On the one hand, if I take the lifetime payments we have made to SolarCity (which includes two months _before_ I got my i3) and add that to payments made to EverSource formerly known as Nstar from the time I got my i3 to today, I get a little under $2K. If I add up the previous year's Nstar bill for the same time frame, I get about $1500. So, you could argue that getting the solar panels was a Bad Deal, because it cost us between 400 and 500 more in energy -- if you took out the extra months in the SolarCity calculation, it's still over $300 more expensive.
On the next hand, if I look at the odometer in my i3 (5165 miles) and divide it by 30 (average mpg in a 2011 automatic Honda Fit, my previous vehicle), I get about 170. Multiply that by AAA's idea of the current national average for gas (which is lower than it was over the last few months), and I get a little under $500. That makes it look sort of like a wash -- I don't know what I actually drove in 2014, but gas was cost more, and I don't think I've been driving _more_ miles, so I probably spent at least $500 on gas last year (during the 8-9 month period under comparison).
But I think a more compelling observation to make involves this past winter.
It'll be interesting to see what happens next winter. If it wasn't so cold we would have run the furnace less. If there had been less snow, the panels would have produced more. But even with the panels completely unproductive and very high usage on the furnaces, we still came out about even on the total energy cost calculation. And that's being very conservative, because I probably spent more than $500 on gas during that period in 2014.
Certainly, the solar panels make it Real Easy to justify running the A/C during the summer. We've been running a big credit at EverSource, and I notice that they recently upgraded their "Delivery Charge", possibly to reflect the existence of people like us and the costs we put on the grid.