The 1/3rd each solution

Let's consider a simple thought experiment to see how we can run a grid using a lot of renewables: 1/3rd wind, 1/3rd solar and 1/3rd gas.

The chart below is a diagram of the way demand moves (more or less) every day.  Demand is shown the the pink line, solar supply by the yellow, and wind by the blue.  The vertical axis is demand or supply, the horizontal axis the time of day, starting at midnight.  These curves are schematic and don't represent actual data (how did you guess?)

Starting at midnight, wind produces more than demand.  In the absence of storage (we'll come back to that) wind output will have to be curtailed, because otherwise  the grid would burn out.  But then demand starts picking up as people get up, put on kettles and heating, run showers, iron their shirts, etc.  Now, solar is starting to pick up but the sun is still low in the sky, so there might not be enough supply.  The gas turbines need to be turned on.  Later in the morning, once again there is too much supply.  Solar will have to be curtailed.  At midday (ignoring daylight saving time) insolation peaks, and the supply of solar power to the grid starts to decline.  But demand doesn't peak until later.  Without rooftop solar (which is often netted out from official electricity demand data) demand peaks between 2 and 4 p.m., even as solar supply is diminishing.  So gas is needed to cover the shortfall.  By 11 at night, demand has fallen enough that wind can now supply it.

Source: me

Some points:

• If you assume away storage of whatever kind, then there is waste.  During parts of the day, output from renewables has to be curtailed (which reduces wind and solar farm revenue) but then during other parts of the day, gas has to be burnt.  If we could store the power from the surplus periods to release in the deficit periods, less gas would be needed.  To cover the afternoon peak, 4 hours of storage would suffice.  This is called "time shifting" supply.  
• The schematic (despite my poor artistic ability) represents relatively smooth supply changes.  How smooth depends on how broad spread (geographically) the grid is.  Your rooftop solar panels might not get sunlight when a cloud passes over your house, but for the grid as a whole, your drop off in supply is a blip, compensated for by all the other solar panels.  All the same, grid-wide, there will be short periods when supply drops or is excessive.  Storage will be better for these small fluctuations than gas, but gas combined with curtailment could still be used.  It just less efficient: storage works when supply is too high as well as when it is too low.  Gas only works when it is too low.  For these smaller fluctuations one hour of storage should be enough.
• As the cost of storage falls, using batteries for time shifting supply for the afternoon peak, instead of gas peaker plants, will become progressively more attractive.
• A 1/3rd wind, 1/3rd solar and 1/3rd gas model is what the US electricity market is currently transitioning to.   That will reduce CO2 emissions from power generation by 80%.  However, "fugitive emissions" (=leaks) of methane skew that calculation.  Over 20 years, methane is 100 times as potent a greenhouse gas as CO2.  If leakage is more than 3% of the gas burnt, then total effective CO2 emissions from a 1/3rd each model would be higher than from a 100% coal model--though, of course, coal still needs gas peaker plants, and mining coal produces lots of fugitive methane emissions.
• But the 1/3rd model is still superior, because as battery costs fall, it will be relatively easy to replace gas with storage.  Even with current battery costs, it now makes sense to have 1 hour of storage to back up renewables.  This will cover short duration fluctuations in renewable supply.  In 5 years' time, it will make financial sense to have 4 hours of storage, and gas peaker plants will then be reserved for exceptional supply or demand issues and for emergencies.
• Four hours of storage with a mixed wind/solar grid won't cover longer-term fluctuations in supply--for example winter vs summer with solar.  For that we'll need seasonal storage.  For now, gas will do.  Later on we can make synthetic natural gas for this purpose, using the Sabatier process.
• You don't have to accept that batteries are going to get ultra cheap.  Even without batteries, the 1/3rd each model will cut emissions, especially after 10 years, which is about how long it takes for methane to decay into CO2.  
• This is why grids everywhere are moving to some variant of the 1/3rd each model.  It's why gas is still seen as a "bridge" fuel.  From an investment perspective, the key question is how rapidly storage costs fall.  If they fall very fast, even gas peaker plants will become stranded assets.