How accurate are OSRS GP per hour estimates?

GP per hour is the output of a model

A headline rate compresses many inputs into one convenient number. For processing, the model usually multiplies outputs per hour by expected sale value, then subtracts inputs, tax and fees. For combat, it may add common loot and long-run expected value from rare drops, then subtract supplies and death-related cost. For recurring activities, it may divide a small amount of active work across a longer cooldown.

None of those models are inherently dishonest, but the assumptions must be visible. Two guides can publish different rates for the same method because one uses maximum actions, another uses an ordinary rate, one counts rare drops and the other shows common loot only. Compare the model before comparing the headline.

Price freshness is necessary but not sufficient

Current market observations improve an estimate, especially when a method consumes and produces tradeable items. They do not guarantee that the advertised quantities can be bought or sold at those values. A low-volume output may take hours to liquidate, while a required input can hit a buy limit before one hour of processing is supplied.

Use recent high and low observations, timestamps and volume, then apply the appropriate side conservatively. Buying every input at an optimistic low and selling every output at an optimistic high exaggerates profit. Include Grand Exchange tax on taxable outputs and test a representative batch before scaling.

Throughput usually creates the largest personal difference

A guide may assume a specific number of kills, trips or processing actions per hour. Your rate depends on stats, gear, route, banking, mistakes, attention and downtime. A player who achieves 70% of the assumed throughput will not earn the headline rate even if every market price is perfect.

Measure completed cycles and active minutes rather than guessing. If a method has a long setup, include it. If supplies are prepared in advance, decide whether preparation belongs in the comparison. Consistent definitions let you compare methods fairly; changing the clock for each method produces a flattering but useless table.

• Record actions, trips or kills completed.
• Record active time and unavoidable setup time.
• Track inputs actually consumed, not the ideal inventory.
• Use actual sales or a conservative liquidation value.

Rare drops make short sessions noisy

Many PvM rates use expected value over a large number of kills. A rare unique contributes a fraction of its value to every theoretical kill, even though a player can go dry for a long time or receive the drop immediately. The long-run average can be mathematically useful while still being a poor forecast for tonight's cash flow.

Show common-loot profit and unique-inclusive expected profit separately when the distinction affects the decision. A player funding supplies may care more about stable cash flow than long-run expected value. Deaths, failed trips and reclaim costs also belong in the personal sample, not outside the model because they are inconvenient.

Active profit and effective hourly profit are different

A recurring method can require ten minutes of work and then hours of waiting. Dividing profit by the active ten minutes creates a high active rate; dividing by the full recurrence period creates a different effective rate. Both can be valid if they are labelled. Neither tells you how many times the activity fits into the player's actual schedule.

Similarly, an AFK method may produce lower logged-in GP/hour but fit alongside another activity. Attention is a limited resource. LootPath lets players filter by intensity because a lower headline rate can have greater practical value when it matches the session they will actually complete.

Run a small profitability sample before investing

For processing, buy enough inputs for a short batch rather than an hour at maximum throughput. Record acquisition cost, completed actions, outputs, sale proceeds, tax and total active time. For combat, sample multiple trips, record supplies and ordinary loot, and keep rare uniques separate. One short sample will not eliminate variance, but it exposes assumptions that are impossible on your account.

Scale only if the method remains attractive after conservative prices and your measured pace. Retest after major price movement, gear changes or route improvements. A saved sample also helps explain why your personal result differs from a public guide without assuming either source is broken.

Compare money makers across more than one column

A useful shortlist includes membership, requirements, starting capital, intensity, risk, time to sell, common profit, expected profit and personal throughput. Eliminate methods that fail hard constraints before sorting by GP/hour. There is no benefit in ranking a method first when the account lacks the quest, gear, attention or cash to perform it.

Use the remaining rate as a range, not a single precise promise. A base case, conservative case and strong-execution case communicate more than an exact-looking number. The goal is to choose a repeatable method that funds the next account decision, not to win a spreadsheet comparison.