Agreed, That is the best way if the application will allow it. Most buck drivers have a current sense resistor which can be modded to give even higher currents. Efficient buck circuits can give you almost a flat lumens output curve through out the entire run time(constant current).
Did you bypass the battery springs? What kind of DMM are you use, and what kind of leads?
Just a note. The zy-t08 has a good switch, its resistance is lower than 5milliohm (<0.005ohm). Usually a DMM self resistance (with leads) is between 0.05-1ohm. You can imagine how big will be the current difference through the DMM and through the switch. My DMM has 0.03ohm with thick and short leads, this is not even good enough for accurate measurement, at least at this light. The measured current will be lower in any case, than the real current when the light works with its own switch.
I have an Solarforce M8, with QLite driver, stacked chips totalling 16, pushing an XM-L2 at 4.90A using a rested Samsung 20R cell. I’ve seen it hit 5.34A with a fresh charge.
Thank you all for helping me to understand a bit more of this dark art 
The springs are braided and the wire changed to 22g.
My DMM has heavy duty speaker wire leads only 8” long, and measures 6.5A on my Kung.
Those figures are close to what I got so,
As I am going to buy a new driver anyway going to 2S will only be a different driver and a bit of dremel work J)
Cheers David