cost of speakers in relation to the rest of the system

My price ratio is always going to be a little ridiculous because of my resource limitations and inexplicable urge to learn, build, test, and experiment. At some point I’m going to dig into cables and build better versions of my 11.5g 6 way round braid. I only built these to learn the difference a cable design can make. And a difference it certainly makes. One of the things that I kept reading that reaffirmed my purchase of the Focal 936 is that they have a tendency to exploit the best gear you put behind them. And exploiting a Pass F5 they certainly are! The thing with building stuff is I can build a stack of Pass designs, hunt down a good DAC, make some pretty nice cables, tweak every component to perfection, and still not spend what I did on those speakers. Speakers, if you ask me, are way too complicated for me to build. I have no tools to make and finish cabs. I have no supercomputers to model how they behave. I don't have the money to buy piles of caps, resistors, and inductors to voice the things with. I see a tremendous value in buying well designed speakers because, unlike an amp which if you just clone well upholds the design specs, it takes tremendous resources to design drivers, cabs, and crossovers to sound like anything good. $4000 buys you and obscene amount of quality and R&D in a speaker these days. Minor alterations to cloning a speaker design can destroy it's sound. 

kosst_amojan makes an excellent point: Designing and building good speakers is insanely difficult. In current dollars my guess is you'd need to spend $5,000 retail for something truly great, which would represent 50% of a decent system, however you can get thirty-year-old speakers that sound incredible for a fraction of that thanks to the fact that: There are minimal electronics inside speakers (so easily serviced/upgraded); speaker design/construction become mature technology by the eighties (thanks to a new design hegemony of aiming for a flat frequency response instead of trying to "voice" them, along with breakthroughs in driver construction); and, not least, audiophiles are collectively insane so the second-hand market is flooded. My point being: Buy a superb-sounding, twenty-year-old pair of Vandersteen 2Cs for $600 and all of a sudden the speaker to amp/source ratio goes crazy. There are no hard and fast rules to this.

These ratios are at best guidelines and will vary depending upon overall budget as well.  Ratios at $20-50k are generally going to be more speaker leaning in terms of total percentage than a $200k system.  There's also a general (incorrect) belief that good speakers can somehow fix shortcomings in front end and amp, which is in my experience far from the case.  Ultimately, unless the system builder is relatively new (and needs a helpful guide as a baseline) cost allocation is not a consideration that will be relevant; rather, it should be building the best system possible, and that is almost always a system where the relative performance of all components is similar since one cannot "fix" shortcomings in a poorer (relatively speaking) component and such component will then hamstring and limit the ultimate potential of the system.

@kosst "inexplicable urge to learn, build, test, and experiment."  +1.  I lack the confidence and skill to try to build anything electrical, even from a kit with clear instruction.  And I have a weird fear of electricity.  So, "mad respect," or whatever kids are saying on tv.

How often does, "I just like these speakers.  I know they aren't the best at X, but these make my ears happy" come into play?  I think some folks alluded to it.

I have an odd fascination with Infinity Primus 150s.  To me, they are one of the best bang for bucks I've run into.  I think they were about $100 a pair.  Bought four new pair years ago, maybe thought I'd use them in a 6.1 second multichannel (pretty sure at least one box has never even opened).  Still have them.  Still occasionally pull a pair out of the closet.  Still impressed.  Had a pair in a small system for several years and probably represented about 5-10% at the height of trickle-down equipment.  Just sounded more "right" to me than several others that cost several times more.

Let's not forget that many high end expensive speakers cost at least ten times what they cost to make. Ten times, think about it. I never heard this about electronics, this doesn't mean there aren't any. Those hypothetical $50k speakers are not that much high-fi as anyone would expect. No wonder that to get more or less realistic sound in a big room, especially with big orchestra playing, one has to go over $100k. For medium size rooms, as many said, forget about this level of realism, there is no need for $100k speakers. Yeah, $50k should be enough. I'd say, buying all new, that speakers' cost would be something like 30% - 35% in a $150k set-up.