Imagine a watch where you could replace the hands yourself, but it doesn't only display the date, hours, minutes and seconds, it displays milli-seconds as well. That's the equivalent of medium format vs small format.
DSLR's have been around for ages, the only thing you have to do is to put in a sensor and a CF card-slot, feed the whole camera from one battery and that's it. But that wasn't always the case, in the early days there were exchangeable backs with sensors for small format cameras- but it's much easier to just produce a compact, non-removable design, which is why most manufacturers do just that.
An as I already stated, medium format is a different beast because of the sensor size and the resulting magnification, smaller DoF etc. - if the focus is off by just a fraction you'll immediately notice and the photo will be out of focus.
With small format - not so much.
I'm not saying that small format (D)SLRs are all cheap trash, I'm just saying that the requirements and challenges for medium format are much higher.
edit: and of course changing the battery on an expensive watch (which is most likely waterproof) costs more. First of all the person who changes the battery always risks damaging the case (which is expensive to replace) and he/she has to exchange the water seals every time the battery is changed. It is a much more complicated task and takes longer too - that's why it's cheaper for a cheap watch and not so cheap for an expensive one.
After all a formula 1 race car is much more expensive than a Ford Fiesta and having someone change the oil or getting new tires will definitely be more expensive as well.
I have photographed with 3 different makes of MFDB backs for 10 years now, and I still default to my old 22mpxl back when it comes to purchase and upgrade thoughts. I paid a lot for it at the time..at the time when digital was sort of new in the SLR world (Canon 1Ds cost me $8500) I most likely will be going to Canon, but perhaps to Sony, maybe by the time, Nikon will have something, who knows?
Maybe Ricoh was onto something with the interchangable sensors. Unless Fuji Honeycomb pattern, or Foveon technology is not exploited into larger mpixels, I see no major reason to upgrade as a improvement across the board. When you get rid of the Bayer, and its limitations, maybe then we can see something new in terms of pixel to pixel gains. Look at the Light.co camera, sure it maybe noverlty in comparison at this time, but its innovative, as well as Lytro, THese are completely different an new technologies that ARE available for under $2K, they maybe matured and find the right marketing to triple the price...but there is nothing new in the MFDB except some ISO gains, LiveView improvemets.
Anyway, as long as some photographers are finding a way to get advertising dollars to pay for owning such gear in the $15-30K...my hats off to you.
If Doug or Steve would help me land such jobs that pay off this gear, I wouldn't have a problem to get in line to buy one :-)
But if you think getting a MFDB will land you such jobs...be prepared to have your a$$ handed back to you.
I have a feeling if and when I take the dive in the direction of the Sony camera, and adapt the best lens I can for my macro work, I will shake my head in regret on why I didn't do it sooner. I have mentioned this before, but its been a while....
When I had the Kodak dSLR 14,mp camera with one of the higher resolving lenses, the quality of file was easily there (larger files was something I was looking for, but at the time Kodak backs were 16mp).
At the time I thought it was the Kodaks CCD sensor, non-AA, etc/ISO-6!, and these did have much influence, but using the right matching lens is ever so important.