I believe in business it is called a cost / benefit analysis.
We don't always realize we are doing it but we do.
There are two types of overclockers:
1) People who overclock to get the fastest everything
2) People who buy slow and overclock to extend the value of their dollar.
Most people here are type 2)
Personally, I do benchmarks in the applications I use most where performance matters to me (specific games I play mostly) and my testing has shown that at the resolutions I play at I simply cannot get better MINIMUM frame rates by increasing RAM speed or CPU speed. In some cases there was zero difference at all for a 700 MHz overclock becuase it's a video card limited situation. So it is totally pointless for me to have spent any more money than I did (3000+ @ 2.5GHz). In most cases I can't even tell the difference between 400 MHz of CPU speed because I'm in a video card limited situation.
Take a look at this example for BF2 by the Inquirer, this is typical of what I see:
http://www.theinquirer.net/?article=23973
scroll down to 'memory and CPU' see how 1.6 GHz and 2.4 GHz are almost identical except for one part where the frame rates are already well above 70 FPS. That is the kind of thing I see when I do FRAPS, except I usually set my video settings so my minimum in closer to the 40-45 FPS range. Faster is only gaining in the highest FPS areas. I don't care if I get higher framerates in the higher FPS areas, I want higher framerates in the lower FPS areas, and I find that does not typically come with CPU and memory speed improvements, but with graphics card improvements.
Anyway you can get very good memory for near the price of value RAM that you don't have to compromise much on RAM performance at all. Crucial generic RAM is failed Ballistix and overclocks VERY well. Twinmos is often UTT that will run 2-2-2 with enough voltage and 2-3-2 1GB sticks are less than $120 per gig. There is just no longer any reason to buy the expensive RAM anymore. You don't lose 2-3% you lose less than 1% because cheap RAM overclocks to within inches of the performance of the stuff that costs twice as much.
The bottom line is that that extra couple percent is totally expendable in most cases because you can get a $150 A64 in the 2.5-2.7 GHz range at which point it's overkill for most applications.
I know that not everyone here is a gamer, but many are. The point is that we evaluate our needs, comparing the cost in $ to the performance gain we will get. In most cases cheap and OC will easily prevail over adding $50 more for a better processor, or $50 per gig more for better RAM. However, if you feel that the cost / benefit analysis gives more benefit to the more expensive parts, then buy the more expensive parts. There is no ONE answer that fits every situation, but for most enthusiasts even, it doesn't make a whole lot of sense to get a (maybe) 3% performance improvement for $50 per gig. Much of the overclocking and upgrading industry is hyped up about what amounts to very small overall improvements in real world applications. With the recent DDR500 article on AT, finally it is getting some pretty widespread publicity.
