The 15" cMBP wasn't selling that much compared to the 13" cMBP and MBAs, mainly because it was too expensive at its $1,800+ price point. Around 15% of Mac notebooks sold were 15" MBPs. Then Apple released an even more expensive model, the 15" rMBP, starting at $2,200+. It sold out extremely fast. The day of the announcement, shipping dates slipped to 2-3 weeks to later slip to 3-4 weeks. That usually doesn't happen, even for much more popular models. This shows how low the initial supply was.
According to analyst Ming-Chi Kuo (that guy that was right so many times this year), the reason the release of the 13" rMBP was pushed back to October was due to even lower yields with the 13" Retina display. He later reiterated that 13" rMBP yields were going to be low:
[...] However, as supply is limited due to panel and assembly yield rate issues, shipment of the 13 Retina MacBook Pro wont be able to satisfy demand. And we think consumers budgets will transfer to new iOS products after Apple launches them, which suggests that overall MacBook shipments will not grow meaningfully on the 13 Retina MacBook Pro.
Demand for the 13" cMBP was around 3 times that of the 15" cMBP, so if you extrapolate that trend to rMBPs and factor in the fact that yields of 13" rMBPs were going to be even lower, you do the math: it was going to sell out super fast. And how can you make more money when you know sales are going to be limited by supply? If you know demand is going to be strong anyway, you just raise the price. Basic supply and demand. We know by comparing rMBPs to their respective cMBPs that Apple makes around $400 more in profits on a 256GB 13" rMBP than on a 15" rMBP, which is just insane, that's a whole 20% more in profits. It's like Apple doesn't want to sell too much of them, and that's probably what it is. They know they can't produce much of them anyway.
My hypothesis is that the 13" rMBP's price will be reduced once Retina display yields are good enough. Obviously not mid-generation though, that would generate too much negative reception, but rather when the next specs bump (Haswell) occurs. By the time that happens, they'll be able to produce lots of 13" Retina displays in advance given low sales of the 13" rMBP at $1699+, and be prepared for the 13" rMBP to actually sell.
tl;dr: Ming-Chi Kuo says yields of 13" rMBPs were way too low for demand. He's always been right so far so it's probably true. Apple had no point in pricing it aggressively since it would be limited by supply anyway. Expect a better price next year when yields are good.
According to analyst Ming-Chi Kuo (that guy that was right so many times this year), the reason the release of the 13" rMBP was pushed back to October was due to even lower yields with the 13" Retina display. He later reiterated that 13" rMBP yields were going to be low:
[...] However, as supply is limited due to panel and assembly yield rate issues, shipment of the 13 Retina MacBook Pro wont be able to satisfy demand. And we think consumers budgets will transfer to new iOS products after Apple launches them, which suggests that overall MacBook shipments will not grow meaningfully on the 13 Retina MacBook Pro.
Demand for the 13" cMBP was around 3 times that of the 15" cMBP, so if you extrapolate that trend to rMBPs and factor in the fact that yields of 13" rMBPs were going to be even lower, you do the math: it was going to sell out super fast. And how can you make more money when you know sales are going to be limited by supply? If you know demand is going to be strong anyway, you just raise the price. Basic supply and demand. We know by comparing rMBPs to their respective cMBPs that Apple makes around $400 more in profits on a 256GB 13" rMBP than on a 15" rMBP, which is just insane, that's a whole 20% more in profits. It's like Apple doesn't want to sell too much of them, and that's probably what it is. They know they can't produce much of them anyway.
My hypothesis is that the 13" rMBP's price will be reduced once Retina display yields are good enough. Obviously not mid-generation though, that would generate too much negative reception, but rather when the next specs bump (Haswell) occurs. By the time that happens, they'll be able to produce lots of 13" Retina displays in advance given low sales of the 13" rMBP at $1699+, and be prepared for the 13" rMBP to actually sell.
tl;dr: Ming-Chi Kuo says yields of 13" rMBPs were way too low for demand. He's always been right so far so it's probably true. Apple had no point in pricing it aggressively since it would be limited by supply anyway. Expect a better price next year when yields are good.