My best guess:
By the time the race is over, the tires will be red, which will cost a roughly 1.5 seconds/lap, based on a guess of 1 second per minute of laptime. Assume a pit about lap 15, and that there is no loss until lap 15, at which point it begins linearly approaching 1.5 seconds/lap. Assume a mean loss of .75 seconds per lap for the last ten laps. That's 7.5 seconds. An average pit stop loses roughly 20-30 seconds at Laguna, I'm guessing? Assuming you can keep from stuffing it really badly, running without a pit stop would be a far superior strategy.
The above is just my best guess, and is probably fairly conservative in its estimation of time lost from bad tires. However, using the above concept of (mean X 10), the average loss would have to be 6 seconds per lap if the pit stop costs 30 seconds. It would be four seconds per lap at 20 seconds. I doubt that the cost is anywhere near that.
Also, if anyone has a best guess about a non-linear effect of tire wear, or any real-world experience with this sort of thing, I'm very interested. This is me taking a break from game theory for five minutes to think about something else. Sorry if it's a bit garbled.
