Galaxy S5 mobile phone, claims to support 802.11 a/b/g/n/ac VHT80, MIMO(2x2).
Ah, that English language again. Too many pronouns. I was speaking in the statistical sense, that is that with a relatively small population of tests sets the standard deviation is a bit wide and therefore so is the confidence interval. It's like when political pollsters say 25% support a candidate, +/- 3%.
Yes, a raw 10% increase in range would be useful if it could be verified with a narrower confidence interval.
I would go further and say the the features of the newer hardware and firmware support beamforming. I was hoping that better microcircuit techniques would actually provide better radio S/N ratio and thus effectively improve the range even before new tricks were added.