Bill, it isn't just in "storing" energy that we have losses. It is every single time you change the form of energy.

Let's take the above example - storing wind power as compressed air, so that we can pull the power back out of the compressed air and turn it into power we feed into the grid at some later time.

Converting from wind to mechanical power - you have a conversion loss. Converting from mechanical power to potential power in compressed air - another loss - and the big one is the heat loss Gerald mentioned, but there are air compressor/pump inefficiencies, as well as air leakage if we're talking of pumping into salt caverns. One of the proposals I've seen mentioned in the salt cavern idea is to use the heat from compressing the air to heat salt. When the air is then pulled out of the cavern(s), it is run through a heat exchanger that pulls heat back out of the salt and puts it into the air, which increases the overall storage efficiency.

If you don't store the heat, when you decompress the air to turn it back into mechanical energy - you have another loss - both in the actual conversion in the air motor and the missing latent heat that was lost during compression that now can't get back into the air fast enough as it decompresses.

And then when you convert from mechanical to electrical energy, you get the final bill.

This is the nut of the Second Law of Thermodynamics: "You can't break even" -- ie, there will always be a loss in the conversion of energy from one form to another, or trying to store it, etc. You'll never get out 100% of what you put in.

The overall power generation efficiency for this setup is probably quite low.

As for pumping water uphill -- it is probably one of the best (if not the best) utility-scale power storage scheme out there.

Until this nut of "how do we store wind energy?" is cracked, wind power will never supplant large pieces of our power generation portfolio, because it cannot be scheduled, and power generation that cannot be scheduled is useless (or worse than useless) for base load service. We can build wind turbines really well... but we have not made it truly useful by being able to store it. When we do find a viable storage mechanism, I rather suspect that a few thousand acres of turbines in southeast Wyoming could power quite a bit of stuff... if you can keep the turbines from getting ripped apart by these lovely Wyoming zephers that waft across the landscape here...