Take a conventional (coal, gas, oil, hydro, nuclear) power system comprising multiple power stations and, assuming it's properly designed and run, it will have a definite "firm" rating for peak power which can almost certainly be supplied.
It comes down to probability. If you have 100 generating units in a system then the chances of having all 100 in service at any one time are pretty unlikely. Even if they are inherently reliable technology (hydro) and you have 60 of them well then suffice to say that it's pretty rare to have them all actually working at once. Major overhauls, shorter term maintenance works, breakdowns, capacity limitations for various reasons - you can't depend on the system ever working to 100% of its' nameplate capacity.
For example, here in Tas the Paloona power station is currently shut down for major works that will take about 12 months to complete. Wayatinah is offline at the moment too, for about 2 months. Rowallan has been out since November (about to be returned to service) and once it's back then Mackintosh will be taken offline for a few weeks. Meanwhile there are countless much shorter outages for inspections etc. It's not unknown to have every machine in service, but it's certainly not the normal situation to be doing that in a properly maintained system.
But you
can have an extremely high probability that a specific, lower level of capacity can be achieved and it comes down to how often you accept failure. If you need 100% to work then it's going to be a problem quite often. But if you only need half the system to work, because it's massively over-engineered, then you're unlikely to see a single problem with meeting demand in your lifetime.
If we take that a 1 in 20 year failure is acceptable, then realistically you'll be able to depend on about 80% of the installed capacity. So if you have a peak load of 10,000 MW then you need 12,500 MW of capacity to do it. That's pretty much the standard that used to be followed in most states. If you tighten it to a 1 in 50 year failure then you'll still be able to rely on somewhere around 75% of the installed capacity.
So to meet the roughly 10,500 MW of demand in Vic then realistically you could do that with incredibly high reliability using about 13,100 MW of installed capacity based on coal, gas (with oil as backup) and hydro.
Oil as backup? That's because no sane person would rely on the gas supply. There have been plenty of problems historically around the world with gas supply disruptions and it comes down to not being able to store the stuff economically near the point of use. Once something goes wrong upstream, well then you lose fuel supply to the power station such that gas supply can be maintained to other users.
These days in the "competitive" market nobody worries too much about this, if the gas goes off then so too will a decent chunk of the power supply, but certainly in cases where reliability is considered important then you need a backup fuel (almost always oil). So if you look at, say, Tamar Valley power station (Tas) then the largest unit is gas only but the 4 smaller ones have oil as backup fuel. There's a simple logic there - failure of a single unit is always possible for whatever reason, be it loss of gas supply or something else, but losing multiple units due to a single point of failure (ie gas) creates a far bigger risk to the power system. Hence oil as backup and it's the same with the older gas-fired plants in other states - they had backup fuel when the former SECV etc was building them.
But what about wind? Well the problem is simply this. It may well be predictable but it is by no means reliable. We can get 8,000 MW from a 10,000 MW thermal / hydro system with reasonable certainty. But there's no guarantee that you'll get much at all from wind since, in practice, multiple wind farms completely fail at the same time due to a single point of failure - the wind itself isn't reliable.
If the forecast says that we'll get 100 MW from wind in SA tomorrow then we can't simply issue a formal direction to postpone the outage of natural wind and get another 500 MW of generation online. Nor can we build up a stockpile of wind at wind farms to cover such variation. Nor can we arrange interruptions in wind to occur in Autumn and Spring or on weekends. In short, whereas a thermal / hydro system can reliably operate to around 80% of it's nameplate capacity for wind it's roughly 8% and there's the problem. Building 100,000 MW of wind to run Victoria, generating on average 6 times the state's total power requirements, just to get some reliable peak capacity doesn't really work out too well financially, practically or environmentally.
Which brings us to storage. Hydro, batteries, compressed air, heat, hydrogen, whatever. If we're going to use intermittent sources of energy to supply most of the load then (1) there needs to be some over-engineering on the supply side (ie at times wind energy will supply 100% of the load with the rest going to waste) and (2) storage.
I flew over Tas today, with a nice view of the entire Mersey-Forth scheme (Rowallan, Fisher, Lemonthyme, Cethana, Wilmot, Devils Gate and Paloona power stations) plus part of the Great Lake scheme (Great Lake itself and also Lake Augusta) and a bit of the Derwent catchment too (Lake Echo). Now, this was a commercial flight by the way (Melbourne - Hobart) but I don't need to look at any lake level data to see that there's energy sitting there in storage. Simply looking out the window it's clear where the water is and isn't. Likewise a stockpile of coal, or a mine set up ready to quickly extract it right next to the power station, is also energy in storage. Likewise a tank full of oil or even a pile of firewood.
But we've got no wind in storage whatsoever. Right now Woolnorth is operating at 85% of capacity but Musselroe is only running at 7%. Both are constrained by how much wind there is right now. Suffice to say that if the ability to spin hydro turbines was dependent on it raining before people get up tomorrow morning then it would have been front page news long ago.
Ultimately we have to go renewable, no doubt about that so far as I'm concerned, but it's not as simple as just scaling up operations and making forecasts unfortunately. Either we need a non-intermittent renewable or we need a means of storing it in decent quantity.