01
A narrow shell, with a wide interior
Dose and served mass clustered because of the rules. Inside that frame, competitors still built very different extraction and service systems.
WAC 2025 recipe analysis
There was no average winning recipe: what 25 WAC 2025 recipes reveal
The meta-finding from 25 available recipes is that the competition narrowed the frame more than the recipe. Dose and served mass clustered, and several workflow conventions emerged. Inside that frame, water allocation, preparation, pressing, and service still produced very different systems.
The sample covers 25 of 66 finalists. Only first, second, and third place are known. All competitors used the same coffee and could use no more than 18 g.
Key conclusions
02
There were real workflow norms
Twenty of 22 eligible recipes were inverted, and all 25 records used the Comandante grinder ecosystem. Those conventions did not produce one grind or method.
03
Water allocation carried the largest split
Brew-water ratios spanned 5.56:1–13.33:1, while exact bypass amounts spanned 0–88 g. This is where recipe structures separated most clearly.
04
Concentrated is not the same as under-extracted
These inputs are much more concentrated than a batch-brew starting ratio, but beverage TDS is missing. Extraction yield cannot be inferred from recipe ratios.
Two steps back
The competition converged on a format, not a recipe
The sample has a narrow outer frame and a wide interior. Rules and shared equipment created visible convergence; the variables competitors used to shape extraction and service remained much less settled.
Rules and served output
Dose middle 50%: 17.5–18 g · final mass middle 50%: 150–150.5 g
Strong convergence, largely imposed by the competition format.
Workflow conventions
Inverted: 20/22 eligible · Comandante ecosystem: 25/25
A shared equipment and setup culture, not one complete recipe.
Heat and press timing
Temperature: 85–91°C · press start: 89.75–105 s · press duration: 28.75–33.75 s
Soft centers appear when the middle 50% is considered, despite wider edges.
Water allocation
Brew-water ratio: 5.56:1–13.33:1 · exact bypass: 0–88 g
The clearest genuine spread in how competitors constructed extraction and service.
Temperature did not trend cooler
The 2025 sample is one event, so it cannot establish a temperature trend. Its median was 88°C and its middle 50% was 87–91°C: a soft center, not a cold-brew direction. The separate historical podium archive points the other way for recipes with one numeric brew temperature: period medians rose from 80°C (n=13), to 84.5°C (n=10), to 88°C (n=12).
Those historical periods contain different coffees, rules, and methods. The values describe the archive; they do not establish causation or an optimum.
See the separate historical archiveConcentrated recipes are not proof of under-extraction
The official AeroPress signature recipe is an upright, medium-fine brew using 16–18 g of coffee and 85°C water. The 2025 sample shares that dose scale and a broadly moderate temperature band, but it uses more varied water allocation, bypass, cooling, filtration, agitation, and service steps. Its 88°C median is slightly hotter than that AeroPress starting point, not cooler.
The SCA home-brewer Gold Cup procedure starts testing at 55 g/L, or about 18.2 g of water per gram of coffee. The 2025 sample's median brew-water ratio was 9.71:1, so its extraction stage was much more concentrated. The same SCA procedure specifies 92–96°C water at the coffee bed, above the 2025 recipe median, although those temperature measurements and brew methods are not directly equivalent. Gold Cup compliance is decided from measured beverage strength and extraction yield—not input ratio alone. Comparable beverage TDS is unavailable here, so the report cannot classify these cups as under-extracted or outside a “golden” zone.
All variables together
See how each recipe dials the complete system
A recipe is not one ratio or temperature. This profile connects all 13 comparable numeric variables so you can see where a competitor went higher, lower, or stayed near the sample center across the whole brew.
Parallel percentile profile
One recipe, thirteen connected decisions
Follow a recipe across every comparable numeric variable. Each point is its midrank percentile among that variable's primary eligible records, so the line reveals relative shape without pretending grams, seconds, ppm, and counts share one unit.
In focus
Némo Pop
Australia · podium #1
Highlight a recipe group25 highlighted
selected recipehighlighted grouppodium, always visiblesensitivity value
Scroll horizontally to follow the full profile →Selected point
Brew-water ratio
Recipe value
5.56:1
Sample median
9.86:1
Evidence
Primary · safe derived
Grams of brew water per gram of coffee, with bypass excluded. It describes recipe inputs, not measured extraction yield.
