From CEU to Bookable Capacity: Why a RoRo Vessel Rating Is Not a Sales Number

Executive summary

A vessel rated at 8,500 CEU does not have 8,500 sellable slots on any given sailing. The gap between the rating and what a planner can confirm is not a rounding error, and it is not the same gap twice in a row.

Car Equivalent Unit and lane metre are metrics for comparing vessels. They were never planning metrics for voyages. Between the number on the fleet page and the number a commercial team can commit to sit five groups of constraint: the space a unit actually occupies once it is stowed rather than measured, which decks it is eligible for once height and structural loading are applied, whether it can physically get aboard, whether the carrier will accept it in that condition, and what remains available on the specific leg being sold rather than across the voyage as a whole.

Each of those is measurable. Most are governed by published rules or published carrier standards. Almost none of them appear in a CEU figure.

This paper sets out the five groups, shows where each one binds, and gives a method for converting a nominal rating into a defensible number for a specific leg. It does not claim a universal percentage, because there is not one. What the deck loses depends on the vessel, the cargo mix, the port rotation and the operating standards of the carrier doing the stowing.

1. The nameplate illusion

Start with two figures published by the same operator about the same ship.

Höegh Autoliners' New Horizon class carries 8,500 CEU across 61,800 square metres of deck area. Divide one by the other and a nominal CEU slot on that vessel corresponds to about 7.3 square metres of deck.

A CEU is conventionally taken as the footprint of a 1966 Toyota Corona RT43, roughly 4.13 metres by 1.55 metres, which is about 6.4 square metres. No standards body appears to define this. It is convention, repeated consistently and traceable to no primary authority, which is itself worth knowing before anyone builds a commercial commitment on it.

So the operator's own two published numbers already carry roughly 14 percent more deck area per slot than the reference footprint implies. That margin exists before a single real constraint is applied. It isn't waste and it isn't an error. It's the first visible sign that the rating and the deck are describing different things.

The metric doesn't survive a change of vessel type either. Grimaldi's GG5G class is measured in lane metres, not CEU, and quotes 7,800 lane metres and more than 500 trailers. There's no honest conversion between a lane metre figure for a short-sea multipurpose RoRo and a CEU figure for a deep-sea pure car and truck carrier. They count different things for different reasons.

2. Why this matters now

The durable problem is structural rather than cyclical. In most carrier organisations, engineering capacity, cargo acceptance and leg inventory are three separate numbers, held by three different functions, reconciled late or not at all. Naval architecture owns the first, operations owns the second, commercial owns the third, and the vessel sails against whichever one was loudest.

That doesn't expire with the rate cycle. What the rate cycle changes is only where the cost shows up.

When rates soften, capacity errors show up in margin. When rates strengthen, the same errors show up as lost revenue.

There's also a signal worth watching, though it isn't yet an obligation. Amendments to SOLAS Chapter II-2 entering into force on 1 January 2026 tightened fire safety in vehicle spaces, special category spaces and RoRo spaces, including a requirement for video monitoring with cameras positioned high enough to see over cargo and vehicles after loading. Those amendments bind RoRo passenger ships. A pure car and truck carrier is a cargo ship and isn't covered by them. The direction of regulatory attention is clear enough to plan for. The obligation isn't there today, and any paper telling you otherwise hasn't read the applicability clause.

3. The five constraint groups

Group 1: Space

A vehicle's stowed footprint is larger than its measured footprint, because units need working space around them. Drivers have to get in and out. Lashing crews have to reach securing points. Neither is optional and neither appears in a length times width calculation.

The gap between measured area and stowed area is deck space that no CEU count reflects. Section 4 quantifies it against one carrier's published standard.

Broken stowage sits in the same group: the area lost to hull contour, structural intrusion, access lanes and the simple fact that rectangular cargo doesn't tile an irregular space. It's real, and it's routinely buried inside a planning assumption rather than measured per voyage. Published quantitative ranges for vehicle carriers specifically are scarce, and the figures that circulate come from military sealift analysis rather than commercial operation. Treat broken stowage as a parameter to measure on your own vessels, not a number to import.

Group 2: Deck eligibility

Two things determine whether a unit can go on a given deck, and neither is its weight.

Height, adjusted for cargo class. Usable height is not deck height. It is deck height minus a clearance margin that varies by what is being carried. Wallenius Wilhelmsen publishes its requirement as 5 centimetres of clear height to official deck height for a car, 10 centimetres for high and heavy and breakbulk, and 50 centimetres for a chassis without a cabin, with the further instruction that below 10 centimetres of clearance driving speed must be reduced to a minimum. A deck that nominally clears a unit may therefore not accept it.

