Q-Logos

Straight-Line Distance vs Road Distance in Route Planning: What Each One Measures and When It Matters

Haversine (straight-line) distance and road-network distance answer different questions. Learn what each one ignores, when straight-line stop ordering is a reasonable first pass, when it misleads, and how to tell which one a route planner is showing you.

Two numbers that share one name

Ask how far apart two stops are and you can get two honest answers that disagree. One is the distance in a straight line over the earth's surface. The other is the distance a vehicle has to cover on real roads. Route planners use one or the other, they do not always say which, and the choice decides how much weight the figures on the screen can carry.

This page explains what each number measures, what each one leaves out, when the simpler one is good enough, and when it will send you the wrong way. For the wider picture of how a planner turns a stop list into a visiting order, start with how a multi-stop route planner works.

Straight-line distance: two coordinates and a formula

Straight-line distance, often called as the crow flies, is the shortest path between two points on the surface of a sphere. That path is an arc of a great circle, and the usual way to compute its length from two latitude and longitude pairs is the haversine formula. The inputs are the two coordinates and the radius of the earth. Nothing else.

That is its whole appeal. It needs no map, no network connection to a routing service and no knowledge of the area. It gives the same answer every time, and it is quick enough that a planner can compute it for every pair of stops in a long list without the user noticing.

It is also its whole limitation. A formula that knows only two coordinates cannot know what lies between them. It ignores:

  • Water. Rivers, bays, lakes and harbours, and where the bridges, tunnels and ferries are.
  • One-way streets. The road distance from A to B and from B to A can differ. Straight-line distance is the same in both directions.
  • Limited-access roads. Two points on either side of a motorway can be a short walk apart and a long drive apart, because the nearest junction or underpass is far away.
  • Barriers and terrain. Railways, airports, industrial estates, hills and valleys that roads go around rather than through.
  • The street layout itself. Even in a regular grid you drive along the sides of blocks, not across their diagonals.

Road distance: a map, a routing engine and a matrix

Road distance is the length of a drivable path between two points. To compute it, software needs a digital road network, with every road segment, junction, turn restriction and direction of travel, and a routing engine that searches that network for a path.

For a single trip from A to B that is one search. Stop ordering needs much more. To decide which order to visit a list of stops in, a planner needs the distance between every pair of stops, in both directions. That table is called a distance matrix, and its size grows with the square of the number of stops: double the stops and the table is roughly four times as large. Every cell is a separate question for the routing engine.

So road distance has a real cost. Someone has to host or pay for the map data and the engine, keep the data current, and wait for the matrix to be filled before the ordering search can even start. That cost is why straight-line distance remains common in planners, and why it is worth checking which one a given tool uses.

How far apart the two numbers can be

Road distance is never shorter than straight-line distance, since the straight line is by definition the shortest path. How much longer it is has no fixed answer. In a compact, well-connected street grid the two tend to move together: the road figure is longer, but stops that are near each other in a straight line are generally near each other by road. Across a river with few crossings, or in an area cut up by motorways, two points that are neighbours in a straight line can be far apart by road.

People sometimes apply a single multiplier to straight-line distance to approximate road distance. That can give a rough total for a whole region, but it cannot fix an ordering problem, because the error is not spread evenly. It is concentrated in exactly the pairs of stops that have an obstacle between them.

When straight-line ordering is a reasonable first pass

Ordering stops is a different job from measuring a trip. For ordering, what matters is not the exact distance but whether the ranking of distances is roughly right: is stop C closer to stop B than stop D is? Where straight-line and road distance rank pairs the same way, they produce similar visiting orders even though the totals differ.

That tends to hold when:

  • The stops sit in a dense, well-connected street network without large barriers.
  • The area is compact, so no pair of stops is separated by a detour that dwarfs the rest of the trip.
  • You want a quick sequencing of a list to review by eye, not a figure to bill or schedule against.
  • Stops form clear clusters, so the main decision is which cluster to visit first and the fine order inside a cluster matters less.

In those cases a straight-line order is a sensible draft. A person who knows the area can look at it on a map and correct the few places where local knowledge disagrees.

When straight-line ordering misleads

The same reasoning tells you where to be careful. Straight-line ordering goes wrong when the ranking of distances flips:

  • Water crossings. Stops on opposite banks look adjacent. A straight-line order may hop back and forth across the water, and each hop is a trip to the nearest bridge and back.
  • Limited-access roads and railways. Stops on either side of a motorway or a rail line can be ordered as neighbours when the drive between them is long.
  • One-way systems. In a city centre the order that looks short can require looping around blocks to approach each stop from the permitted direction.
  • Mountain and valley roads. Two villages in neighbouring valleys can be close on the map and far apart by road.
  • Sparse rural networks. With few roads, the drivable path between two points can bear little relation to the line between them.

There is a second way to be misled that has nothing to do with ordering. A straight-line total is a lower bound on the trip, not an estimate of it. Any drive time, fuel figure or comparison built on that total inherits the same gap. Treat those as model outputs, useful for comparing one order with another under the same assumptions, and not as what the odometer will read.

How to check which one you are looking at

  • Read the label. A careful planner states the metric next to the distance figure or in a method note.
  • Run the river test. Place two stops on opposite banks of a wide river or bay, away from any crossing. A short reported distance means straight-line.
  • Look at the drawn route. Straight segments from stop to stop, cutting across blocks and water, mean the planner is drawing the order, not a drivable path.
  • Check direction. If A to B and B to A always report identical distances in a city with one-way streets, the figure is probably straight-line.

A worked example: what Q-Logos measures

The Q-Logos web route planner is on the straight-line side of this divide, and it says so on the page.

The ordering uses straight-line distance. When you press the optimize button, the server builds a table of haversine distances between every pair of your stops and searches for a short visiting order on that table. Road networks are not part of that step.

The reported distance is the same metric, and it is labelled. The distance on each route card is the straight-line sum between consecutive stops in the returned order. Each card marks the figure and carries a note that it is the straight-line distance between stops, not road distance. The lines drawn on the map connect the stops in order; they are not a drivable path.

Durations are modelled, not measured. Travel times are derived from the straight-line distance using a fixed average speed, so they are estimates under a stated model. In the current version you can also choose a traffic level, and the planner may apply the current weather at your first stop; these scale the estimated minutes by a fixed planning rule set for the product, not by a live traffic feed. Time and fuel differences shown against the baseline come from the same model and are not measured savings.

What is not there. The web route planner does not compute road distance, does not consult live traffic, and does not give turn-by-turn directions. It also does not model delivery time windows, vehicle capacity or more than one vehicle. If your stops straddle a river, a motorway or a one-way centre, review the suggested order against a road map before you drive it.

Used with that in mind, the planner gives you three alternative orderings to compare for the same stop list, each with the same straight-line label. If your stops are in a spreadsheet, the page on the CSV import path explains the file shape it reads.

A short checklist before you trust a route

  • Find out which distance the planner optimizes, and which one it reports. They are not always the same.
  • If it is straight-line, use the result as an order to review, not as a distance or time to quote.
  • Look for water, motorways, railways and one-way areas between consecutive stops, and adjust those legs by hand.
  • Compare orderings with each other under the same metric, not against a figure from a different tool.
  • For billing, scheduling or driver hours, use road distance from a routing engine.

Plans differ in how many routes you can run and how many stops a route can hold; the limits are on the pricing page. To see the straight-line label on your own stops, create a free account and run the planner.

Create a free Q-Logos account, order your own stops, and read the straight-line label on each route card before you decide how far to trust the figures.

Get started