Read the shape, not a score
Higher means a larger numeric value, never better coffee. The shaded band is the middle 50%; M marks the median. Usable n changes by variable, missing values create gaps, and hollow rings keep sensitivity values separate.
Podium #1UprightUses bypassParticle prepFlow Control cap
| Variable | Value | Midrank percentile | Primary eligible records | Evidence |
|---|---|---|---|---|
| Coffee dose | 18 g | 64th percentile | 21 | Primary · explicit |
| Brew water | 100 g | 7th percentile | 22 | Primary · explicit |
| Brew-water ratio | 5.56:1 | 6th percentile | 18 | Primary · safe derived |
| Bypass amount | 70 g | 82nd percentile | 17 | Primary · explicit |
| Bypass share | 41.2% | 90th percentile | 15 | Primary · safe derived |
| Final beverage mass | 150 g | 37th percentile | 23 | Primary · competition rule default |
| Final-beverage ratio | 8.33:1 | 32nd percentile | 19 | Primary · competition rule default |
| Brew temperature | 84°C | 10th percentile | 21 | Primary · explicit |
| Water TDS | 125 ppm | 92nd percentile | 13 | Primary · explicit |
| Press start | 50 s | 3rd percentile | 20 | Primary · explicit |
| Press duration | 20 s | 13th percentile | 16 | Primary · explicit |
| Stir count | 8 | 68th percentile | 19 | Primary · safe derived |
| Swirl count | 0 | 28th percentile | 20 | Primary · explicit |
Full distributions
Variable atlas
These graphs show where competition practice converged and where it did not. They are the central result of the report.
● Primary value◇ Sensitivity value□ 2025 rule-default targetYellow diamond = podiumBand = middle 50%
Primary and sensitivity values stay visually distinct. Every chart states its own eligible n. Select a point to inspect the recipe.
Water structure
Dose and served mass were rule-constrained. Brew water, bypass, and their ratios reveal more of the recipe design.
Coffee dose
16 g–18 g · median 18 g · primary n=21
Dry coffee placed in the AeroPress chamber. 1 sensitivity value is also shown.
Lower doseMiddle 50%Higher dose
The 18 g competition maximum compressed dose. Most documented recipes sit at or close to the cap.
Brew water
100 g–240 g · median 170 g · primary n=22
Water documented as contacting the coffee during brewing. 2 sensitivity values are also shown.
Less contact waterMiddle 50%More contact water
The sample used a wide range of brew-water amounts.
Brew water excludes bypass and final beverage mass.
Brew-water ratio
5.56:1–13.33:1 · median 9.86:1 · primary n=18
Grams of brew water per gram of coffee, with bypass excluded. 3 sensitivity values are also shown.
Lower brew-water ratioMiddle 50%Higher brew-water ratio
This is the clearest numeric view of concentrate versus fuller-volume brewing.
Brew-water ratio describes recipe inputs, not measured extraction yield.
Bypass amount
0 g–88 g · median 30 g · primary n=17
Water added outside the main coffee-bed extraction phase. 5 sensitivity values are also shown.
Less bypassMiddle 50%More bypass
Bypass ranged from exact zero to large additions and occurred at different times.
Final beverage mass
150 g–205 g · median 150 g · primary n=23
The target mass of coffee served for judging. 1 sensitivity value is also shown.
Smaller served cupMiddle 50%Larger served cup
Most recipes sit at 150 g because that was the minimum beverage mass required for judging. This chart mainly provides service context.
Rule-default targets, explicitly stated targets, and sensitivity values use different marks.
Final-beverage ratio
8.33:1–11.39:1 · median 8.33:1 · primary n=19
Grams of final beverage per gram of coffee dose. 2 sensitivity values are also shown.
Lower served-cup ratioMiddle 50%Higher served-cup ratio
Served-cup ratios are much tighter than brew-water ratios, largely because the rules pulled final beverage mass toward the same minimum.
Final-beverage ratio describes served cup size, not extraction or beverage strength.
Water TDS
48 ppm–125 ppm · median 87 ppm · primary n=13
Total dissolved solids reported for the brew water.
Lower water TDSMiddle 50%Higher water TDS
Water TDS varied where stated, but coverage is limited.
Water TDS describes brew water. It is not beverage TDS.
Heat and time
Temperature, press start, and press duration stayed broad even though every competitor used the same coffee.
Brew temperature
82°C–95°C · median 88°C · primary n=21
The stated single-point temperature used for brewing.
