What we mean by a typical price, and when we refuse to give one

Every price on this site is compared against a typical price, and that phrase is doing a lot of work. This page is what is underneath it, because a number whose method you cannot inspect is a number you should not act on.

What it is

Every time our tracker sees a product listed, it records the price. A typical price is the median of those sightings for that exact product over the last 30 days. The median, not the average, because one mispriced listing at ten times the going rate drags an average somewhere no real club has ever sold, and does nothing to a median.

"That exact product" is stricter than it sounds. A product here is a model, a variant, a flex and a loft. A stiff 9.5 degree driver and a regular 10.5 degree driver of the same model are two products with two separate prices, because treating them as one produces a figure that is wrong for both.

When we refuse

A median of three sightings is not a market price, it is three numbers. So a product only gets a stated typical price once we have seen it at least 8 times at 3 different prices. Below that the page says we do not know, and the product stays out of the rankings entirely.

20%

have a typical price

The other 80% is tracked, listed, and not yet priced.

That is the number most sites would not publish. It says that for the majority of what we track we will not commit to a price, and we would rather print that than pad it out with medians of two.

Why the total keeps moving

The tracker runs through the day, so a product that was short of the bar this morning can clear it this afternoon, and both the typical price and the coverage figure above change as it does. Nothing on this site is written into the page. We have 1,754,300 recorded prices across 71,691 products and the figures are read from that when you load it.

What it is still not

A typical price is what a product has been listed at, not what it sold for. We can see every asking price on the open market and none of the closing ones, and anybody claiming otherwise about a used-goods market is describing data they do not have. Asking prices track selling prices closely enough to be useful and are not the same thing, and that gap is the honest limit of everything here.

We also do not use auction bids. A bid is not a price, and they are stored in a separate table so that they cannot leak into a median by accident.

Why the median and not the average

Worth spelling out, because it is the single most consequential choice in the method. Used-goods listings are full of mistakes: a decimal point in the wrong place, a set priced as a single club, a headcover listed as the driver. An average is moved by every one of those in proportion to how wrong it is, so one listing at ten times the going rate drags the figure somewhere no club has ever sold. A median does not notice it at all, because a median only cares about the order of the values, not their size.

The same logic drives how we measure spread. We use the middle half, the range the second and third quarters of the market sit in, rather than the distance from cheapest to dearest. The extremes of a used market are made of errors and optimism, and a range that includes them describes the errors rather than the market.

One offer, one observation

A subtler correction, and one we got wrong until recently. Our tracker records a price per listing, and bulk sellers list the same club many times over with the shaft or the flex reworded in each title. Counted naively, one seller with nine listings outvoted nine sellers with one each, and the typical price drifted towards whoever had the most rows rather than towards the market.

So a typical price now counts each independent offer once, where an offer is a seller at a price. Nine identical listings from one shop are one offer. Nine shops agreeing on the same price are nine, because that agreement is the strongest evidence a market gives you and throwing it away would be the same mistake in the other direction.

How to check any of this

Every product page shows the listings the figure is built from, grouped so you can see how many independent sellers are behind it, alongside the full price history we have recorded. If a number here looks wrong to you, that page is where to look, and if it still looks wrong after that we would genuinely like to know: a median skewed by one mis-parsed listing is a bug we can fix at the source, and several have been found exactly that way. Contact.

How the grading works, and why we do not grade against list prices.