How it works

Each number on this site is either a published figure, our own estimate, or calculated from those. This page explains which is which.

What the ratings can't tell you

Every stability and speed number on this site is our estimate. None of them comes from the maker or from timed runs, and none is a promise about how a ski will feel for you. Your weight, skill, seat and footplate set-up, and the water on the day all change that. Use the ratings to narrow a shortlist, then paddle the skis on it. If a number looks wrong, tell us and we'll check it.

Where the numbers come from

Length, beam, deck height and layup weights come from the maker. Where a maker doesn't publish a figure, we say so. Prices come from dealers in the markets we cover, in the dealer's own currency, and the footer shows when they were last checked. Converted prices are marked as converted.

Stability

Stability ratings come from the hull's shape, not from reviews. On 39 craft we traced the beam and depth from the maker's side and top-down photos, set the hull at the waterline for its design weight, and calculated its metacentric height. That's the distance from the paddler's centre of gravity up to the point the boat pivots around as it leans, and it describes how steady the ski is in the first few degrees of lean. Traced hulls get a single whole-number rating at medium confidence. A photo trace can't capture every detail of the hull, so we don't show decimals.

The ratings are calibrated against boats we've paddled. At 77 kg on flat water on 3 September 2026, an Epic V10 was at least as steady as a first-generation Fenn Swordfish, both were clearly steadier than a pre-Fennix Fenn Spark, and the 54 cm-wide skis (Epic V6, V8) belong near the top. The rating combines two measurements from the traces, metacentric height and beam at the waterline, weighted 25 to 75. Beam carries most of the weight, because metacentric height alone could not match the boats we paddled. We keep the exact values internal. Pages show the rating or range and how confident we are. Where two craft have overlapping ranges, we can't say which is more stable.

The other 188 craft haven't been traced. For those, we estimate stability from the published beam, length and deck height using a generic hull shape, then adjust it to match the 39 traced hulls. This explains about 70% of the variation and when we leave a brand out and estimate it, metacentric height is off by about 16 mm and the final rating is off by about 0.3 points. Tested on the traced hulls, this estimate is off by half a point on average and within one point on 37 of 39, so these craft show a range of one point either side, labelled dimension estimate. Two older craft, the first-generation Fenn Swordfish and the pre-Fennix Fenn Spark, use the trace of their traced successor, and their pages say so. Outriggers, touring skis, spec skis and craft without a published beam get our own estimate with a wider three-point range, because there's nothing measured behind it.

The ranges show how precise the inputs are. They aren't statistical confidence intervals. Stability on flat water isn't the same as stability downwind: a hull that feels steady sitting still can still be hard to handle on a run, and one that feels tippy at the beach can be settled in a following sea. Each craft page has an Evidence note covering that.

Speed is our own rating, based on hull dimensions, independent reviews and owner reports. On traced craft it's adjusted toward the hull measurement, as explained below. Stability is rated at the hull's design weight. The weight guide gives you a starting point for a test paddle. A paddler 20 kg heavier makes the ski sit lower in the water; a lighter one makes it sit higher, with the effect on handling varying by hull and conditions.

Speed

On 39 craft, the speed rating also uses the traced hull. We scale the trace to the published beam, set it at the waterline for its design weight, and calculate the wetted area, how the volume is spread along the length, and the drag at a fixed 160 watts. That gives a speed based on the shape alone, independent of reviews.

To test it, we left one brand out at a time and had the model rate that brand. Using the hull shape cut the error against our own ratings from 1.14 points to 0.88, and none of 200 tests with random inputs did as well. So the measurement adds real information. But it was fitted to our ratings, not to timed runs, so it can smooth out our inconsistencies but can't fix a consistent bias.

So the rating shown moves a third of the way from our rating toward the modelled one. That matches the likely error in each: our ratings are set in half points and have about that much error, and the model is off by about three quarters of a point. In practice it changes most ratings by a tenth or two, and by half a point at most. Each craft page says when this applies.

Weight fit

The stability rating is for the hull at its design weight and doesn't change with your weight. The weight-fit label is separate: it compares your weight with the maker’s guidance or our estimate. An estimated range does not decide whether you can enjoy the ski.

A heavier paddler makes the hull sit lower in the water; a lighter paddler makes it sit higher. How that changes handling depends on the hull, paddler and conditions, so we describe the possible effects instead of adjusting the rating. The figure below shows how it works; it isn't a second rating.

The paddler weight band

Where a maker publishes a paddler weight range, we show that range as published. Where it gives only one end, such as "best for paddlers up to 80 kg", we fill in the other end with the calculation below. Load capacity is a different number and is never used as a paddler range.

Where a maker publishes nothing, we work the range out from the hull's dimensions:

upper limit = 111.5 × L × B × H − boat weight − 2 kg kit

L, B and H are length, beam and deck height in metres. The constant is fitted to the 33 skis whose makers publish a focused paddler limit. Checked against each of those skis with that ski left out of the fit, it lands within 5 kg of the maker's figure on average. Doubles use a higher constant and are given per seat.

The calculated range runs from 0.70 of the upper limit, which is where makers' own ranges typically start, with an estimated sweet spot from 0.86 to 1.00 and a small heavy margin to 1.04 for the error in the estimate. Estimated deck heights are marked and add about ±5 kg of uncertainty. Hulls under five metres are flagged, because a box is a poorer guide to a short, full hull.

A box can't see volume taken out of the bow and stern. Carbonology's Boost X LV is only 2 cm lower than the Boost X but rated 40 kg lighter, which is why a maker's figure always comes first. The calculation estimates fit from dimensions; it does not establish the designer’s intended paddler weight. Being outside an estimated range does not rule a ski out. Seat fit, conditions and paddling style matter, and lighter paddlers can enjoy larger skis downwind. Check any published load limit and try the ski before choosing it.

What we don't do

We don't sell boats, hold stock or take commission on links. We don't list second-hand prices, because the used market is too local and too small to average reliably. And we don't rank craft overall, because the best surfski depends on who's paddling it.

Craft that are no longer made

Discontinued craft are kept in a separate section for anyone looking at a second-hand boat. They're left out of the finder and the main list so they don't appear in a shortlist of skis you can buy new.

Corrections

If a figure is wrong, please tell us, especially if you're a maker and we have your boat's dimensions wrong. Each craft page lists its sources.