Cooler brew waterMiddle 50%Hotter brew water
Primary single-temperature recipes span cool, middle, and hot editorial bands.
Press start
50 s–160 s · median 92.5 s · primary n=20
Time from the recipe start to the beginning of the press. 3 sensitivity values are also shown.
Earlier pressMiddle 50%Later press
Press start varied widely across the primary records.
Press duration
15 s–60 s · median 30 s · primary n=16
Time spent pressing the AeroPress. 2 sensitivity values are also shown.
Shorter pressMiddle 50%Longer press
Documented press durations range from quick to deliberately slow presses.
Movement and service
Stirring, swirling, cooling, orientation, and particle preparation describe different actions and remain separate.
Stir count
0–25 · median 5 · primary n=19
Count of documented stirring actions. 1 sensitivity value is also shown.
Fewer stirsMiddle 50%More stirs
Stir counts vary, but counts do not capture force or timing.
Swirl count
0–28 · median 0 · primary n=20
Count of documented swirling actions. 2 sensitivity values are also shown.
Fewer swirlsMiddle 50%More swirls
Many recipes state no swirl, while a few use repeated swirling.
Cooling
explicit n=22
Whether the written recipe documents cooling after brewing.
10
12
Cooling documented10/22
No explicit cooling12/22
One additional recipe has cooling inferred from transfer or aeration wording and remains separate from the explicit comparison.
Inverted orientation
Explicitly documented inverted brews.
90.9%3 missing or ineligible
Particle preparation
Sifting, chaff removal, or fines removal documented.
41.7%1 missing or ineligible
Grind and hardware
Grinder settings stay inside compatible systems. Hardware prevalence is descriptive, not evidence of an advantage.
Grinder model breakdown
The available sample uses the reported grinder-model labels without collapsing unresolved or mixed-model records into C40.
reported model n=25
Comandante C4036.0% · 9/25
Comandante C40 / Hammerhead28.0% · 7/25
C40, model unspecified20.0% · 5/25
Comandante, model unresolved12.0% · 3/25
Trailmaster x254.0% · 1/25
These are equipment records, not comparable grind-size or performance results.
Grind within compatible systems
Click settings are device settings, not direct particle-size measurements. No combined grinder scale is created.
Strict C40 comparator
exact n=13 · 12–37.5 clicks12 clicks37.5 clicks
Hammerhead comparator
exact n=6 · 22–52 clicks21 clicks52 clicks
The estimated split-grind midpoint is shown separately and is not part of the exact denominator.
Unresolved Comandante · n=3C40 model-specific, not pooled · n=1Trailmaster/Tigershark · n=1
The strict C40 and exact Hammerhead views show meaningful spread inside each compatible system. Unresolved and model-specific settings remain outside those comparisons.
Flow Control cap
A cap that prevents drip-through before pressing and allows pressure to build during the press.
13.0%2 missing or ineligible
Two of three podium recipes used a Flow Control cap, compared with 1 of 22 rank-unknown recipes. This small descriptive contrast does not show a placement advantage.
Filter setup
Filter and hardware configurations remain recipe-level records in Explore the data.
What connects the variables
Water architecture and bypass
Water architecture is the split between water used during brewing and water added outside that phase as bypass.
Contact water versus bypass
Exact paired values · n=13
Bypass share
High bypass
Low contact water
High contact water
5.56:1 brew-water ratioBrew water per gram of coffee13.33:1
Ring = podium□ Preloaded bypass● Post-brew bypass▲ No bypassFilled = cooling documented
Fourteen recipes have primary paired values for brew-water ratio and bypass share. Before-brew and after-brew bypass remain distinct.
The broad structures are easy to see. Some competitors made a strong concentrate and added a large bypass. Others used a moderate concentrate. A third group brewed with much more water inside the AeroPress and used little or no bypass. Source-specific and unresolved recipes remain separate.
Bypass amount is only part of the decision. Water could be placed in the server before brewing or added after pressing. It could dilute a concentrate, lower serving temperature, or support another service step. That is why “uses bypass” is too broad to describe one method.
No single broad water structure dominated the available sample. Brew-water ratio and bypass are more informative than served mass because the 150 g judging minimum pulled many final beverages toward the same target.
