Why the Shortest Route Between Multiple Stops Is Not Obvious (Boater's Edition, 2026)

Give a skipper two stops and the route plans itself. Give the same skipper six stops - fuel, three anchorages, a lunch dock, home - and something interesting happens: they will produce an order quickly, feel confident about it, and be wrong by double-digit percentages more often than not. This is not a boating failure; it is one of the most studied problems in mathematics, and the reason a whole industry of delivery-routing software exists. This piece explains the problem honestly, why water makes it worse, and what "optimized" actually means when a planner claims it.

#The factorial wall

With a fixed start and a fixed end, the number of ways to order the stops in between is a factorial: 3 stops between the ends give 6 orders, 5 give 120, 7 give 5,040. Factorials are why this problem is famous: the pile of candidates grows faster than any intuition can sample it. People handle 3 or 4 spread-out stops well - the eye finds the sweep. At 6, studies of human route choices show consistent mis-orderings, typically 10-30% over the optimum, and worse when the stops do not form an obvious ring. The skipper picking an order at the helm is sampling perhaps three or four candidate sequences mentally out of hundreds and choosing the best of the sample. The best of four random-ish samples out of 720 is not the best of 720.

#Water is an adversarial map

The road version of this problem at least gives intuition honest inputs: road distances roughly track how far apart things look. Water lies. Two anchorages 3 miles apart across a peninsula are 14 miles apart by boat. A bank that dries at low water turns a "direct" leg into a dogleg through a single charted cut. The consequence is brutal for hand-ordering: even a skipper with perfect ordering instinct gets wrong answers when the perceived distances are wrong. Any real optimization must first route every stop-to-stop leg over the navigable network - the way you would actually run it - and only then compare orders. Optimizing over straight lines produces orders that are optimal for a seaplane.

#What the payoff looks like

The savings depend on geometry. Stops strung along a coastline in a line: small savings, any sane order sweeps them. Stops spread around a bay, an island group, or both sides of a channel: this is where hand-picked orders backtrack, and computed orders routinely return 20-50% of the running distance. The interesting property is that the worst hand-ordering cases are precisely the ones that feel fine at the helm - the order matches the sequence in which the stops were suggested, or puts the "main event" first, and the backtracking hides in legs that each look reasonable alone. The total is what gives it away, and totals are what optimizers minimize.

The Marine OS Multi-Stop Optimizer implements exactly the honest version of this: up to eight stops, every leg routed over the global sea-lane network first, every visiting order checked against those routed distances, exact best order returned - and, deliberately, the distance of the order you originally typed shown next to it. That delta, "saves 11.4 nm versus your order," is the whole argument in one number. The result applies to your route as one editable line, and the Voyage Timeline prices each stop in ETA and forecast terms.

#Frequently asked questions

Trust the sweep for three stops; compute it for six. The practical method is in the multi-stop trip planning guide, applied setting by setting in the island hopping guide and the day cruise guide.