If the query string is being passed through, which I suspect it is, you can use the query string to easily locate the corresponding entry in your own logs. Or, if the query string isn't being passed through, you can use a path instead:
You probably already thought of this technique. I decided to post it anyway in case you hadn't, or in case anyone else is facing a similar challenge.
A unique sequence of legitimate requests might be more difficult to for the other side to detect and it won't result in 404s. Could randomise the sequence and each can come from a different IP as long as they were synchronised properly.
He very well might. But my estimation of the thief's skills is low. I could be wrong, of course.
> A unique sequence of legitimate requests might be more difficult to for the other side to detect and it won't result in 404s. Could randomise the sequence and each can come from a different IP as long as they were synchronised properly.
That's probably the best bet. A legitimate but very winding path through actual links on the site would work quite well. Given enough steps, it would almost certainly be unique. Because you'd be varying the path each time, the thief would find it hard or impossible to block you.
(for now)
Put a few more branding elements into your page design, so that visitors to their site will understand that they're redirecting your data. Then take advantage of the extra publicity that they're giving you.
Second idea: Javascript redirect all of your pages to your own subdomain. Again, its just a step in an arms race, but this would be a little too hard/expensive to take to court. You can win an arms race if you try.