Three case studies
Three podium systems
The podium shows why an average recipe would be misleading.
| Variable | |||
|---|---|---|---|
| Dose | 18 g | 18 g | 16 g |
| Brew water | 100 g | 100 g | 208 g |
| Brew-water ratio | 5.56:1 | 5.56:1 | 13:1 |
| Bypass | 70 g | 86 g | 12 g |
| Temperature | 84°C | 88°C | 88°C |
| Press duration | 20 s | ≈44 s | 60 s |
- Dose
- 18 g
- Brew water
- 100 g
- Brew-water ratio
- 5.56:1
- Bypass
- 70 g
- Temperature
- 84°C
- Press duration
- 20 s
- Dose
- 18 g
- Brew water
- 100 g
- Brew-water ratio
- 5.56:1
- Bypass
- 86 g
- Temperature
- 88°C
- Press duration
- ≈44 s
- Dose
- 16 g
- Brew water
- 208 g
- Brew-water ratio
- 13:1
- Bypass
- 12 g
- Temperature
- 88°C
- Press duration
- 60 s
Approximate values are visually distinct. Jan Ahrend’s ≈44-second press is derived from the stated range.
First and second shared a concentrated 100 g brew-water structure, but differed in bypass timing, orientation, filters, particle preparation, cooling, and pressing. Third used a fundamentally different structure: 208 g of brew water and only 12 g of bypass.
Némo Pop’s recipe also separated from the other 22 in the strongest statistical split. That makes it a useful structural outlier, not an explanation for winning. With only three known placements, the report cannot connect any recipe feature to final rank.
Compare the three 2025 podium recipes side by side with exact, derived, rule-default, estimated, and ranged values kept visibly separate.
Supporting findings
Three additional patterns worth noticing
Temperature, cooling, and press documentation add important context to the water story.
Heat and timing had soft centers, with wide edges
The middle half of temperatures sat at 87–91°C, press starts at 89.5–102.5 seconds, and press durations at 27.5–32.5 seconds. Those centers are meaningful, even though the complete ranges still contain much earlier, later, cooler, and hotter recipes.
Cooling appeared as a meaningful branch
Cooling is the clearest documented difference between the two largest post-brew neighborhoods.
Cooled neighborhood7/7
recipes with documented cooling
Mostly non-cooled neighborhood1/8
recipes with documented cooling
Neighborhoods identify nearby recipes. They are not fixed recipe types, and this pattern does not show a placement advantage.
Pressing was under-described
What the written recipes leave out
All 25 written recipe records
12
6
3
Timed press; endpoint unstated
12/25 · 48.0%
Press under-specified
6/25 · 24.0%
Timed press + hiss endpoint
3/25 · 12.0%
Estimated timing + hiss endpoint
1/25 · 4.0%
Output- and rate-controlled press
1/25 · 4.0%
Stop-before-hiss; duration unstated
1/25 · 4.0%
Timed press + stop-before-hiss endpoint
1/25 · 4.0%
This measures what the written recipes document. It does not measure how precisely competitors executed their presses.
For competitors
What competitors can test
Use the distributions to choose a controlled experiment, not to copy one isolated setting.
Start with a water structure: concentrate plus bypass, a moderate concentrate, or a fuller-volume brew. Hold final beverage mass constant. Then change one main variable at a time, such as brew water, bypass timing, temperature, or press start.
Record stirring and swirling separately. Define press start, duration, rate, output, endpoint, and whether the press reaches the hiss. Include cooling and service in the recipe. Those details make the next cup easier to interpret.
Important limits
Limitations
This is a descriptive report about the available recipes, not a model of competition success.
The sample covers 25 of 66 finalists. Only first, second, and third place are known. The missing 41 recipes may contain patterns that do not appear here, and the 22 rank-unknown records cannot be ordered by performance.
Usable counts vary by variable. Missing, estimated, and conflicting values remain separate. Written gaps may reflect incomplete records, not what happened on stage. Water TDS is not beverage TDS, so the report does not calculate extraction yield. No chart identifies a cause of podium placement.
Conclusion
The shared format created islands of convergence, not one formula
Dose, served mass, orientation, grinder ecosystem, temperature, and press timing each show some clustering. Water allocation and finish-stage decisions remain the clearest sources of structural difference.
The high-level finding is not that every variable was random, or that competitors converged on one recipe. They shared a competition grammar, then used water, bypass, particles, agitation, pressing, cooling, and service to write different brews inside it. Compare those complete systems, then test one decision at a time.
Related reading