As part of its broader effort to transform PV economics and finance (www.rmi.org/solar_market_transformation_program), Rocky Mountain Institute recently published its second report on dramatically reducing balance-of-system costs in general and soft costs in particular. It's at www.rmi.org/simple.
1) I recommend using the term "variable" for resources with forecastable variations (particularly wind and PVs) and the term "intermittent" for terms with unforecastable forced outages (particularly large thermal stations).
2) There is emerging evidence that 80–90+% renewable scenarios may require *less* storage and backup than achieving the same reliable supply with large thermal stations, due to their less graceful failure (losing GW blocks in milliseconds, often for weeks or months and without warning). For example, LBL-5559e compiled many US utilities' studies showing typically ?5% balancing reserves required for high (?40%) wind penetration; the 80%-renewable, 50%-distributed, highly resilient "Transform" scenario in "Reinventing Fire" (www.reinventingfire.com), using NREL's ReEDS model, found bulk storage needs of only 5% of renewable capacity; and RMI's hourly dispatch modeling found ERCOT (Texas) could be 100% renewable with no bulk storage. How? Diversify variable resources by type and location; forecast; integrate with dispatchable renewables, distributed storage (smart bidirectional electric vehicles and ice-storage air conditioning), and demand response.
3) Modern variable renewables can often be forecast more accurately than demand: see e.g.www.renewablesinternational.net/wind-power-statistics-by-the-hour/150/505/61845/, Fig. 7.
4) Some parts of this otherwise very helpful post do not seem symmetrical in their treatment of the grid integration needs and costs of different kinds of variable or intermittent resources. Symmetrical comparisons remain very rare. But we must always bear in mind that large thermal stations too have integration needs and costs, including the canonical ~15-20% reserve margin and its spinning reserve element.