Repair or Replace Calculator
Compare a repair bill against a replacement on cost per year, and see how often the familiar 50% rule gets it wrong.
"Years that buys" is how much longer you expect the thing to last after the repair, which is a guess and the figure the answer turns on most. If you are unsure how long an appliance should last in the first place, the lifespan planner has typical spans. Nothing is uploaded.
The 50% rule never asks how long the repair buys
The familiar advice is to skip a repair costing more than half the price of a new one. It compares two prices and stops — leaving out the only thing that makes them comparable, which is how many years each buys. Every case below is one where following the rule gets it wrong, and they fail in both directions.
| Repair, as a share of new | The 50% rule says | Cost per year | Which is actually cheaper | |
|---|---|---|---|---|
| Washer, small fix | 44.4% | repair | $100 vs $75 | replace |
| Washer, big fix | 66.7% | replace | $75 vs $75 | repair |
| Fridge | 29.2% | repair | $117 vs $86 | replace |
| Car, running costs counted | 20.0% | repair | $1500 vs $1240 | replace |
A $400 repair buying four years is $100 a year against $75 for a machine that lasts twelve — cheap on the sticker, dearer to live with. A $600 repair buying eight years matches the new machine exactly, and at 67% of its price the rule turns it down. These four were chosen to disagree, so they show the rule can fail and say nothing about how often.
How often it is wrong, measured
Across every combination of repair bill, years bought, replacement price and replacement lifespan in plausible ranges — 12,960 of them, with no selection at all — the rule disagrees with the cost-per-year test 22.0% of the time. About one case in five.
| Across 12,960 combinations | Share |
|---|---|
| The rule says repair, and repairing is dearer per year | 11.9% |
| The rule says replace, and repairing was cheaper | 10.2% |
| Disagrees either way | 22.0% |
Which is the part worth noticing: the two kinds of mistake are almost the same size. If the rule failed mostly in one direction it would at least be a safe default — err towards replacing, waste some money, never get stranded. It errs both ways roughly equally, so there is no bias to lean on and no sense in which it is the cautious choice.
The number the rule should have been
There is a largest bill worth paying, and it is not a share of the new price — it is whatever makes the repair match the new machine's cost per year. For the first washer above that ceiling is $300: pay a dollar more and repairing becomes the dearer choice, a dollar less and it is the cheaper one. The calculator works it out for whatever you put in, which is more use than a percentage, because it moves when the running costs and the trade-in value move and a percentage cannot see either.
How to use
- Enter the repair bill and how many more years it buys.
- Enter what a new one costs and how long it should last.
- Add running costs and any trade-in value.
- Read the cost per year for each, and the largest bill worth paying.
Frequently asked questions
Is the 50% rule reliable?
No. Across 12,960 combinations of repair bill, years bought, replacement price and replacement lifespan, it disagrees with the cost-per-year test 22% of the time — about one case in five. It compares two prices and never asks how long either one buys, which is the only thing that makes them comparable.
Does it at least err on the safe side?
It does not, and that is the part worth knowing. The two kinds of mistake are almost the same size: 11.9% of cases where it says repair and repairing is dearer per year, and 10.2% where it says replace and repairing was cheaper. A rule that failed in one direction would at least be a usable default.
What should I compare instead?
Cost per year of service. A $400 repair keeping a washer going four more years is $100 a year; a $900 machine lasting twelve is $75. The repair is 44% of the new price, so the old rule waves it through, and it is the more expensive option.
Can a repair be worth more than half the price of a new one?
Easily. A $600 repair buying eight years works out at $75 a year, exactly matching a $900 machine that lasts twelve — and at 67% of the new price the rule rejects it. What matters is the years, not the share.
What is the most I should pay for a repair?
Whatever makes the repair match the new machine cost per year. The calculator works it out for your figures. It is more useful than a percentage because it moves when running costs and trade-in value move, and a percentage cannot see either.
Should I include running costs?
Yes, where they differ. A new appliance that uses less power or fuel earns that difference every year, and it can change the answer on its own. A price comparison is blind to it, which is one reason the old rule drifts on cars and heating equipment in particular.
How do I know how many more years the repair buys?
You do not, and it is the weakest input here. Work backwards instead: the calculator gives the largest bill worth paying for a given number of years, and it is easier to judge whether four more years is plausible than to guess a total lifespan.
Does this send anything anywhere?
No. Every figure is computed in your browser, and nothing is uploaded or stored.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.