Great start, but one important factor is missing from this analysis: The surface area (for heat exchange) of a closed-loop borehole is tiny compared to the surface area in a natural hydrothermal system. The walls of the borehole will cool off relatively quickly due to the low thermal conductivity of rock, and you'll soon be unable to make power. Conventional geothermal gets around this by pulling from very large reservoirs that include natural convection and huge fracture surface area.
This means that a closed loop system needs many many many miles of boreholes in order to last long enough to pay off.
You may be right. Back of the envelope kinds of computation though, 6 miles of 12" diameter bore hole is 99,525 sq ft of surface area for thermal transfer. If you think of the lower half of that as the heat injection surface (so 3 miles -> 6 miles and then back up to 3 miles on the way out) call it 100k sqft of rock surface. Now some people have spent a lot of time studying rock (https://pubs.usgs.gov/of/1988/0441/report.pdf) I'm not one of them (more of a casual geologist because I like hiking in the mountains) but if I'm reading the heat transfer equations from pg 91 of that report it still seems like you'd get a decent amount of heat transfer (order of 10's of MJ) into the water.
This means that a closed loop system needs many many many miles of boreholes in order to last long enough to pay off.