You apparently haven't understood anything I've said. Put it this way: If I offered you 50-1 odds on your ability to roll three sixes in a row on a die, would you take it?
More like you don't understand what you're saying. This has nothing to do with your earlier "profits" rationale, which you cited as the
reason and proof positive that something has "poor value" if it's
profitable for the sellers. The fact that one makes a profit does not ipso facto mean it is poor value. That was a dumb argument, and easily disposed of.
Since you're no longer defending it, you've effectively discarded it.
The argument you're now trying to make is a different one -- that the odds of a lemon are so low that there's not much point in insuring against it. Unfortunately, you're just pulling improbable numbers out of thin air.
For one, I don't think that the probablity of something going wrong with your Mac during the warranty period is 1 in 216, or the chance of "rolling 3 sixes in a row" on a six-sided die. That's 215 in 216 Macs having no warranty-covered problems within its lifetime. Really? Even you don't believe that.
Even if just 5% of Macs (a far more reasonable number even at a guess) have a problem within the relevant period, that's "odds" of 19-1. If this were a bookmakers' and you were offering 50-1 odds that your computer would fail at a rate of 2% (do the math) -- then it would be a very good deal indeed.
The problem is you're not being offered "odds". The insurance "payout" is not $50 for every dollar you put in. Rather, the payout depends only on the severity of the repair job and the cost of replacement parts. They could be anything from below even returns (less than the cost of AppleCare) to better than even returns (more than the cost of AppleCare). This is not a function of the "odds" offered but rather a function of the extent of the repair job.
You oversimplify (wrongly), and confuse gambling odds with the type of odds at play in insuring an item. Once again, your analogy is completely inapt. The more apposite argument is: if you KNOW that the failure rate is 5%, 1 in 20, or 19-1, is it good value to insure, given that that you can buy a Mac 20 times and have it fail just once? That's a legit question, and most might conclude, "No."
The problem is, you DON'T KNOW the failure rate. Your argument independent of the actual rate of failure is thus based on pure guesswork, and nothing more. And this is the nub of your mathematical and logical fallacy: you're assuming before the event that you know what the probable failure rate of your machine is. Unless you're claiming to know the likelihood of component failure before the event, your argument is fruitless.
