Interestingly (for me) ive just fount the correction chart for my hydrometer.
As mentioned, its calibrated at 20c. I was testing at more like 15c. The chart states that, at 85%+, at 15c, 1% should be added to the read out. Which, specific density errors between ethanol and methanol aside, would tie in with the 99% reading that i got from my 'virgin' sample.
I would therefore be happy to assume my recovered to be 97%, however, when im feeling keen, i will do as you say, and test by adding water to a 100% sample.