You shouldn't need three browser tabs, a forum deep-dive, and a spreadsheet to know whether a cruise is worth booking. We did that work already. This page explains exactly how, because a score you can't interrogate is just marketing.
Every ship is graded out of ten. The grade judges the vessel itself: how good is the ship, regardless of what any particular sailing costs?
The average rating from people who actually sailed, with the number of reviews behind it shown honestly. A bare average without its sample size is a bluff.
Every sailing gets a Value Score out of 10: how much ship you get per dollar, ranked against genuinely comparable cruises — same cabin type, same region, same length. 5.0 is the going rate for that much cruise; above it the fare beats the going rate, below it it does not. A 9 or better is rare by design.
Quality counts the ship and the passenger reviews together, so a well reviewed ship is worth more per dollar than an identically specced one people did not enjoy.
Two things follow from that, and both are deliberate. A better ship at the same price always scores higher, and the same ship for less always scores higher. But the comparison is always like for like: ranking every cruise on earth by cost per star would just list the cheapest, since prices vary about twice as much as quality does. Seabourn would lose to a discount sailing every time, which tells you nothing.
Six numbers, weighted in the open, and one that can only help. Every one is a published measurement of the ship, not an opinion about it, so you can check our arithmetic against a brochure, a deck plan or a government inspection report.
| 25% | Space per guest. Gross tonnage divided by guests, counted at full occupancy. Roughly 25 tons each at the crowded end, 68 at the most generous. The single hardest thing for a ship to fake, because it is fixed in steel. |
| 25% | Cabins with a balcony. What share of this ship’s cabins open onto private outdoor space — from 16% on the oldest hulls to every cabin aboard the newest. The fleet median is 67%. It is the one measurement that separates a modern ship from an old one in the way a passenger actually feels, and it cannot be retrofitted into a hull that was not built for it. |
| 20% | Condition. Age, with a refit counted as erasing half a ship’s years, not all of them. A 1996 hull refitted last year is not a one-year-old ship, and a never-refurbished 2015 ship is not young. Announced future refits do not count until they happen. |
| 15% | Guests per crew member. Between 1.25 and 4.0, again at full occupancy. It is the closest honest proxy for how much attention there is to go round, and unlike opinions about service it is a headcount. It carries less weight than it once did for a measurable reason, given below. |
| 10% | Average cabin size. Square feet, averaged across every cabin grade aboard and weighted by how many of each the ship carries — 144 at the tightest, 486 at the most generous, median 204. Read off the published cabin tables, and withheld entirely for any ship whose tables do not add up to its own cabin count. |
| 5% | Health inspection. The US Public Health sanitation score, recent inspections averaged. The pass mark is 86 and every ship in the index clears it — the fleet runs 88.5 to 100 — so this does not measure whether a ship is safe to sail on. It measures how far above merely passing it inspects, which is a real difference in how a ship is run. Blank, and the other weights rescaled, for the 18 ships that do not call at a US port and so are never inspected by this authority. |
| +0.5 | Facilities on board. What is actually
there, with a water coaster counted above a shuffleboard court. This one is a
bonus, not a weight: a well-equipped ship earns up to half a point on
top of its measured grade, and a ship without those things is not marked down
for the omission.
It used to be a weighted input, and that was wrong. Facility count tracks hull size rather than quality, so every boutique luxury ship sat near the floor of it — the highest-graded ship in the fleet was losing more of a point to facilities than to space, staffing and sanitation combined. We already say elsewhere that a small luxury ship intentionally carries fewer venues; it cannot then be docked half a point for doing so. A roller coaster is a real thing to offer, so it still counts — upward only. |
Where a figure is genuinely unavailable, the remaining terms are re-weighted rather than the missing one being scored as zero. All 177 published ships carry the first six; 159 of them carry a health score as well.
The health term used to be a hidden deduction applied after the grade was computed, appearing in no column — so the published inputs did not add up to the published number, and 34 ships sat below their own measurements with nothing on the page to explain it. One ship lost nearly a full point that way. It is an input now, weighted and printed like the rest.
Both ratios above divide by the number of people a ship can actually carry, not by the two-per-cabin figure the industry quotes. This is the harder standard and we think it is the honest one.
Almost every published passenger count is “lower berth” — two people per cabin, upper bunks empty. That number is not neutral between lines. Compare a ship’s maximum capacity to its lower-berth capacity and Disney comes out at 1.40, the mainstream lines cluster near 1.20, Regent is 1.11 and Virgin is 1.03, because a family line sells third and fourth berths that an adults-only line does not build. Dividing by lower berth credits a ship for space and crew it only has while those bunks stay empty — and credits it most generously exactly where they are least likely to be.
So a ship that can carry 3,466 people is graded as a ship for 3,466 people. A full Disney sailing is a ship of four-berth family cabins and it is measured that way. Eleven ships in the fleet are unaffected, because their maximum and their lower berth are the same number: the Viking Ocean hulls and Celebrity Flora carry exactly two guests per cabin, having built no third berths at all.
This costs a little agreement with the inspection guide we calibrate against. We changed it anyway, because which number you divide by is a question about honesty rather than about fit.
