Warranty and Repair Reserve Calculator

Two thousand units, three per cent coming back, eighty-five dollars a repair. The average year costs five thousand one hundred.

Reserve that and you are short in about half of all years.

Not through bad luck — that is simply what an average is. Failures vary from one year to the next, and half of all variation is upward. The figure you actually want is one that covers a bad year, and how much more that costs is computable rather than a matter of nerve: here it is six thousand one hundred and sixty-seven, which is one and a fifth times the average. Rather less of a premium than most people assume for the difference between usually solvent and usually fine.

Per unit sold that is three dollars and eight cents instead of two dollars fifty-five, which is the form you need it in, because that is the number that goes into a price.

Each unit either comes back or it does not, so the count of failures follows a binomial distribution and its spread is calculable rather than guessable. That spread is the whole reason this is a decision rather than a shrug.

The page offers three levels of cover — ninety, ninety-five and ninety-nine per cent of years — priced side by side, so what you are actually choosing is how often you are willing to be caught short. Ninety-five per cent means about nineteen years in twenty. The twentieth year still happens. No reserve computed from a distribution can promise otherwise and this one does not pretend to.

And it checks its own working. The normal curve stands in for the binomial here, which is fine when you expect at least about ten failures and at least about ten non-failures. Below that it stops describing reality — so the page says so and refuses to give you a number, rather than handing over a confident figure from a model that has quietly stopped applying. Sell eighty units at three per cent and it will tell you to argue the reserve from your worst case instead.

There is no industry-standard percentage anywhere on it. The reserve falls out of your failure rate, your repair cost and the spread, or it is invented.

Save as many versions as you like and put any two side by side, figure against figure, with the better one on each line marked and the difference in its own column. That screen is the one that decides things, and it is the part a free calculator cannot do at all. Failure rates differ enormously between products, and a single reserve across everything is right for nothing.

One HTML file. No account, no subscription, no internet needed. Opens on Windows, Mac and Android; the readme explains the extra step on iPhone and iPad, where Apple blocks JavaScript in downloaded files and every tool of this kind is affected.

Everything lives in your browser on your device. Nothing is uploaded, which is why it works offline and why nobody else can see your figures — and it also means clearing your browsing data deletes it. The tool does not pretend otherwise: the header shows how long since your last export at all times, turns amber after three days and red after seven, and export is one click that produces a plain file you keep.