Why Boats Roll — A Comfort Primer for Solar Yacht Design

1 · Your four claims, examined & quantified

Your intuitions are substantially correct — but each one runs on a different piece of physics, and two of them are missing a mechanism that matters more than the one named. Here is the honest scorecard, with the key numbers.

Claim	Verdict	What's right, what's missing, and the number that settles it
Sailboats roll less because of the sail pushing on the air.	Mostly true	The sail doesn't mainly push — it damps. Rolling changes the sails' angle of attack and the apparent wind, producing a force that opposes the roll velocity. Worked estimate below: aerodynamic damping adds roughly ζ ≈ 0.05–0.15 of damping ratio (bare hull ≈ 0.03–0.08). Second effect: the boat sails at a steady heel where the righting-moment curve is steeper, effectively raising stiffness by ×1.1–1.4. Net result near resonance: roughly half the roll angle of the same hull motorsailing bare (28° → 14° in the worked example of §4).
Powerboats roll less because they go fast and average out many waves; on average the ocean is level.	Partly true	The "average is level" idea is a red herring — the mean sea surface is always level; what matters is the spectral energy near your roll frequency. The real mechanisms: (a) Doppler detuning — in head seas the encounter frequency rises (a 6 s wave met at 8 kn arrives every 4.2 s; at 20 kn every 2.9 s), moving excitation away from a typical 4–6 s roll period; (b) faster hulls carry more hydrodynamic damping; (c) most importantly, fast powerboats run active fin stabilizers, whose lift scales with U² and simply doesn't work slowly. Caveat: roll is excited mostly by beam seas, where speed changes almost nothing — which is why fast boats still roll in beam swell without stabilizers.
Trawlers are moving fast enough that they can have stabilizers and be very stable.	True, with nuance	Fins become effective above roughly 7–8 kn (lift ∝ U²), and full-displacement trawlers cruise at 7–9 kn — just enough. But note the design philosophy: trawlers are deliberately tender (low GM, long gentle roll period, high inertia) and then buy back amplitude with bilge keels + fins or paravanes. A stabilized trawler stacks damping ratios to ζ ≈ 0.2–0.3, holding typical rolls to ~3–6° (§4). "Very stable" here means controlled motion, not high initial stability.
A slow, sail-less solar boat is noticeably less comfortable than either.	Generally true …but fixable	It loses all three advantages at once: no sail damping (−0.05–0.15 ζ), no heel-stiffening, and too slow to detune waves or power fins. The power budget kills gyros outright: a gyro suited to a 12–15 m boat draws 2.5–4.5 kW continuously ≈ 60–90 kWh/day, versus a realistic solar harvest of 25–35 kWh/day. In steep short chop the worked example puts an unstabilized solar monohull at ~28° rolls versus ~6° for a planing powerboat. Yet the gap is closable: bilge keels (0 W), an anti-roll tank (0 W), paravanes (~0.2–0.6 kW), and smart period-tuning recover most of it (§6).

Common confusion "Stability" ≠ "comfort." Stability (statically) is resistance to capsize — GM, righting arms, vanishing stability angle. Comfort is motion: roll angle, roll acceleration, vertical acceleration. A very stable boat (big GM) can be un-comfortable because it is stiff and snappy; a tender boat can be lovely if well damped. Everything on this page is about the second kind, though the two share vocabulary.

2 · Crash course: the physics of rolling

2.1 The motions that matter

A boat moves in six degrees of freedom. Comfort at cruising speeds is dominated by three:

Roll (φ) — rotation about the fore-aft axis. The big one for discomfort: it tilts the horizon, swings lamps, throws galleyware.
Pitch (θ) — rotation about the athwartships axis; matters in head seas.
Heave (z) — vertical translation; its acceleration is what triggers seasickness.

This page focuses on roll, because that is where sails, speed, and stabilizers make their most visible difference.

2.2 Righting moment, GZ, and the metacentric concept

Heel a boat and buoyancy shifts; gravity does not. The horizontal offset between the buoyant upward force and the weight downward force is the righting arm, GZ. Multiply by displacement weight and you get the righting moment:

Righting moment RM(φ) = Δ · GZ(φ)

where Δ = displacement force = m·g (newtons). Plot GZ against heel and you get the famous GZ curve: rising from zero, peaking somewhere around 35–60° for a monohull, falling to zero at the vanishing-stability angle.

For small angles the curve is nearly a straight line through the origin. Its slope is governed by the metacentric height GM:

Initial stability GM = KB + BM − KG, BM = I_WP / ∇

and for small heel, GZ ≈ GM·sin φ ≈ GM·φ. So:

KB — height of the centre of buoyancy (roughly half the canoe-body draft).
BM — the metacentric radius: how fast the centre of buoyancy slides sideways as you heel. Set entirely by the waterplane: I_WP is the waterplane's second moment of area about the centreline, ∇ the displaced volume.
KG — height of the centre of gravity. Every kilogram up top (flybridge, antennas, dinghy on davits) pushes KG up and GM down.

