[ Data Exploration ]

The Volume Gap

How much should you trust surfboard volume calculators?

Surfboard volume calculators promise a simple answer: enter a few details and receive the number of liters you should ride. But different recommendation systems use different assumptions, variables, and definitions. This project compares how they behave and makes their disagreement visible.

One surfer, many answers

Enter a profile and see what every system in this comparison recommends for it, including where a system's answer is directly observed, interpolated between collected data points, or a community heuristic rather than a branded calculator.

Age, fitness, and frequency only change a system's answer where that exact variation was actually collected for it. We don't invent adjustments a system never published (see Methodology).

Not all calculators work the same way

Some systems ask for two inputs and look up a table. Others ask for ten and run an undisclosed formula. Neither approach is inherently better, but they're not answering exactly the same question, and it's worth seeing which is which before comparing their numbers.

SystemWeightHeightAgeGenderFitnessFrequencySkillWave typeWave sizeWetsuitInputsOutput
Firewire2point liters
Surf Simply2range liters
Roberts Surfboards2range liters
Lost Surfboards2range liters
Rusty Surfboards6range liters
JS Industries9point liters
Guild Factor heuristic2range liters
Weight-based rule of thumb2point liters
NSP3point liters
Matta Surfboards2point liters
Surftech2point liters
Surfboard Empire5range liters
Pyzel3point liters

Methodology families

Systems in the same family share a conceptual foundation, not necessarily an organization. Treat calculators in the same family as more correlated evidence, not fully independent opinions.

Weight-and-skill tables

A published lookup table indexed by body weight and skill level. No formula to reverse-engineer, just pre-computed values the brand publishes directly.

  • Firewire
  • Surftech
  • Pyzel

3 systems share this methodology, not fully independent

Volume-to-weight ratio reference

A published minimum-functional ratio framework rather than a single recommended number. It describes the smallest board that is still practical to ride, not the most comfortable one, which is why it is excluded from the consensus.

  • Surf Simply

Guild Factor systems

A ratio of board volume to body weight ("Guild Factor" = liters ÷ kg), popularized by pro surfer Whitney Guild. The surfer picks a ratio band based on skill and fitness, then multiplies it by their weight.

  • Roberts Surfboards
  • Lost Surfboards
  • Guild Factor heuristic

3 systems share this methodology, not fully independent

Rich profile calculators

A multi-input formula that adjusts a weight-based baseline using factors like age, fitness, surf frequency, and wave conditions, rather than reading straight off a table.

  • Rusty Surfboards
  • JS Industries

2 systems share this methodology, not fully independent

Weight-ratio folk heuristic

A simple weight-times-coefficient rule repeated across surf-education content, without one clearly identifiable original source.

  • Weight-based rule of thumb

Board-type-weighted ratio

A weight-based ratio multiplied again by a board-type factor (shortboard, fish, midlength, longboard). The only family here where board type is an input to the volume, not just an output of it.

  • NSP

Additive skill offset

Adds or subtracts a flat number of liters by skill level on top of a weight-based baseline, rather than multiplying by a ratio. A structurally different kind of formula from every ratio-based family above.

  • Matta Surfboards

Weight/skill table with multipliers

Its own published lookup table indexed jointly by weight and skill (not just weight), then further scaled by age and fitness multipliers on top. A hybrid of the table-lookup and multi-factor approaches above.

  • Surfboard Empire

How recommendations change with weight and skill

Each line traces one system's recommendation across the core weight range for a given skill level, against the dashed cross-system consensus curve. Toggle sources on or off to isolate one at a time.

Sensitivity explorer

Pick a system and vary one input at a time around a starting profile, holding everything else fixed.

Which systems tend to recommend more volume?

Each cell shows how far a system's recommendation sits above or below the cross-system median for that specific profile, a relative tendency, not an accuracy or bias measurement. Systems marked not directly comparable (see Methodology) are left out here rather than compared against a median they were never part of.

Where do calculators agree?

The shaded band is the middle 50% of systems around the consensus median at each weight, faceted by skill. Disagreement here is useful uncertainty to know about, not a flaw in the comparison.

Key findings

Recommendation-system profiles

These scores characterize how each system behaves, not how good it is. A high complexity score isn't a flaw and a high transparency score isn't a guarantee of accuracy. Where a score reads "not measurable," that means we don't yet have enough data to compute it responsibly, never a hidden zero.

