Tuesday, November 06, 2007

The Alpha Effect

In any business, one of the most important and often most difficult tasks that you have is predicting your future sales. It's particularly important in the electric business. But in other businesses, if you fall short of demand, your product may be valued more highly and you might even make money by not meeting your actual sales potential.

However, in the electric business, you need to have the product ready to go. And in the electric business, that historically means building a big expensive plant, well in advance of your present sales. Sure, you can buy power from the grid, but that leaves you pretty much in the clutches of a market that is sometimes quite volatile, which definitely takes your ability to control your bottom line out of your hands.

Consequently, utility companies spend a great deal of time and effort trying to predict the future so they can be prepared to meet their load growth. We do this by looking at our past growth, at the prosperity of our community, and by trying to keep tabs on what is happening in the world in general.

Last week, I was sitting before a large whiteboard and we were discussing the year 2020. Now, in our community, we have made a commitment to meet 30% of our needs with renewables by that date. The question is, would that amount of new renewables be enough to meet the growth that we expect, and how would we make up that difference, and perhaps more importantly, how would we manage those sources to insure plenty of product during our peak demand.

When it became time for me to comment, I responded with two words.

Alpha Error

As I looked at these charts and graphs, I saw that we were doing what any responsible organization would do. We were looking at the past and using it to predict the future.

In statistics, there is this troublesome concept called alpha. If you look at any political poll for example, you see that the results are correct plus or minus say 3%, depending on the size of the sample. But there is also this chance that the data that you have collected means absolutely nothing about the situation you are trying to model. This is alpha.

But I don't mean alpha in the purely statistical sense though. I mean "alpha error" in the sense that the data you have, and the situation that you presume, may not fairly represent the future you are getting ready to face.

What if oil supplies grow tight because of War or even excessive demand and the price of oil soars? (it's in the mid 90s today)

What if that causes a rip snorting recession.

What if the dollar collapses?

What if next year, the sea ice in the Arctic completely melts and even Russ Fatball can't spew out his goofyiness any longer and that sea change in understanding makes burning coal an international crime?

What if that brings President Edwards to mobilize the country, much like FDR did, and within 2 years, we have 10 cent distributed PV and 7 cent centralized solar?

What if General Motors launches their new Volt and it is an amazing success, and plug in hybrids are suddenly the only sensible vehicle to buy given the super high price of gasoline. (You also are happy to not be constrained by your gasoline ration booklet) And with this new fleet of electric vehicles, the utility is not only able to buy more wind power, it is able to buy it back during peak demand.

What if a super thin flat battery or an ultra cap device is proven to be an effective way to store energy both for cars, homes, and utility systems in general?

What if a climate stabilization premium is inacted, and a new wave of clean energy technologies come on the scene; and, in the same time that horseshoes and buggy whips were relegated to the game pits and museums just 100 years ago, coal plants and petroleum driven devices become as important to our well being as the hay barn?

What if the corporate consumer society that has taken over our lives is transformed by a enraged public that is no longer willing to allow our democracy and our domestic tranquility to be hijacked by this rogue group of corporate psychopaths?

Clearly, you cannot run a utility on these "what ifs."

But also, in these times of exponential change and extreme environmental global challenges, you cannot blindly assume that everything is going to be the same either.

So you begin to craft polices of least regret. You don't want to build a new coal plant if there is even a small chance that it will need to be prematurely shut down. You don't want to find yourself in another nuclear project that ends up costing more than a blend of wind, central station solar, and a basket of firming strategies.

You try to make your community as efficient, as walkable, as car free, and as electronically advanced as you can imagine. You put in bus stop cut-outs, and you build bike avenues. You make the city a quilt of villages. You make plans for each home and office to be a zero energy structure, and you do that by first making them energy efficient and environmentally aligned.

In most polls, when alpha is 5%, the chance that the data you have is meaningful is 95%. In today's world, the chance that the data we had yesterday will accurately portray tomorrow may be just the reverse.

When Alpha is 95%,

The past is the past.

And the future is ours.


"alpha alternatives" courtesy of Marina Hanacek

Labels: ,


Anonymous Anonymous said...



1:06 PM  

Post a Comment

<< Home