Train, ferry, ride: building routes you can't do as a loop

There is a category of ride you cannot build as a loop. The 40 km descent off a pass that would be miserable ridden uphill first. The coastal road that runs one way along a peninsula and stops. The route that follows a river from source to sea. Loops are the default because loops end where the car is, and that default quietly deletes some of the best riding in any region.

A train or a ferry removes the constraint. You ride out in a straight line and let something else carry you back, or you let it carry you out and ride home. The planning is more involved than a loop — there are timetables, bike rules and a connection you can miss — but it is not difficult, and the reward is a class of route that is otherwise unavailable.

What the linear ride actually gives you

Three things, mainly.

Net descent. Take the train up the valley, ride down. A route that gains 200 m and loses 1,400 m is not a soft option — long descents demand concentration and cost you body heat — but it is a completely different day from the same road ridden in both directions. Mountain rail lines and lake ferries exist in exactly the places where this works best.

A guaranteed tailwind. On a loop, wind is a tax you pay in both directions and mostly get back. On a point-to-point you can simply choose the direction. This is worth more than most riders assume, because the arithmetic is asymmetric: in the models that predict road cycling power accurately, aerodynamic drag scales with the square of the air speed over the bike, so a 20 km/h headwind costs considerably more power than the same 20 km/h tailwind gives back. Riding both ways is a net loss. Riding one way is not. Our guide to reading the wind forecast covers how to pick the direction with reasonable confidence a day out.

Roads that do not connect. Peninsulas, islands, dead-end valleys, one coastal corniche with no parallel road inland. A ferry turns a dead end into a through route and opens a map that looked closed.

Plan the return leg first

This is the whole discipline of multimodal riding in one sentence. The timetable is fixed and your legs are not, so the fixed thing goes on the page first.

Find the last service of the day that takes bikes and is not a replacement bus. Work backwards from it. Subtract a buffer — the amount depends on how often the service runs, but for anything less frequent than hourly, 60 to 90 minutes is not excessive. What remains is your ride window, and the route has to fit inside it with the climbing it contains, not with your best-case average speed on a flat day.

The bail-out points on that timeline are the second half of the discipline. A route that runs roughly parallel to the rail line — or crosses it twice — gives you intermediate stations where the day can end early without ending badly. This is one of the quiet advantages of following a river valley: rivers, railways and roads all want the same gradient, so they tend to run together.

Bike rules, which vary more than anything else

There is no single answer to "can I take my bike". It varies by country, by operator, by train type, by day of the week and by time of day. Check the specific service, not the national brand.

In the EU the floor has been raised. The rail passenger rights regulation in force since 2021 requires operators to carry assembled bicycles where the service allows it, and sets a minimum of four bicycle spaces on new and substantially upgraded rolling stock. That is a floor, not a promise: it does not apply to older carriages, and on many long-distance services the four spaces are reservable and go early.

Two practical notes. First, a bike reservation and a passenger reservation are often separate transactions on separate screens, and buying one does not buy the other. Second, where the bike space is unreserved, arriving ten minutes early is worth more than any amount of planning, because loading is first-come and the carriage is usually at one specific end of the platform.

Why this works so well where it is normal

Places that have taken bike-and-rail seriously show what the combination is capable of. In the Netherlands, a large share of rail passengers reach the station by bicycle — the bike is treated as part of the journey rather than an awkward item of luggage — and research on the combination describes it as producing a genuinely distinct mode with a reach neither has alone. Comparative work on Dutch bike-and-ride, and on the slower North American integration of racks, carriages and station parking, both point at the same conclusion: the constraint is almost never the riding. It is the interface.

That matters for you because the interface is where the planning effort belongs. Research on public transport transfers finds that a connection carries a cost beyond the minutes it takes, because of the uncertainty attached to it. A single-connection day is much easier to hold together than a two-connection day, and a plan with one transfer and a generous buffer will feel better than a clever plan with three tight ones, even when the clever plan is faster on paper.

What goes wrong

A workable template

The version that has survived contact with reality looks like this. Ride out from home or from a near station in the morning, with the wind. Take the train or ferry back. This ordering has two advantages over the reverse: your fixed deadline sits at the end of the day when you know how your legs are, and if you are running late you can shorten the ride rather than miss the start of it.

Build the route on the planner first, then check it against the timetable rather than the other way round. In Moveee's route library you can keep the outbound leg and a shorter bail-out version of it side by side, which makes the "cut to the earlier station" decision a matter of switching files rather than improvising on the roadside. Save both to the head unit before you leave.

Pack for the transport as well as the ride. A cable lock small enough to secure the bike in the carriage while you find a seat. A spare layer, because station platforms in the evening are colder than the ride was. Enough food to cover a delay, since the plan already assumes you are at your limit of daylight. If the route ends somewhere unfamiliar, how to find good roads in a place you have never ridden is the companion problem, and for multi-day versions of the same idea, bikepacking route planning covers resupply and daylight in more depth.

The first time you do it, the timetable will feel like the point of the day. By the third time it will feel like what it is: a piece of admin that buys you a road you could not otherwise ride. That trade is a good one almost every time.