Sunday, February 1, 2009

Projections, Predictions, and Prognostications

This is a collection of ramblings about measuring and monitoring the right things that will help you maintian your process. Thoughts, please...



There was a radio personality in the town where I grew up that would read the weather at the top of every hour. When he finished reading, he’d play a little drum roll and fanfare thing and then scream, “And now for Kenny’s ‘lookin’ out the window’ forecast!” It never failed to crack me up because it put such perspective on the entire idea of forecasting anything. I remember listening to him one afternoon predicting rain with the report and then make a big deal yelling and fussing that the weather guys were crazy because his ‘lookin’ out the window’ report saw nothing but blue sky and little white puffy clouds.

I took the time to actually look out my window that day… and he was right. It was a perfect “Chamber of Commerce” kind of day. It was indeed blue sky with little white puffy clouds. I realized then that there was a huge difference between projections and reality, no matter how “scientific” or sophisticated the projection mechanism was.

When the weather report predicts a 40% chance of rain they base their prediction on history. The actual formula goes something like this. In all of recorded weather history, when conditions were exactly as we think they’ll be tomorrow at a given time, it rained 40% of the time. They’re basically measuring the past to predict future conditions that they’ll base their prediction on. (You might wanna read that twice)

And for some reason we let these weather predictor guys off the hook when they get it wrong. And they get it wrong a lot! Why on earth do we continue to believe what they tell us? How can we trust anything that predicts what will happen next by measuring what happened yesterday?




A rain gauge can only tell you how much rain you got, not how much rain you’re gonna get!

The best bookies don’t make their bets based on the scores from the previous week’s games. They watch the teams practice if at all possible and they hold off on placing the bet until the last possible moment. That’s the closest they can get to make a real-time projection. What if the star running back got injured in an accident on the way to the game? That won’t show up in the stat sheets.

What if you sold 20 bikes last week and nineteen the week before? Did you do better in week two? Not if your showroom floor was three times as busy in week two than in week one. In fact, you did worse. But your stat sheet (your financials) won’t tell you that. The only thing that’ll tell you that is a real-time look at what’s happening now.

Farmers measure what’s happening now to tell them the weather. Are the cows facing the same direction – right now – right at this very moment? That means it’s gonna rain. Just west of Corvallis, Oregon in the coast mountain range, you’ll see Mary’s Peak. The locals say that if you can see Mary’s Peak, it’s gonna rain. If you can’t see it, well then it’s raining. As silly as it is, it’s at least a real-time look at things.

And that’s where I run out of examples. I need feedback, people. Someone please tell me the message that this brain dump is trying to get across so I can see if I've captured what I wanted to. Thanks for your help.


3 comments:

  1. Who knew the weather man and the average Wall Street pundit had so much in common. I would add one caveat to your commentary... If you do take the weather man's doom and gloom forecast rather than looking out your own window (or perhaps into your dealership's showroom), it can become a self-fulfilling prophecy. Back in his old Dealernews columns, Uncle Paul Wunsch always said: "Perception is reality." The powersports market and the weather are the same way: If you see it as a chamber of commerce day, why not make it a reality rather than complaining we are all going to wash away in a flood of red ink?

    ReplyDelete
  2. An Indian chief started the tribe gathering wood for the winter. Then he called the Weather Bureau to ask their opinion, and they said it would be a cold winter. So he had them gather more wood, and asked the Weather Bureau again. They said it looked like a really cold winter. How can you tell, he asked. They said the Indians are gathering wood like crazy.

    The point is to figure how much to believe. If the weatherman says 50% chance of rain and skies are blue, then it looks like the forecast was wrong. If it's cloudy, then maybe it will rain.
    But you can be fooled. Blue sky in the morning can give way to clouds later in the day. The weather.com site has radar pictures that can show storms approaching beyond the horizon, but rain might fall only in patches. Forecasts now can give you 10% chance before 12, 30% chance between 12 and 5, and 80% chance between 5 and 10. There is uncertainty both when and how much.

    When it comes to economic forecasts, right now there is no one who knows what will happen.
    Nobody knows what Congress will do and which companies will prosper.

    ReplyDelete
  3. I love the last comment. I wonder who left it???

    ReplyDelete