Hoistable decks make this sharper rather than softer. Raising an intermediate deck to create height for high and heavy cargo does not redistribute that deck's area elsewhere. It removes that deck's area from inventory. The Höegh Aurora class carries 14 decks including five liftable, with a maximum deck height of 6.5 metres. Reaching that 6.5 metres means a deck's worth of car slots ceases to exist for that voyage. The trade is zero sum and it is a commercial decision disguised as an engineering one.

Structural loading, which is not about mass. A vehicle deck is not loaded by the weight of a vehicle. Under Bureau Veritas rules for the classification of steel ships, static wheel force is derived from the axle load divided by the number of wheels on that axle. That force is treated as pressure distributed over the tyre print area, whose dimensions the designer must supply along with the arrangement of wheels on axles, the load per axle and the tyre pressure. For assessing stiffeners and primary supporting members the force is taken as concentrated at the centre of the tyre print, and the forces on primary supporting members must account for the area of influence of each member and the way ordinary stiffeners transfer the load. Tracked vehicles are treated the same way, with the print taken below each wheel or wheelwork.

The consequence is direct: two units of identical total weight load a deck differently if they differ in axle count, wheels per axle, tyre pressure, print size, or position relative to the supporting structure. A unit can be within a deck's distributed load and still be inadmissible where it needs to stand.

This is not a marginal technicality in classification terms. To classify a vehicle-carrying ship, the owner must submit a plan of design loads on decks covering fork lift areas, axle loads and tyre print areas of wheeled loads, a plan of vehicle arrangement indicating securing and load bearing arrangements, and the characteristics of every vehicle type intended to be carried, specifically the axle load, axle spacing, number of wheels per axle, wheel spacing and size of tyre print. The ship is classed against a declared cargo profile. Cargo outside that profile is a question, not a booking.

Group 3: Access

Access binds before any deck question, and a unit that fails here fails completely regardless of what the decks could have taken.

Ramp capacity varies enormously across current tonnage, and the variation is not proportional to vessel size. The Höegh Aurora class publishes a maximum ramp capacity of 375 tonnes. The New Horizon class, at 8,500 CEU on a similar overall length, publishes a stern ramp of approximately 250 tonnes and a side ramp of 22 tonnes. Grimaldi's GG5G quarter ramp is rated to 150 tonnes. A 200 tonne unit is loadable on one of these and not on the others, and the decks below have nothing to do with it.

Ramp geometry is also not static on the day. Approach angle changes with tide, vessel draft and quay height, so a low unit that fits in principle can ground out in practice. The same physical cargo and the same physical ship can produce different answers at different hours.

Group 4: Cargo acceptance

A unit can fit, be eligible for a deck, and still not be accepted.

Cargo class changes the space arithmetic. Under the same published standard cited in Section 4, high and heavy cargo requires 30 centimetres side to side against 10 centimetres for a car, and 50 centimetres bumper to bumper against 30. Cargo class is not a label on a booking. It is a different area consumption per unit.

Securing geometry constrains density. For road vehicles with a maximum total mass between 3.5 and 40 tonnes, IMO's revised guidelines recommend that the distance between securing points should in general not exceed 2.5 metres longitudinally, and that the maximum securing load of lashings should in general not be less than 100 kilonewtons, with lower values permitted for vehicles not exceeding 15 tonnes gross vehicle mass. A unit whose tie-down geometry does not align with the deck's fixed securing grid has to be repositioned, and repositioning creates void space.

Two qualifications matter and are frequently dropped. These are recommendatory guidelines, not mandatory requirements, and they apply to road vehicles, not passenger cars. Anyone extending them to car deck density is citing a real document about something else.

Condition constrains acceptance. Carrier electric vehicle policy now caps state of charge on acceptance: Wallenius Wilhelmsen states 50 percent or lower and preferably below 30 percent, while UECC specifies a minimum of 20 percent and a maximum of 50 percent. These are carrier and insurer positions, not IMO regulation, and they bind commercially rather than legally. They bind nonetheless.

Group 5: Network inventory

The final constraint is not physical at all, and it is the one most often measured wrongly.

Capacity is per leg. A vessel on a multi-port rotation can be full between the second and third calls and half empty between the fourth and fifth. Only the leg being sold matters to the next booking. Voyage-level utilisation, calculated as total cargo moved divided by nominal capacity, averages across exactly the variation a commercial team needs to see, and reliably reports comfort where there is none.