Firewire

Complexity0.00
Transparency0.50
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+12%

Surf Simply

Complexity0.00
Transparency1.00
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendencynot measurable

Roberts Surfboards

Complexity0.00
Transparency1.00
Coverage0.30
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency-4%

Lost Surfboards

Complexity0.00
Transparency1.00
Coverage0.30
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency-10%

Rusty Surfboards

Complexity0.57
Transparency0.00
Coverage0.60
Consistency1.00
Sensitivity0.02
Relative-volume tendency+2%

JS Industries

Complexity1.00
Transparency1.00
Coverage0.90
Consistency1.00
Sensitivity0.02
Relative-volume tendency-5%

Guild Factor heuristic

Complexity0.00
Transparency1.00
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency-6%

Weight-based rule of thumb

Complexity0.00
Transparency0.75
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+10%

NSP

Complexity0.14
Transparency0.50
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+1%

Matta Surfboards

Complexity0.00
Transparency0.50
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency-8%

Surftech

Complexity0.00
Transparency0.50
Coverage0.20
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+13%

Surfboard Empire

Complexity0.43
Transparency0.50
Coverage0.40
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+6%

Pyzel

Complexity0.14
Transparency1.00
Coverage0.30
Consistency1.00
Sensitivitynot measurable
Relative-volume tendency+4%

The consensus model

The consensus shown throughout this project is the median recommended volume across directly comparable systems for a given profile, not an average, and not any single system's answer.

Alongside it we report the middle 50% of systems (the interquartile range) and the full observed range, so the consensus is never presented without its own uncertainty.

This consensus is not claimed to be correct, ideal, optimal, or scientifically validated. It's a transparent summary of what the systems we sampled currently say, nothing more.

Secondary layer: explanatory regression

An interpretable model (weight, skill, their interaction, and per-source fixed effects) fit across the core comparison grid: an explanatory lens on the data, not the consensus calculator itself.

0.908
CV RMSE
4.14 L
Observations
252
TermEstimate95% CI
intercept51.27[47.79, 54.70]
weight_kg_centered0.53[0.47, 0.60]
skill_intermediate-11.53[-12.84, -10.29]
skill_advanced-16.99[-18.50, -15.59]
weight_kg_centered:skill_intermediate-0.11[-0.19, -0.04]
weight_kg_centered:skill_advanced-0.17[-0.25, -0.09]
source_guild_factor_heuristic-7.20[-10.03, -4.39]
source_js_industries-7.66[-11.07, -4.23]
source_lost-9.11[-12.60, -6.01]
source_matta-7.77[-10.68, -4.93]
source_nsp-5.02[-8.11, -2.11]
source_pyzel-3.77[-6.81, -1.03]
source_roberts-6.93[-9.90, -3.96]
source_rusty-4.56[-7.58, -1.61]
source_surfboard_empire-2.66[-5.69, 0.29]
source_surftech0.37[-3.15, 3.75]
source_weight_rule_heuristic-1.21[-4.14, 1.91]

What liters leave out

A single liter figure is a useful starting point, not a full board prescription. None of the systems compared here account for the factors below, and each one changes how a board actually rides independent of volume.

Length
Longer boards paddle into waves earlier and catch more of them; shorter boards turn and accelerate faster. Two boards can carry the same volume at very different lengths by trading length for width or thickness.
Width
Wider boards are more stable and plane sooner, which is why beginner boards run wide. Width also slows down rail-to-rail transitions, so the same volume in a narrower board turns quicker at the cost of some stability.
Outline
The curve of the edge from nose to tail. A fuller, rounder outline paddles easily and holds speed through flat sections; a curvier outline turns tighter but demands more precise footwork.
Rocker
The curve from nose to tail measured through the bottom of the board. Flatter rocker paddles fast and planes early, good for weak waves, but catches on steep faces; more rocker suits powerful waves but paddles slower.
Rails
The edges of the board. Thick, soft rails float more and forgive mistakes; thin, hard rails bite into the wave face for sharper turns but punish poor timing.
Foil
How thickness is distributed from nose to tail. Two boards can share a volume number while carrying it in very different places, changing how the board pivots and how forgiving the nose feels.
Volume distribution
Related to foil: where the volume sits, not just how much of it there is, shapes how a board paddles and turns. A board that's volume-heavy under the chest behaves differently than one that's volume-heavy underfoot, even at the same total liters.
Construction
PU/polyester, epoxy, and foam constructions flex, float, and feel different in the water even at identical volume. Epoxy boards in particular often paddle and float livelier than PU at the same liters.
Board category
A 32-liter shortboard and a 32-liter fish express that volume through completely different shapes, and surf completely differently, despite sharing a number.
Wave type
A board tuned for soft, mushy waves handles differently than one tuned for punchy, powerful surf, independent of whether its volume matches a calculator's recommendation.
Surfer goals
Someone chasing performance and responsiveness might deliberately ride less volume than recommended; someone prioritizing wave count might want more. "Recommended" assumes a goal the calculator never actually asked about.
Technique
Paddling technique, positioning, and timing can compensate for, or amplify, the effects of volume. Two surfers with identical stats can need different volumes depending on how well they execute in the water.
Paddling ability
Upper-body strength and paddling endurance affect how much volume a surfer actually needs to catch waves comfortably, somewhat independent of the "skill level" category a calculator asks for.

Two boards with the same volume can paddle and surf very differently.

