A journey with a fixed end date should carry its slack at the end

Some dates move and one usually does not

Almost everything in a long journey is negotiable. A night in one town can become a night in the next, a walk can be cut short, a museum can be missed. What is not negotiable is generally a single fixed point at the far end: a flight home, a return to work, a wedding, a visa expiry.

Recognising that asymmetry early changes how a route is designed. The middle of the trip is a soft structure that can absorb almost anything, while the last link in the chain has no give in it at all and everything upstream is feeding risk into it.

It is worth naming the fixed point explicitly rather than assuming it. Sometimes there are two, such as a flight and a hotel that will not refund, and sometimes the one that looked fixed turns out to be changeable for a modest fee.

Delay accumulates, it does not average out

There is a persistent intuition that late arrivals and early ones cancel each other. They don’t, for a simple structural reason: a train that runs early doesn’t let you take an earlier connection that has already gone, so the gains are discarded while the losses are carried forward.

This asymmetry is why a chain of tight connections behaves so much worse than the individual probabilities suggest. Each link only has to fail once for everything downstream to shift, and the failures do not need to be dramatic. A twenty-minute delay is enough when the transfer was designed around fifteen.

It is also why the number of moves matters more than the length of the journey. A single long leg is exposed to one set of risks; four short legs that add up to the same hours are exposed to four, and each one hands its problems to the next.

Slack placed early gets spent on the early part

A buffer day in the middle of a trip is not really a buffer. It’s a spare day, and spare days are absorbed by whatever is in front of them: a place worth another night, a walk that overran, a morning that was too pleasant to leave. That is a fine way to use a day, but it isn’t protection.

The only slack that reliably protects the fixed point is the slack sitting directly before it, because nothing downstream can consume it. Everything else has an opportunity waiting to spend it, and travellers are extremely good at finding those opportunities in the moment.

This is the argument for the unglamorous last night. Arriving in the departure city the evening before rather than the morning of a flight converts every upstream delay from a crisis into a story, and it does so more effectively than several extra hours distributed through the previous fortnight.

Redundancy is a different tool from buffer

Buffer buys time. Redundancy buys an alternative, and the two solve different failure modes. If a route depends on one ferry a day, no realistic amount of buffer helps when it does not sail; what helps is knowing whether a bus goes the long way round and roughly when.

Single points of failure are usually easy to spot once you look for them. A daily service, a seasonal crossing, a single bridge, a border post with limited hours, a last connection with nothing behind it. Each of these deserves a moment of thought about what the fallback actually is.

The fallback does not have to be booked or even good. It only has to be known, because the expensive part of a disruption is usually the hour spent discovering the options rather than the options themselves, and that hour tends to arrive at night in an unfamiliar station.

The cost of a buffer is the point of it

A day held in reserve at the end of a trip is a real day, spent in a city near an airport rather than somewhere you chose. Nobody should pretend that is costless. The question is whether that day is worth less than the consequence of missing the fixed point, and usually it plainly is.

That calculation shifts with the stakes. A cheap flight that could be rebooked for a modest sum justifies far less protection than a long-haul ticket, a visa run or a date somebody else is depending on, and the buffer should be sized against the consequence rather than applied uniformly.

It also shifts with the route. A journey along a dense network with hourly departures needs very little protection, while the same distance across a region with one daily bus and a border in the middle deserves considerably more, and that is a judgement about the network rather than about caution.

Where the discipline breaks down

The plan almost always survives contact with the first week and fails somewhere in the last few days, when the accumulated small overruns arrive together and the remaining legs are the least flexible. That is the predictable shape, and knowing it is most of the defence against it.

The other common failure is the buffer that gets quietly reallocated, because a spare day at the end is tempting to spend on one more place, and that decision is taken a week earlier when the risk still feels theoretical.

A modest habit helps more than a rule. Decide at the start which day is the protected one, and treat any proposal to spend it as a decision that has to be made deliberately rather than as an obvious improvement to the itinerary.