The academic literature is clear that this is hard rather than merely fiddly. Optimising stowage across a multi-port rotation is a recognised and computationally difficult problem, treated in the operations research literature since Øvstebø, Hvattum and Fagerholt's 2011 work on stowage plan optimisation for RoRo ships and still active today. What that literature largely addresses is feasibility and algorithms. What it does not resolve is the commercial question this paper is about: what can still be confirmed, for a specific leg, once cargo, deck and reservation constraints interact.

On top of the physical picture sits commercial segmentation. Space reserved for a port pair, a customer, a contract or a cargo class is not available to the next enquiry even when the deck is empty, and reservations across all segments cannot exceed what the deck can physically take. A reservation book that has drifted past physical capacity is not a plan.

4. Working the numbers

A necessary caution about whose standards

Operating standards are carrier-specific. The figures below are Wallenius Wilhelmsen's published stowage standard, used here as an illustrative worked example because that carrier publishes it openly. They are not an industry benchmark, and they must not be applied to another operator's vessel to produce a combined result.

That caveat is itself part of the argument. Searching the major deep-sea vehicle carriers, Wallenius Wilhelmsen appears to be the only one publishing full cargo handling instructions publicly. The others keep them behind customer accounts. If the operating standards that determine how much of a deck is usable are mostly unpublished, then there is no shared industry basis for converting a nominal CEU rating into sellable slots. Every carrier's answer is different, and most are not visible from outside.

The published standard

Stowage distance, car	Requirement
Side to side, body or mirror	10 cm
Passenger side to fixed objects	10 cm
Driver side to fixed objects	50 cm
Width of passage or walkway	60 cm
Bumper to bumper	30 cm
Bumper to fixed objects	30 cm
Auto to breakbulk	50 cm

Stowage distance, high and heavy	Requirement
Side to side	30 cm
Bumper to bumper	50 cm
High and heavy to breakbulk	50 cm

Fixed objects here means bulkheads, stanchions, void spaces and panel edges. The standard also notes that where a manufacturer states its own requirements, those take priority.

What it does to a footprint

Take a car of 4.50 metres by 1.85 metres, a net footprint of 8.33 square metres. Apply 30 centimetres bumper to bumper and 10 centimetres side to side and the stowed cell becomes 4.80 metres by 1.95 metres, or 9.36 square metres.

That is about 12 percent more deck consumed than the vehicle measures, under one carrier's standard, for one vehicle size, before broken stowage, before deck eligibility, before anything else.

Why that number on its own is not the point

Twelve percent on a single vehicle is a technical observation. It becomes a commercial problem only when it meets a leg where the remaining inventory is small, and it is misleading if extrapolated naively across a whole vessel.

The denominator that matters is remaining bookable capacity on the leg you are selling, not vessel nameplate. A five percent area movement against a full ship is noise. The same movement against the last 400 slots on a constrained leg is the difference between confirming a booking and rolling it.

So the method is:

Run that and the output is not a percentage. It is a bounded answer to a question a commercial team can act on: how much of this leg can we still sell, and how confident are we in the edges of that number.

One scope warning worth its own paragraph

Both figures are real, published and correct. Applied to the wrong place, either one produces a capacity answer that is not slightly wrong but categorically wrong. When a figure is quoted for capacity work, the question is never whether it exists. It is what it governs.

5. What it costs to leave this unmodelled

The failure modes are specific and they show up in ordinary operating life:

• Deck space that sails empty on a leg that was reported as comfortable because utilisation was measured across the voyage.
• Cargo rolled to a later sailing after a commitment was made against a number that was never true for that leg.
• Reshuffles at the quay when a unit that was accepted commercially turns out to be inadmissible on the deck assigned to it, or on the ramp before that.
• Space sold to two segments because reservations were tracked separately from the physical picture.
• High and heavy accepted at car area assumptions, quietly consuming several times the deck it was priced for.

None of these requires anything unusual to go wrong. They're the normal result of 3 functions holding 3 numbers that were never reconciled.

6. In practice

A vehicle carrier operating both pure car carriers and pure car and truck carriers has planned its cargo this way since 2025, across multi-port rotations spanning US, European, African and Far East trades, at more than 4,000 units per sailing.

Allocation is made at deck level, against confirmed bookings and forecast demand together, so space held for cargo that hasn't been booked yet sits visibly alongside space already committed. Competing stowage scenarios are run as simulations against the same voyage before one is chosen, which turns the hoistable deck question from an assumption into a comparison. Capacity is read per leg rather than per voyage.

