EVE Online Round-Trip Trading Guide: Never Fly Empty
Why the empty backhaul is a hidden pay cut: round-trip budgeting, the ISK snowball, chain routes between the smaller hubs, and when one-way genuinely wins.
Here's a number most haulers never compute: what their return trip earns. For half of every round trip, the typical trader is flying an empty hold — which means their real income is half what the outbound leg's margin suggested. The one-line takeaway: you're not paid per jump, you're paid per hour, and an empty backhaul cuts your hourly rate nearly in half for free.
This guide covers round-trip and chain trading: filling both directions, budgeting the legs so profit compounds, and recognizing when a one-way run genuinely is the better play. It assumes you know basic hub-to-hub arbitrage and which hubs anchor the map.
The empty-hold tax
Say Jita → Dodixie carries 40M of realizable margin over 15 jumps, about 25 minutes with docking and trading. One-way, you fly home empty: 40M over ~50 minutes, call it 48M/hour. Now suppose Dodixie → Jita offers even a mediocre 25M return load. Same 50 minutes, 65M total — your hourly rate just rose 35% without flying a single extra jump.
That's the entire thesis. The return leg doesn't need to be great; it needs to beat zero, which is what an empty hold earns. Traders reject 20M backhauls as "not worth it" while flying that identical route empty — the comparison isn't against your best route, it's against the nothing you'd otherwise earn on jumps you're already making.
The ISK snowball
Round trips have a second, quieter edge: the outbound profit funds a bigger return buy. Leave with 500M of budget, sell for 540M at the destination, and the return leg shops with 540M. On a chain — A→B→C→A — the compounding stacks three times.
Small numbers per leg, but they're multiplicative and they're free. A 6% leg margin run as a compounding round trip beats the same 6% run as two independently-budgeted one-ways, every time, and the gap widens with your capital. (This is exactly how ISK Scout's synergy board simulates its plans — each leg's buy list is optimized against the budget the previous leg's sales just topped up.)
Chains: when two hubs aren't enough
Sometimes the A→B margin is good and B→A is a desert. Before accepting the empty ride home, check the triangle: B→C might be strong, and C→A decent — a chain that keeps the hold earning on every segment even though no single pair justified a round trip.
The classic geography helps here: Dodixie, Rens, and Hek cluster within a modest loop of each other and back toward Jita's feeder routes, so triangle routes among the smaller hubs often out-earn the obvious Jita out-and-back on an hourly basis. The margins per leg look unimpressive; the sum over an hour is the number that matters.
Chains cost more attention, though: three shopping lists, three sell sessions, three chances for a market to have moved since you planned. Which is an honest segue —
When one-way actually wins
Round trips are not a religion. The return leg has real costs: buying and selling time (a market session per leg adds minutes that erode the hourly math on short routes), risk exposure (a loaded hold through a chokepoint twice instead of once), and stale-margin risk on the leg you planned an hour ago.
One-way is correct when the outbound margin is exceptional and the return board is genuinely bare; when the return route detours through danger the direct route avoids; or when your hold is specialized (an ore hauler's return cargo options are limited by the ship itself). The discipline is simply to check — flying empty by decision is strategy, flying empty by default is a pay cut.
A worked sanity check before any multi-leg commitment: total profit ÷ total minutes, versus your best simple route's number. Mode bonuses in your head ("chains feel efficient") don't count; the division does.
Reading a synergy board
However you plan multi-leg routes — spreadsheet or precomputed board — the numbers that deserve your eyes, in order: ISK per hour (the only honest ranking metric across modes), ISK per jump (fatigue-adjusted reality), per-leg danger (a round trip through trouble is trouble twice), and cargo utilization per leg (a "round trip" whose return fills 15% of the hold is a one-way with a souvenir).
Hourly rate first, always. A glamorous 120M chain that takes 2.5 hours loses to a boring 60M round trip that takes 55 minutes — and to noticing that in advance.
Recap — never fly empty (on purpose)
- You're paid per hour. Empty backhauls halve the denominator's earnings.
- A mediocre return load beats a zero return load — that's the actual comparison.
- Round-trip profit compounds: outbound earnings fund the return buy.
- Chains rescue routes where the direct backhaul is dead — at an attention cost.
- One-way by decision is fine; one-way by default is a standing pay cut.
- Rank all multi-leg plans by ISK/hour. Everything else is decoration.
Related posts
Other reads you might enjoy.

Reprocessing Report #5, August 2026: Ore Mistakes and the Day the Scanner Lied
August 2026 reprocessing report: 3.9B ISK in flagged spread, a month of mispriced compressed ore, Dodixie's 1-ISK capital hardware, and the 3.6-trillion-ISK glitch we cut from the books.

Reprocessing Report #4, July 2026: The Month Jita Learned to Repeat Itself
July 2026 reprocessing report: 6.4B ISK across 1,083 flagged trades, a 1B armor-hardener dump in Jita, and one Hek structure mispriced for 15 days straight.

Reprocessing Report #3, June 2026: The Small Hubs Run Assembly Lines
June 2026 reprocessing data across all five trade hubs: 1,127 flagged trades worth 5.6B ISK, a stasis-grappler assembly line in Rens, and two capital modules in Jita that paid 800M.