You're ignoring a factor in your long-term prediction. While MS may or may not have addressed all issues that can contribute to the RROD, they have taken steps over time to attenuate the problem: an additional heatsink has been added, the CPU is now contributing less to the total heat of the system, and they have improved the conductivity points regarding the heatsink.
If you were to actually have a reliable way to graph this, you would see the failure rate declining while the total failure count declines as well.
Unless you have hard data that shows how effective or ineffective MS' steps to improve the situation have been, when those steps where taken, how many of the older units were still on shelves when they were taken, etc., then you have no way to reliably predict what the total failure percentage will be over the life of the product. Common sense, however, would tend to dictate that the total failure rate is dropping and has been dropping for some period of time unless every attempt to improve the heat-sink issue has had zero effect, and I seriously doubt that anyone posting on this site has data that demonstrates such a thing.
Yes, you have a point, and I should have covered that a little more carefully. As I said, the newer 360s may be more reliable -- on the short term, at least, if not the long term -- but only time will tell if Microsoft's efforts have made a significant difference. If the new 360s are more reliable on the long term, then the failure rate for the console as a whole (for its lifetime) will drop, but the initial machines will still be hugely unreliable, and they will continue fail for the life-time of the 360.
Yes, we could assume that Microsoft's engineers have got it right this time. But, then again, I would have assumed they had it right the first time around, and I would have been wrong.
The same arguement you throw at me could be returned to you: Unless you have hard data to show how effective or ineffecitive MS's steps to improve the situation have been on the long term life of the console, then you have no way to predict what the total failure percentage will be for the life of the product. You state common sense says one thing, I'd argue that common sense says no such thing. Common sense seems to suggest we should be cautious, and wait to see what happens, not that we should trust Microsoft's console-building skills.
Just because many hope Microsoft have resolved the issue on the long term does not mean they have. However, give it a few years; we'll all know in the end.