GM is the roll spring stiffness. Double GM → the roll restoring torque doubles → the boat is "stiffer," rolls faster, and (all else equal) rolls to smaller angles in a given wave but with harder accelerations.

Worked example: where a catamaran's enormous GM comes from (click)

Take a 12 m solar catamaran, 8 t displacement, hull centres 4.2 m apart, each slender hull with waterplane area A_h ≈ 5 m², own waterplane inertia ≈ 0.3 m⁴, hull draft 0.7 m, KG ≈ 1.5 m.

I_WP = 2·[ i_hull + A_h·(s/2)² ] = 2·[0.3 + 5×(2.1)²] ≈ 45 m⁴
BM = I_WP/∇ = 45/8 ≈ 5.6 m
GM = KB + BM − KG = 0.4 + 5.6 − 1.5 ≈ 4.5 m

Compare a 12 m monohull: I_WP ≈ 6–9 m⁴, ∇ ≈ 9 m³ → BM ≈ 0.8 m, GM ≈ 0.9–1.2 m. The catamaran buys stiffness purely by spacing waterplane area away from the centreline — the (s/2)² term is brutal. Trade-off: form stability saturates by ~8–12° of heel (one hull's waterplane starts leaving the water), and the huge GM makes the roll period very short — "snappy."

2.3 Centre of gravity, loading, and free surfaces

KG is a live number. Fuel and water burn off, stores move, crew sit topsides. A 3% rise in KG can cut GM 10–15% on a fin-keel monohull.
Free-surface effect: liquid sloshing in a partially filled tank shifts its CG as the boat heels, acting like a virtual rise in G:

Free-surface GM loss ΔGM_fs = Σ (ρ_liquid/ρ_sea) · i_tank-surface / ∇

A wide, shallow tank is far worse than a deep, narrow one. Baffles help; keeping tanks full or empty helps most. On a solar yacht with big battery mass, battery placement is your best ballast tool: low and outboard (in the hulls of a cat) lowers KG and spreads mass from the roll axis, both of which lengthen and calm the roll.

2.4 Roll inertia, radius of gyration, added mass

Stiffness alone doesn't set the motion; inertia is the other half of the pendulum. Roll response is governed by the total mass moment of inertia about the roll axis:

Roll inertia I_tot = (1 + j) · m · k², j ≈ 0.15–0.30 (added inertia of entrained water)

where k, the radius of gyration, is the radius at which the boat's mass could be concentrated to give the same inertia. Practical ranges:

Monohull: k ≈ 0.35–0.45 × B (rig and topweight push it up; low ballast pulls it down).
Catamaran: k ≈ 0.35–0.50 × BOA.

Heavier-for-their-size boats (high k, high m) roll slower and more sluggishly — one reason old heavy cruisers feel "steady" even though they roll to large angles.

2.5 Natural roll period — the single most useful number

Natural roll period T_n = 2π √( I_tot / (Δ · GM) ) ≈ 2π k / √( g · GM )

(The shortcut folds added mass into k.) Typical values:

Vessel	T_n (s)	Comment
Dinghy / small dayboat	1–2	Too quick to synchronize with anything but tiny ripple
40 ft catamaran	2.3–3.0	Huge GM, short period — "twitchy" in chop
30–40 ft monohull yacht	3.3–4.5	Sits squarely inside wind-sea periods — the danger zone
45–55 ft trawler / heavy cruiser	4.5–7	Tender by design; relies on stabilizers for amplitude
Large ship	15–30	Far above ordinary sea periods; rolls only in long swell

Key idea Resonance happens when the sea's period equals T_n. Ocean wind waves live at 3–8 s and swell at 8–16 s. A 3.5–5 s boat cannot avoid the wind-sea band entirely — which is why damping, not period-tuning alone, is the solar yacht's best friend.

2.6 Damping — what actually bleeds the energy out

A pendulum in vacuum swings forever; a boat doesn't, because of several energy sinks. Naval architects lump them into an equivalent damping ratio ζ (fraction of critical damping):

Source	Mechanism	Typical ζ contribution	Works at zero speed?
Wave radiation	Rolling radiates waves; energy carried away	0.02–0.05	Yes
Viscous hull / eddies	Skin friction, separation at bilges	0.01–0.04	Yes
Bilge keels / fins (passive)	Eddy shedding + lift from roll-relative flow	+0.03–0.08	Yes
Sails & rig	Aerodynamic force opposing roll velocity	+0.05–0.15	No (needs wind + sails set)
Active fins (driven)	Powered lift opposing roll	+0.10–0.25	No (needs U ≳ 7–8 kn)
Gyro stabilizer	Precession torque opposing roll	+0.10–0.25	Yes, but kW-scale power
Anti-roll (flume/U-tube) tank	Tuned water mass lagging the roll	+0.04–0.10	Yes
Paravanes / flopperstoppers	Towed birds resisting roll velocity	+0.04–0.10	Needs way on (any speed)

Bare monohulls total ζ ≈ 0.03–0.08; a well-appendaged boat reaches 0.10–0.15 passively, 0.2–0.3 actively. Skin friction is a minor player — eddy-making and wave radiation dominate, which is why blunt appendages near the bilge work so well.

