What mileage does to a classic car's value

The obvious page would state a coefficient: so many dollars off for every thousand miles. We are not going to state one, because the data cannot support it. Of the sales behind this site, 25,963 carry no usable odometer reading at all: a number with no unit, or a zero that means "not recorded" rather than "never driven". Adjusting a median with a coefficient derived from the other 6.6% would be inventing a number and dressing it in arithmetic.

There is a second reason. Mileage does not act on a market as a constant. On a 6,000-mile car the first 20,000 miles matter enormously; on a 200,000-mile driver they barely register. A single coefficient applied across the range would be wrong at both ends.

Here is what can be measured honestly. Take the sales in a model that carry a mileage reading in miles, sort them by miles, and compare the lowest third with the highest third. On the 12 models with enough such sales to split, the highest-mileage third sold for between 24% and 95% of the lowest-mileage third.

Read the direction, not the percentage. The direction holds on all 12 models here: the high-mileage third sold for less than the low-mileage third every time. The size of the gap does not hold, running from 24% to 95%, and it is not a coefficient, because the two groups differ in more than mileage — a 40,000-mile car is usually also the better-kept car, with better paint and a fuller history file, and those come with the low reading rather than being caused by it.

The confound column is where a naive reading goes wrong. If most of a model's low-mileage cars are priced above its own 75th percentile, they are not ordinary examples with fewer miles — they are a different, more valuable variant that happens to be sold with low miles. The Porsche 991 shows it plainly: 61% of its low-mileage third sits above its own 75th percentile, because the low-mileage cars in that set are GT2 RS and 911 R specials rather than Carreras. Set those variants aside and the 991's split moves from 24% to 67% on 56 sales. A page quoting the raw 24% would be telling you low mileage roughly quadruples a 991, which is not what it does.

Setting variants aside is not a cure, and the Porsche 964 shows why: even after removing the Singer cars, the Turbo and Carrera Cup models still dominate its low-mileage group, so its controlled figure stays at 26%. The remaining gap is still mostly variant, not odometer. That is the honest state of the data: mileage moves price in the direction you would expect, the size of the effect varies by model, and a single number for it cannot be extracted from this set without controlling for which cars get driven.

On a collector car, mileage is evidence rather than a price input. A low reading supports originality: original paint, an unworn interior, an engine that has not been apart. A high reading is not disqualifying — a well-maintained 150,000-mile car can be the better buy than a 30,000-mile car that has sat — but it narrows the buyers who will bid.

The two questions that matter more than the number are whether the mileage is documented, and whether it is consistent with the rest of the car. A 12,000-mile car with a worn driver's seat and a replacement engine has a story to explain, and that story is worth more to you than the figure.

Each model page shows mileage in its sales table wherever the auction record carries it, and leaves the cell blank where it does not — a blank is a record that did not state the mileage, not a car with none. The split by body style and transmission is on the same page. Start from model values, or read how the figures are computed.