I'm trying to be my very own Devil's Advocate here so forgive what seems like a bit of a troll thread title. I'm just very interested in the opinions of veteran overclockers on here. So here goes...
If I decide to upgrade my PC in the next year or so I will buy parts that are of decent quality and speed. Not the top - bleeding edge - stuff but good nonetheless. My point is that if I do this, is there any reason for me to overclock my PC given that it will most likely run everything perfectly anyway? And, if overclocking it gives me no noticeable differences in-game then is it all just a numbers game for benchmarking enthusiasts? Given that stressing anything out over and above the norm' will shorten it's life and perhaps bring hidden issues to the fore wouldn't it just be better to save that extra cash and buy faster equipment? Like I said, this isn't a troll, just and interested party.
If it helps you to do something that you either need or want to do, yes.
As far as buying a faster computer, try to buy a stock 4.5 GHz system. The only way you can is to buy a system that someone else overclocked.
I have overclocked Q9550 and Q6600 systems. They are still fast enough to do and play what I want. And they are still fast enough to let me wait for Ivy Bridge.
As far as reliability is concerned, if you stay inside recommended voltage and thermal limits, except for a random fault which all electronics are subject to, the system will most likely be obsolete long before it fails for overclocking side effects.
My best example of that: in 1998, I built a Pentium 233MMX system. Eventually (around 2002), I put a better than stock Socket 7 cooler on it and found an undocumented set of motherboard jumper settings that let me run at 333 MHz. Turned it into a DOS/Win98 play system. I discarded the system in 2009. Couldn't give it away.
I have been doing this for quite a while and I have never had a system fail for overclocking effects.
---------- Overclocking since 1978 - Z80 (TRS-80) from 1.77 MHz to 2.01 MHz