More or less - there is always some variability from sample to sample (due to accepted variances in circuit, LED output bin, Vf, etc). And protected 14500s typically have slightly higher capacity than protected RCR, despite being rated the same (typically ~10-30% more in my testing).
At special request, I have now done the 17670 runtime on the QAA head (using the 123-2 battery tube). The result is interesting:
As you can see, you get much longer runtime than RCR or 14500, as expected. But the buck/boost QAA circuit doesn't outperform the buck-only Q123-2 on 17670.
The reason for this appears to be in the regulation - the Q123/QAA head is fully regulated on 1x3.7V Li-ion, but the Q123-2 drops out of regulation and into something approaching direct drive after ~40 mins in the 17670 run. Direct drive is always more efficient that flat-line regulation, so it's not surprising to me that the Q123-2 wins out on 17670.
When the 123^2 goes into direct drive (from too low voltage), will the different modes still work? Or would you lose everything except high, like when a Fenix PD20 goes to direct-drive on 16340 from too high voltage?