Measuring Fleets

What is the world's largest navy? The second largest? How has this changed over time? Can't we just count how many ships each has?

Oh. Well, it's true that not every vessel is equal, but all navies are a mix of large and small vessels, and so it should come out in the wash, right? What? Some navies expect to operate close to home and buy a lot of small ships, while others buy only big ships? So just counting hulls is going to make you look really silly, isn't it? That seems like something to avoid.

OK, you're the expert. Tell me how we're going to do this.

Clearly, measuring navies is a problem that is at least moderately difficult, placing it outside the capabilities of the typical journalist. As a result, we've had to put up with stories about the PLAN overtaking the USN as the world's largest, thanks in large part to about 100 missile boats and 70 corvettes, vessels smaller than almost anything the USN has in service. Astonishingly, several of the articles turned up on Google when I went looking actually mention this and explain it, but it still remains beyond the grasp of, for instance, professors at the Naval War College.

So what do we do instead? There are several options we can pick from that give results that are closer to what we'd expect and more closely approximate actual combat power. By far the easiest of these is to count warship hulls of frigate size and above, along with all non-midget submarines. While this still counts a 3,000 ton frigate or a 1,600 ton submarine the same as a CVN, in practice this method screens out the small stuff that really messes up these comparisons. Frigates and submarines are about the smallest really useful warships, and nobody is actually trying to build a "frigate swarm". Everyone who has a lot of frigates also has other, bigger ships roughly in proportion. Obviously, this isn't a perfect model. Not all ships are actually equal, and in particular, there are several countries with quite large diesel submarine fleets that are overcounted by this method. But SSKs are still important units in a way that missile boats aren't, so cutting them out entirely isn't really a good plan. At the very least, the frigate+ method is a decent way of getting a rough idea of relative naval power from easily-available data, although it has obvious flaws.

The obvious alternative to just declaring certain ships out of bounds is to take the size of ships into account directly by counting tonnage. This has two major downsides. First, tonnage is terrible, and you need to make sure that all of your tonnage measurements are to a common baseline. Second, you need a lot more numbers. That is a manageable problem if you just want to compare the US and China today, but let's say you're asked when the Soviet Union overtook Britain as the 2nd largest naval power. Using the Frigate+ method, I can generate an answer (1949) within about 3 minutes from data in the book The Changing Face of the World's Navies. Doing a tonnage calculation easily would involve a fairly complete set of Jane's from the late 40s and early 50s and a complicated spreadsheet to keep track of all the different classes. Or if you don't have access to those, you could spend hours trying to reconstruct the fleet from other reference books. That would be a massive pain, and while it would push back the date somewhat due to the lack of Soviet capital ships, it doesn't seem like a fun way to spend your time.

Nor does using tonnage free us from the tyranny of having to decide what to count. Sure, we can safely include missile boats without skewing the results, but there are still a lot of judgement calls to make. Do we include the US Coast Guard, and if so, do we include all cutters, or just those over a certain size? What about the Chinese Coast Guard, which is a lot harder to find information on? What about auxiliaries, which are often very large compared to warships? The Iranian Navy operates IRIS Makran, a converted oil tanker which displaces more than every other ship in their fleet put together, including other auxiliaries. Obviously, you can make calls to deal with each of these problems, but there isn't an obviously right answer.

There are various other ways you could try to generate a single number, counting VLS cells/missile tubes being the most obvious, but the best approach is to abandon the quest for a single number and instead rely on type-by-type comparison. Even this can be messed up by getting the types wrong, and my approach would be to try and split them into as few categories as you can while still keeping the categories fully meaningful. For instance, you want to make sure true aircraft carriers are separated from helicopter carriers and SSNs and SSKs don't end up under the same heading, but lumping all fleet escorts (CG/DDG/FFG) together is basically fine.¹ Obviously, this isn't perfect, as there will be considerable variation in capability within each category, but it's at least something which can be done reasonably quickly and will give results which actually make sense. For instance, here's the US v China under that scheme as of mid-2024:²

From this, we can easily see that the US has huge advantages in carriers and SSNs, the two sides are about even in fleet escorts (although the Chinese figure includes a lot of Type 54 FFGs, while the USN figure is basically all Ticos and Burkes) and the Chinese advantage is entirely in SSKs (dangerous in coastal waters) and small stuff that we shouldn't really care too much about. This gives a vastly more accurate picture than the simple hull comparison, even if the people doing hull counting managed to avoid sweeping in really silly things like patrol vessels that basically just have a gun for reminding merchant ships that they are serious.

And if you're too lazy to look this stuff up on your own, you can do what I did to get the data and grab a copy of the annual Seaforth World Naval Review. Their classification scheme is basically what I described above, although I of course might have minor quibbles with where certain ships go. In any case, even new editions are quite reasonably priced, and used copies from a couple years back are very cheap.

Obviously, even this method conceals a lot of variation within each type of ship, for instance counting the Russian Kuznetzov as equivalent to the Ford, or inflating the Chinese "fleet escort" count with about 40 Type 054A frigates, while the USN list is almost entirely Burkes and Ticos of twice the size and far greater capability. But it's probably the best that can be done without spending a ton of time and effort on a ship-by-ship comparison, and it gives far better answers than simply counting hulls, as so many make the mistake of doing.