How to measure your own boat's T_n and ζ — the roll-decay test (click)

Calm water, no traffic. Gather crew on one side or winch the boom to heel the boat 5–10°.
Release cleanly. Time ten full periods with a stopwatch; divide by ten → T_n. Repeat three times.
Record successive extreme angles a₁, a₂, a₃… The logarithmic decrement δ = ln(a₁/a₂); for light damping ζ ≈ δ / 2π.
Repeat with sails set (if applicable) and with bilge keels/paravanes deployed to see what each device really adds.

A smartphone strapped below decks logging accelerometer + gyro data does the same job and gives you spectra you can compare against buoy wave data.

2.7 Resonance — the amplifier nobody installed

Drive a damped pendulum with a sinusoid and the response amplitude is the excitation times a magnification factor:

Roll response to wave slope φ_a = M · γ · ε_a, M(Λ, ζ) = 1 / √[ (1 − Λ²)² + (2ζΛ)² ], Λ = ω_e/ω_n = T_n/T_e

where ε_a is the wave-slope amplitude, γ ≈ 0.6–0.9 an "effective wave slope" coefficient accounting for hull shape and diffraction, and Λ the tuning ratio. At resonance (Λ = 1):

Magnification at resonance M ≈ 1 / (2ζ)

So ζ = 0.05 → ×10; ζ = 0.10 → ×5; ζ = 0.20 → ×2.5. Damping is the only thing standing between you and a ten-fold amplifier.

Figure 1 — The resonance amplifier. Away from resonance every boat responds about 1:1 to wave slope. Near Λ = 1 the response is amplified by ≈ 1/(2ζ). Moving ζ from 0.05 to 0.15 cuts peak roll by a factor of three — the entire business case for bilge keels, sails, fins, and tanks.

2.8 Waves 101 — period, length, steepness

Deep-water waves λ = 1.56 · T² (m), c = λ/T = 1.56 · T (m/s), slope amplitude ε = π·(H/2)/λ

A 6 s wave is 56 m long and travels at 9.4 m/s. Its maximum surface slope for H_s = 1.5 m is π×0.75/56 ≈ 0.042 rad ≈ 2.4°. That tiny tilt, multiplied by the resonance amplifier, is what rolls you.

Figure 2 — Period placement. Every small-boat roll period overlaps at least one sea band; the art is choosing which band you overlap (matching your home waters) and buying damping for the overlap you can't avoid. A 40 ft solar catamaran (2.6 s) overlaps only short chop; a 40 ft solar monohull (3.4 s) sits inside both chop and wind sea.

2.9 Encounter frequency — what speed really does

You don't meet waves at their own period; you meet them at the encounter period. In deep water the Doppler shift is beautifully simple:

Encounter period T_e = T_w / ( 1 + f · U/c )

with f = +1 head seas, +0.707 bow-quarter, 0 beam, −0.707 stern-quarter, −1 following. Worked table for a 6 s wave (c = 9.4 m/s):

Boat speed	Head seas T_e	Bow quarter	Beam	Following
0 kn	6.0 s	6.0 s	6.0 s	6.0 s
4 kn	4.9 s	5.4 s	6.0 s	7.7 s
8 kn	4.2 s	4.9 s	6.0 s	10.7 s
20 kn	2.9 s	3.6 s	6.0 s	>50 s (overtaking)

Two lessons: (1) speed is a powerful detuner in head seas — a 20 kn boat turns a 6 s roller into a 2.9 s flutter its inertia ignores. (2) in beam seas — the strongest roll excitation — speed does nothing. That asymmetry is why the "fast boats average the waves" story is only half-true, and why slow boats suffer most in beam and quartering seas.

Slow-boat hazards in following seas When U approaches wave celerity from astern, T_e stretches enormously: the boat sits on one wave face for ages. Two failure modes to respect: surfing/broaching and parametric rolling, which ignites when waves are encountered at about twice the roll frequency (T_e ≈ ½T_n) so that the waterplane — and thus GM — pumps in sync with the roll. Remedy: slow down, change course, move the rendezvous off the magic ratio.

