I wonder if I'm the only person in the world who analyzes extended warranties on electronics like prop bets.
I'm buying a laptop for my wife. The extended warranty cost is a 1-to-16 bet against the price today of the laptop.
So, i sit and think: "If I were to buy this laptop 16 times would at least one of them fail in the second year? Third year?"
I have concluded that it's a bad bet, for two reasons: (0) Companies employ acctuaries to analyze these situations with tons of data to ensure it's a bad bet, and (1) when I factor in the odds that I can repurpose a partly broken laptop (e.g., if only the wifi subsystem fails) into a useful task anyway, there is added value that isn't inherent when I consider "broken to point of unusable". I'm not sure how to quantify its numeric effect on the odds, but I suspect it puts well out of the 1-to-16 per-year area that the company is offering for warranty pricing.
But yeah, I don't buy extended warranties. I try to make sure there's no invaluable data on the machine, so that it is trivial to chuck out and replace if need be.
For laptops that we loan out to students at our library, however, we buy extended warranties that cover damage due to drops, etc. This protects the students from heinous charges when the inevitable accidents happen.