DNMs volumes should approach 0.1% of the global black markets by the end of 2018, I.e. roughly 25bn$/an.  In mid-2013 SR1 was turning over about $1bn/an.  Since no major DNM can risk BTC, ZEC, or ETH in the future, the only coin with the liquidity to support DNMs will be XMR.  Since pq=mv, and pq and m are known (estimated) factors, that leaves v.  If the velocity of XMR is roughly the same as SR1 era BTC, then at the end of 2018 XMR would trade at roughly 2500 $us, averaging over any periods of excess volatility.  I prefer to assume that either half of the DNMs are honeypots using BTC or ZEC, or that the velocity of XMR is more than twice the velocity of BTC in the SR1 era, in which case $1000 would be a better estimate.

Rational discounting in a negative real interest rate environment isn't very...relevant; thus, the only discounting factor is risk of model error.