Methodology

Sources & collection

Each system is collected via the method most faithful to how it actually works: published tables and ratios are extracted directly, interactive calculators are automated with cached, dated captures, and hard-to-automate sources fall back to documented manual entry. Every observation records its collection method and date.

Sampling design

A core comparison grid of 7 weights (120–240 lb) × 3 skill levels (21 profiles) is collected from every source, plus 6 named archetype profiles and one-factor-at-a-time variations for systems that support richer inputs. This is a deliberately staged design, not a full Cartesian product, so richer calculators don't dominate the comparison.

Normalization

Skill, fitness, frequency, conditions, and wetsuit inputs are mapped to a small shared taxonomy. Every mapping from a source's original label to the normalized category is documented with a rationale and a confidence level. Ambiguous mappings are never hidden.

Interpolation policy

Missing weights are linearly interpolated only within the same source and skill level, and only between two actually-collected weights, with no extrapolation beyond the collected range and no interpolation across skill levels or unsupported inputs like age or fitness.

Comparability rules

A system is excluded from the consensus for a given profile when it returns only a board model, assumes a materially different board category, can't be aligned to the requested inputs responsibly, or its recommendation can't be translated into liters. The number of contributing systems is always shown alongside any consensus figure.

Consensus definition

The consensus shown throughout is the median recommended volume across directly comparable systems, reported with the interquartile range, full range, and median absolute deviation. Agreement labels (high/moderate/low) are a presentation convention based on relative IQR thresholds, not a scientific claim.

Score definitions

Complexity (input count, min-max normalized), transparency (a 4-factor rubric: published methodology, published formula, stated assumptions, reproducible outputs), sensitivity (mean % change across collected one-factor variations), relative-volume tendency (mean deviation from the cross-system median), consistency (monotonicity across adjacent weights, with outsized jumps flagged but never auto-labeled errors), and coverage (fraction of supported input types). All descriptive, none a quality ranking.

Heuristic provenance

Community heuristics are included only once their exact formula, source, intended surfer type, and limitations are documented (see the Sources table below for citations).

Limitations

Reflection & next steps

What was surprising
How much the brief's guesses about complexity were wrong in both directions. Lost was expected to be a rich multi-input calculator but turned out to be the same simple Guild Factor system as Roberts, right down to sharing an explicit "Whitney Guild 2013" credit on the page. JS Industries was expected to be complex and opaque but shipped a fully commented formula in plain client-side JavaScript, more transparent in practice than several sources that only publish a black-box table.
Easiest vs. hardest to compare
Easiest: the four sources with a published table, ratio, or formula (Firewire, Surf Simply, Roberts, Pyzel), no automation needed, just careful reading. Hardest, in a different sense than expected: not Rusty's genuine multi-input form (Playwright handled it cleanly once the hidden-select quirk was found), but resolving ambiguous unit and skill-tier mappings. For example, figuring out whether Roberts' "Guild Factor" meant liters-per-pound or liters-per-kilogram required cross-checking the resulting volumes against Firewire's independently-collected real numbers.
What real surfer outcomes would take
Nothing here is validated against how a real board with that volume actually paddled or performed for a real surfer. A genuinely validated version would need outcome data: surfers trying boards at the consensus volume (and above/below it) and reporting back on wave count, paddling effort, and how it felt, not just what thirteen calculators say before a single wave is caught.

Next highest-value improvements

  1. 1.Resolve DHD's calculator: it has a genuinely rich 5-input form (age, height, weight, 5-tier experience, 4-tier fitness) with dual soft-wave/powerful-wave outputs, but its formula lives behind a code path this pass couldn't locate in the site's shipped JS. Worth a targeted network-traffic capture rather than static source-reading. SMTH Shapes (a distinct "Shape3D BoardAdvisor" third-party engine) is a similar unresolved case.
  2. 2.Revisit Boardcave and Chemistry Surfboards specifically to check whether their Cavewire tuning diverges meaningfully from Rusty's, now that the shared-engine finding is documented rather than assumed.
  3. 3.Add outcome data (real surfers, real boards, real feedback) as a second, independent axis to compare against the calculator consensus. That's the single biggest gap this project can't close with public calculator pages alone.

Sources

SourceCollectedMethod
Firewire2026-08-05Published table
Surf Simply2026-08-05Derived from published ratio
Roberts Surfboards2026-08-05Derived from published formula
Lost Surfboards2026-08-05Derived from published formula
Rusty Surfboards2026-08-06Automated collection
JS Industries2026-08-06Derived from published formula
Guild Factor heuristic2026-08-06Derived from published ratio
Weight-based rule of thumb2026-08-06Community heuristic
NSP2026-08-06Derived from published formula
Matta Surfboards2026-08-06Derived from published formula
Surftech2026-08-06Published table
Surfboard Empire2026-08-06Derived from published formula
Pyzel2026-08-05Derived from published formula