A weighting that cannot be checked is just a preference with a percent sign on it. So we checked ours against the one external benchmark that exists: the Berlitz guide’s professional ship ratings, which cover 166 of our 177 ships and are compiled by inspection rather than by survey. We searched every whole-5% combination of the six inputs and kept the one that best reproduced that guide’s ordering. Rank agreement rose from 0.49 to 0.74. Splitting the fleet in half and fitting each half separately picked the same weighting and held up on the half it had never seen, which is the test that tells you a number is real rather than fitted to noise.
Two results from that exercise are worth stating plainly, because both of them went against what we expected.
Space per guest and guests per crew measure almost the same thing. They correlate at 0.83 across the fleet. Left to itself the arithmetic therefore drops the crew term to nothing — not because staffing does not matter, but because space has already counted it. We hold it at 15% anyway. That costs a little accuracy against the benchmark and keeps the formula honest about what it is made of.
Counting facilities makes the score slightly worse, and we do it anyway. Facility-dense ships are space-poor ships, so a facility count runs against the benchmark rather than with it. Weighted at 5% it costs about as much agreement as the crew term does. We keep it because a ship with a water coaster, an ice rink and a go-kart track genuinely offers something a ship without them does not, and a grade that refused to say so would be measuring only restraint. It is 5% and not 20% because that is what the evidence supports.
The benchmark is used to set the weights and nothing else. It is never shown here, never republished, and never any part of a ship’s score.
Two reasons, and we would rather show our working than assert a number.
What a refit costs. Royal Caribbean’s Amplified programme ran about 15% of a ship’s original build price — $120m on Mariner of the Seas against $650m to build her, $116m on Freedom against $800m. Allow for inflation between a 2003 build and a 2018 refit, and for the roughly half of any new ship that is hull and machinery no passenger ever sees, and a refit buys something like a quarter to a third of the part you actually experience.
What a refit changes. When Ovation of the Seas went into dry dock, the venues were swapped inside the rooms that already existed: one restaurant became another, the casino spread into the old music hall, and forty new cabins were cut out of a fitness centre and some technical space. Everything a guest touches was renewed. Nothing a guest moves through changed. Cabin dimensions, ceiling heights and deck geometry are fixed in steel on the day a ship is built and stay that way for her whole life.
Cost alone would argue for less than half. Scope argues for more, because recarpeting a room costs a fraction of building it. Half is where we landed, and it is a judgement, so it is written here rather than hidden.
One consequence worth naming: a refit can make a ship worse. Ovation gained forty cabins without gaining a single ton, so every guest aboard now has slightly less ship to themselves. That shows up as a loss on space per guest even as the ship gains on condition, which is the right answer.
Thin data shrinks a ship's grade toward the fleet middle rather than bluffing confidence, and shows an “est.” or “ltd data” badge. A score without error bars is a guess in a suit.
The Vessel Class grade is fixed and objective: you can't bend it, and neither can we. Instead, every sailing earns descriptive tags from its own data, and you filter by them:
Great food Recently refreshed Adults-only Great for families Ultra-luxury
A tag has a threshold behind it: “Great food” means the ship's food signal clears a fixed bar, not that a marketer chose an adjective. Pick any tags; a sailing must carry all of them to stay on your list.
The trap in “value” scores: luxury prices grow much faster than quality. The step from a great ship to the absolute best of the best can cost 2–3× more. Rank every sailing by simple price-per-quality and the finest ships in the world sink to the bottom of the list forever, which tells you nothing.
So we never compare across classes. For each sailing we work out its cost per point of quality — the fare per night divided by a quality figure that counts the vessel and its passenger reviews together — and set that against the median for its own group: sailings of the same class of ship, in the same region, of similar length. 5.0 means this sailing costs what that much cruise typically costs right now. Roughly every 20% below the going rate is worth about another point. A top-graded ship at a mid-market price scores high. A mid ship at a luxury price scores low.
Traveler reviews are inside that quality figure rather than bolted on afterwards: they carry about a third of its weight, the vessel itself carries the rest. A ship people love is worth a little more at the same fare. Reviews stay out of the Vessel Class star entirely, which has to keep meaning the vessel.
The fair price shown against some sailings is a different number: a fitted estimate of what a sailing like this one should cost, there to show you the gap. The Value Score is not derived from it.
Price never touches the stars. A discount doesn't make a ship better; it makes a deal better. And a sailing that is both better and cheaper than another can never score below it.
Sources, in general terms: shipbuilder and fleet registry data (tonnage, guest capacity, crew complement, build and refit years — the three inputs to the Vessel Class); US CDC Vessel Sanitation Program (public inspection data); public incident records including Wikipedia-sourced event histories; aggregate traveler ratings from major cruise-review communities, shown as their own figures rather than folded into a grade; professional ship-inspection literature, used privately to sanity-check the formula and never published as a score; ship and port photography from Wikimedia Commons and Wikipedia, used under their own licences and credited on the image. Sanitation data courtesy of the CDC VSP; scores are public records; interpretation is ours.
CruiseIndex links to cruise lines and travel agencies and may earn a commission when you book through those links; it never changes a score, a rank, or a fair price. Full affiliate disclosure · Privacy policy.
About the fares. Every price shown is a real fare, not an estimate. Illustrative sample fares were switched off entirely: a sailing we cannot price carries no price. Fares come from CruiseDirect, the same page the booking link sends you to, so the number on the card is the number at checkout. A handful of sailings carry a fare that is some weeks old and are labelled as such.
· Terms