The payoff is the absence of the failures in Section 5. No leg reported as comfortable because the voyage average said so. No commitment made against a number that was never true for the leg being sold. No reservation book that has quietly drifted past what the decks can carry.

7. What to do next

The five groups above are not a checklist to admire. They are the inputs to one number that most carrier commercial teams currently estimate rather than compute: how much of this leg can still be sold, and how confident are we in the edges of it.

You can build that number in-house. The inputs are knowable, the rules are published or held internally, and the arithmetic isn't exotic. What makes it hard is that the inputs live in 3 places. The vessel profile sits with naval architecture and marine operations. Stowage standards and cargo acceptance rules sit with operations. The reservation book and the leg-by-leg commitments sit with commercial. Reconciling them is a data problem before it's a planning problem, and that's the reason the number is usually estimated rather than computed.

The leg capacity review

If you want the reconciliation done against your own fleet rather than a worked example, that is a defined exercise rather than a product demonstration.

What we take in

• Vessel profiles for the ships on the trade: decks, clear heights, hoistable arrangements, structural and axle limits, ramp geometry and safe working load.
• Your own stowage standard. Your clearances and height margins by cargo class, not a published example borrowed from another carrier.
• The port rotation under review, loading and discharge calls in sequence.
• Whatever cargo is actually being offered, down to dimensions, weight, axle configuration and cargo class.
• The reservation book: what's committed, to whom, for which port pair, and in what priority.

What we return

• Bookable capacity per leg, as a range with the assumptions stated, rather than a single figure that hides them.
• The binding constraint on each leg, named. Not "the deck is full" but which of the five groups closed the door, because the commercial response differs entirely depending on the answer.
• Where reservations exceed what the leg can physically carry, and by how much.
• What a hoistable deck decision on that voyage costs in car slots, so the high and heavy trade is priced rather than assumed.

Which decision improves

Whether to confirm the next booking on a constrained leg, at what price, and against which reservation. That's the decision the number exists for.

The review runs on one trade and a handful of voyages. It's built to be checked against what actually happened on those sailings, which is the only test worth passing.

8. Sources

All sources below are publicly accessible without payment or registration.

Vessel specifications

• Höegh Autoliners, Aurora Class: 9,100 CEU, 14 decks including 5 liftable, maximum deck height 6.5 m, ramp capacity 375 t
• Ship Technology, New Horizon Class Pure Car and Truck Carrier: 8,500 CEU, 61,800 m² deck area, 14 decks including 5 hoistable, stern ramp approximately 250 t, side ramp 22 t
• Ship Technology, Grimaldi Green 5th Generation Class: 7,800 lane metres, quarter ramp 150 t

Carrier operating standards

• Wallenius Wilhelmsen, Cargo Handling Instructions: Stowage distances
• Wallenius Wilhelmsen, Cargo Handling Instructions: Height clearance
• Wallenius Wilhelmsen, "The State of State of Charge"
• UECC, Electric Vehicle Carriage

Classification and regulation

• Bureau Veritas, NR467 Rules for the Classification of Steel Ships, Part B: Hull and Stability, July 2026
• Bureau Veritas, NR467 Rules for the Classification of Steel Ships, Part D: Service Notations, July 2026
• IMO Resolution MSC.479(102), Revised Guidelines for Securing Arrangements for the Transport of Road Vehicles on Ro-Ro Ships, adopted 11 November 2020 (recommendatory)
• IMO, amendments in force from 1 January 2026: SOLAS Chapter II-2 fire safety for vehicle, special category and ro-ro spaces (applicability: ro-ro passenger ships)

Industry standards

• ECG Association of European Vehicle Logistics, Operations Quality Manual for Cars and LCVs, version 10, Section 5.2 Storage (compound storage clearances)

Literature

• Øvstebø, B.O., Hvattum, L.M. and Fagerholt, K., "Optimization of stowage plans for RoRo ships", Computers & Operations Research, 38(10), 2011, pp. 1425 to 1434. DOI 10.1016/j.cor.2011.01.004

About Logisoft

Logisoft is a logistics software platform built specifically for RoRo and finished-vehicle logistics, used by ocean vehicle carriers, forwarders, and terminal operators to run bookings, yard and terminal operations, customs, and billing on one shipment record. Cargo planning runs against that same record, so the vessel profile, the cargo actually booked, and the space reserved against each port pair are reconciled from one source rather than three. Logisoft is developed by INTRA Logisoft Labs, based in Nicosia, Cyprus. Learn more at logisoft.io.

To see bookable capacity computed per leg against your own fleet, request a leg capacity review.