2.10 The proper bookkeeping: spectra and RAOs

Real seas are sums of hundreds of waves. The modern workflow:

Describe the sea by a wave spectrum S_η(ω) (Pierson–Moskovitz / JONSWAP family), characterized by significant height H_s and peak period T_p.
Compute the boat's response amplitude operator |RAO(ω)|² — essentially the M of §2.7 across all frequencies.
Multiply and integrate: the roll variance is m₀ = ∫ |RAO|² · S_η(ω_e) dω, and the significant roll amplitude is φ₁/₃ = 4√m₀ ≈ 2×RMS.

The punchline for designers: only the spectral energy landing near ω_n matters. The ocean's average level is irrelevant; its behaviour in a ±15% band around your roll frequency is your whole comfort forecast. This is also the correct version of "averaging lots of waves": at high encounter frequency the RAO is small, so each wave contributes little — the integrand, not the wave count, does the work.

2.11 Quantifying "comfort"

Comfort is measurable. Four families of metrics:

(a) Motion limits (what crews tolerate)

Metric (RMS)	Typical comfort limit	Source / note
Vertical acceleration	≈ 0.15–0.28 m/s²	NORDFORSK 1987 criteria (widely cited)
Lateral acceleration	≈ 0.18 m/s²	NORDFORSK
Roll angle	≈ 3.0°	NORDFORSK (working/comfort)
Pitch angle	≈ 1.5°	NORDFORSK
Weighted whole-body accel.	<0.315 "not uncomfortable" … >1.6 "very uncomfortable" m/s²	ISO 2631-1 bands

Handy conversion: for narrow-band rolling, significant amplitude ≈ 2 × RMS. So the NORDFORSK 3° RMS roll limit corresponds to about 6° significant amplitude — a useful sanity check on the numbers in §4.

(b) Seasickness band

Motion sickness from vertical oscillation peaks sharply around 0.167 Hz — a 6-second period (O'Hanlon & McCauley). Annoyingly, 6 s is also one of the most common ocean periods. Roll periods of 3–8 s sit right in the nausea band 0.1–0.5 Hz; another reason small slow boats are punishing.

(c) Perception thresholds (approximate)

Visible roll: from ~0.5–1°; distinctly felt: ~2°; annoying sustained: 4°+.
Lateral acceleration at the head: noticeable from ~0.1 m/s².
Angular acceleration becomes unpleasant when sustained above a few °/s².

(d) Size-scaling indices: Brewer Comfort Ratio, DLR, and RAD

Because big boats meet proportionally smaller waves, comfort grows with size. Popular one-number indices encode this:

Displacement–Length Ratio: DLR = D_lt / (0.01·LWL_ft)³. Light <200, moderate 200–300, heavy 300–400, very heavy >400. Higher DLR → slower, lazier motion.
Ted Brewer's Comfort Ratio — same spirit (displacement weighted heavily against length cubed); consult current published tables for the exact constant and ratings, as versions circulate.
RAD — "Ratio of Acceleration to Displacement" (Steve Dashew): compares motion-induced acceleration against displacement^(1/3), so vessels of different sizes can be ranked; lower = gentler. Conventions vary between sources, so use it comparatively (same sea state, same computation) rather than as an absolute target. For design targets, prefer the RMS criteria in table (a).

Micro-example: lateral acceleration at the deck edge (click)

Total sideways acceleration felt at height h above the roll axis, at the deck edge:

a_lat ≈ g·sin φ + (B/2)·φ̈, with φ̈ ≈ φ·ω_e²

Take the 3.5 s chop scenario of §4 (ω_e = 1.80 rad/s):

Solar monohull at 28° (0.49 rad): a = 9.81×0.47 + 1.9×0.49×3.22 ≈ 4.6 + 3.0 = 7.6 m/s² ≈ 0.78 g — gear-breaking, crew-injury territory.
Solar catamaran at 8.3°: ≈ 1.4 + 1.6 = 3.1 m/s² ≈ 0.31 g.
Stabilized trawler at 5.2°: ≈ 0.9 + 0.7 = 1.6 m/s² ≈ 0.16 g.

(Peak estimates; halve them for RMS.) This single calculation explains most "my boat feels violent" complaints.

3 · Why each platform feels the way it does

3.1 Sailboats: the sail as a damper, not a thruster

Three mechanisms, in order of importance:

Aerodynamic damping. When the boat rolls at rate p, the rig sees a changing apparent wind. At the sail's centroid (height h above the roll axis) the flow-angle perturbation is δα ≈ p·h/V_AW. The resulting lift change opposes the motion. Worked estimate for a 40-footer (70 m² of sail, centroid 7 m up, 16 kn apparent wind, rolling at 0.15 rad/s): δα ≈ 0.13 rad, force modulation ≈ 1.1 kN, moment ≈ 7.5 kN·m per