sensitive panellists are an order of magnitude better. For this class of fault the panel is the analysis, and the 20% ABV dilution is the condition under which it performs. Sub-threshold TCA raises the stakes further, because it does not merely go unnoticed: very low levels “can disrupt olfactory signal transduction, suppressing odour detection.” A batch can be quietly dulled by a contaminant nobody consciously smells.

What does any of this have to do with the water?

Everything, because at 20% ABV roughly half the liquid in the nosing glass is the water you added.

Published protocols do not leave the diluent to chance. SWRI’s method states: “Unless otherwise stated, samples were diluted using distilled water on the day of testing” (Baxter et al., 2024). The 2023 Foods study used ultrapure water at 18.2 MΩ·cm (Ashmore et al.) — about 0.055 µS/cm, effectively the practical ceiling of purity. ISO maintains a dedicated standard, ISO 3696, specifying three grades of water for analytical laboratory use, precisely because analytical results are only as clean as the water behind them.

The logic is the one that governs reduction water in the bottling hall, with a sharper failure mode: a taint carried in by the diluent does not read as a bad diluent — it reads as a bad cask, and the panel cannot tell the difference. Chlorine is the clearest case. The compounds formed when chlorine meets phenols are perceptible at around 0.1 µg/L (WHO), and peated spirit is phenol-rich by definition, so a chlorinated tap-water diluent is a chlorophenol generator pointed at the samples you most need to read accurately. The connection is not incidental to TCA either: its precursor, 2,4,6-trichlorophenol, is described in the same SWRI paper as “a by-product of the chlorination of paper pulp.”

Set against a panel threshold of 4.40 ng/L — 4.4 parts per trillion — no ordinary drinking water is fit to be a sensory diluent, however pleasant it is to drink. One caution applies with particular force here: purified water is not sterile water, and stripping the chlorine residual removes what was suppressing growth. A carboy left standing beside the sensory booth is a plausible source of exactly the musty note the panel is convened to find — a trap we cover in pure is not sterile.

How is a nosing sample actually prepared?

The protocol is short, and every element is doing work.

Glassware and coding. “20 mL portions were presented in 100 mL blue nosing glasses and covered with watch-glasses” (Daute et al., 2021); SWRI’s TCA study used the same arrangement. The watch glass holds the headspace until assessment; blue glass — or the black Glencairn glasses of the Foods study (Ashmore et al., 2023) — hides colour, so a darker sample cannot be read as an older or better one. Samples carry “a 3-digit random code.”

Timing. Dilution is done “on the day of testing.” SWRI’s stability trial found nosing scores held across an average interval of four hours 56 minutes, while GC-MS showed “a marked decline in headspace TCA after three days.” Samples “are stable for the typical working day,” but should be “assessed promptly.”

Temperature. Normally ambient — but the same study found TCA “easier to detect … when nosed at 5°C compared to room temperature,” and recommends chilling “where panel agreement is difficult to achieve.”

The panel itself. SWRI’s expert panel of 22 was “pre-screened using an odour recognition test,” trained on the SWRI Scotch Whisky Flavour Wheel, and monitored through the FlavorActiv Whisky Sensory Proficiency Scheme (Baxter et al., 2024). Panels are built and audited like instruments, with an international framework alongside: ISO 8586 for assessors, ISO 8589 for test rooms, ISO 4120 for the triangle test.

Should you drink whisky at 20% ABV, then?

No — conflating the two is the most common misreading of this protocol. A panel dilutes to analyse, trading the integrated experience of a dram for sensitivity to individual compounds. The Foods study of consumer-style dilution found that beyond roughly 20% added water — far gentler than a panel’s cut — further dilution had “a deleterious effect on whisky aroma” (Ashmore et al., 2023). For a drinker the lesson is a principle, not a target strength, as our ratio guide by style sets out.

What should a QC laboratory specify for panel water?

Four lines, following directly from the numbers above.

1. Demineralised or distilled quality, not merely potable — the benchmark is laboratory water, not a drinking-water standard. Our demineralised water specification guide sets out the parameters.
2. Zero chlorine residual, justified by the 0.1 µg/L chlorophenol threshold rather than the taste threshold for chlorine itself.
3. A microbiological specification and a dispensing plan, on the understanding that purified water left standing does not stay pure.
4. Consistency across sessions and sites. If the diluent varies between batches, some of the variance the panel reports is yours, not the spirit’s.

That last point is the one most often missed. Sensory data is comparative by nature — this cask against last month’s, this site against another — so every uncontrolled variable in sample preparation becomes noise in a dataset that quality decisions rest on. The discipline that governs proofing on the production floor belongs in the sensory booth for the same reason: the water is supposed to be the thing that changes nothing.

Glossary

• Nosing — assessment of a spirit by aroma alone, without tasting. The standard mode of quality assessment in the Scotch whisky industry.
• Nosing strength / 20% ABV — the standard dilution for sensory samples, adopted to reduce alcohol burn and sensory fatigue and to improve perception of volatile flavours.
• Headspace — the air above a liquid in a covered glass, where volatiles accumulate and from which aroma is perceived. A watch glass is the cover that retains it until assessment.
• Trained / expert panel — assessors selected by screening and trained to a common vocabulary, with performance monitored by proficiency testing. Governed internationally by ISO 8586.
• TCA (2,4,6-trichloroanisole) — a contaminant giving musty, damp-cardboard off-notes, detectable by trained panellists at single-digit ng/L and linked to contaminated wood-based materials.
• Chlorophenol — a medicinal-smelling compound formed when chlorine reacts with phenols; perceptible at around 0.1 µg/L, and a precursor route to anisole taints.
• SPME-GC-MS — the routine instrumental method for volatiles; limits of detection around 10 ng/L for TCA.

Frequently asked questions

Why is whisky diluted to 20% ABV for tasting panels? Because ethanol interferes with the assessment. Diluting to 20% ABV “reduces alcohol burn, reduces sensory fatigue and improves the perception of volatile flavours” (Jack, 2003), and it is described in the current literature as “the standard industry practice for ‘nosing’” (Baxter et al., 2024).

Is 20% ABV better than 30% or 40% for detecting faults? For TCA taint, yes, and it was tested against exactly those strengths. Panellists found the off-note easier to detect at 20% ABV, and headspace analysis showed TCA availability falling sharply between 30% and 40% ABV, leading the authors to conclude that “dilution to 20% ABV is optimal for the detection of TCA” (Baxter et al., 2024).

What water do sensory laboratories use to dilute whisky samples? Purified laboratory water. Published protocols specify distilled water prepared on the day of testing (Baxter et al., 2024) or ultrapure 18.2 MΩ·cm water (Ashmore et al., 2023). At 20% ABV the diluent is around half the sample, so any taint it carries is attributed to the spirit.

Can a trained panel really out-detect a GC-MS? For TCA, yes. SWRI panellists averaged a detection threshold of 4.40 ng/L in 20% ABV spirit, with the most sensitive at 0.25 ng/L, while conventional SPME-GC-MS limits of detection are around 10 ng/L (Baxter et al., 2024).

Does dilution just make everything weaker? No. Rum cut with water to 20% ABV kept flavour profiles “very similar” to the original, while the same dilution performed with 40% ethanol solution lost intensity across nearly all attributes (Ickes & Cadwallader, 2018).

Blenders International · Water Science series