Linear regression: How much can a used SLR camera cost?


Since the Canon 5d Mark IV has just been released, the 5d Mark III will also be affordable. I was advised to pay €1,500 for a maximum of 30,000 releases, but if you look at the cameras on offer on eBay and the relevant forums, the price seems to be much higher. But what is the fair price? With sufficient data, this can be determined by regression.

First of all, let’s assume that the more shutter releases a camera has, the cheaper it becomes. 150,000 triggers is the expected life of the shutter release on the 5d Mark III, a new shutter release including labor costs about 450€ (not verified). But first, let’s take a look at what the data looks like. I have picked out the prices and releases of almost 30 offers from the various platforms, so this does not take into account how good the camera is on the outside or what accessories are included. This data is read into R and examined for its characteristics:

We have an average price of 1,666€ and an average number of triggers of 85,891. That’s a far from the €1,500 with 30,000 triggers. The median does not look much better either. Now we start the regression and form a regression model with the function lm and look at the details with the command summary(MODELNAME):

From this we can already form a function:

Price = -0.001749x+1816 where x is the number of triggers. So a camera with 30,000 shutters should cost €1,763.53, a camera with 100,000 shutters still €1,614. Let’s take a look at the plot:

As you can see, we have some outliers (by the way, not due to great accessories or maintenance etc.) that somewhat “distort” the model. I had already removed a camera with nearly 400,000 releases and a price of still €1,000.

Unfortunately, the number of triggers cannot be automatically read from the portals because it is always manually entered by the user. Otherwise, you could build a nice tool for each camera model.

Leave a Reply

Your email address will not be published. Required fields are marked *