Short answer, no, probably not.
It will likely be similar to Hawaii (290x) overall in some regards (performance probably the most important to most consumers), and similar to Tonga/Tahiti in compute per clock (but perhaps clocked so 870 is close to Tahiti and 880 similar to 290 vanilla); what you would expect from the performance (gk104/Tahiti) segment of the generation following Hawaii/GK110 on 20nm, only...without the benefits of 20nm because it wasn't ready (and hence 28nm).
Think GTX580 -> 670-680/760-770, only without the drop from 40nm to 28nm. 680 was around 20% faster than 580. 20nm should bring somewhere around 20-25% (or so) performance over 28nm, which all said and done (around 5% or so) is the speed difference between 290x and 780ti. Hence, it likely will perform (very slightly) worse but constructed with a more efficient (die size) ideal that it's predecessor. It would appear this means a very efficient gaming chip for 32 rops and high clocks. On 20nm this may have meant an efficient 48 rop chip (equal to something like Hawaii or greater per clock) at a relatively similar low clock...or iow instead of getting something similar to a shrunk GK110 but using the benefits/setup of maxwell (and low clocks but high transistor count of 20nm), we get a reworked design more similar to gk104, but with the benefits of Maxwell (and high clocks) with obviously higher power consumption and a higher segment than perhaps originally intended.
TLDR: 880 will probably be like 680/770 in design philosophy, but likely use the power of a high-end part (225w+) rather than lower (conceivably intended as a 680 part deux against Hawaii) perhaps as a direct result of 20nm being late.
Hope that makes sense.