Jeff, our volume isn't what we anticipated for tomorrow so we aren't going to need you to come into work." That's the call I received from a product salesperson from a Colorado ski resort that I work for part time. My services aren't needed yet because there are not enough people in town to warrant having me on the hill.
"Go enjoy some personal skiing," she said.
But why, if the snow is pretty good and the holiday is close are bookings lower than expected?
Last year, in case you weren't around, was a lousy snow year. The year before it puked, a record amount in fact. One would expect a correlation between snowfall amounts and bookings that look something like this:
Lots of snow = lots of bookings
Little snow = little bookings
However, as the latter part of this week and last year shows this in not always the case. In fact, it is often the opposite. Skiers from around the country heard about last year's dismal snow fall and failed to book a trip to Colorado for this year.
Yet, last year, despite the snow drought gripping the High Country, bookings were high because people heard about or experienced the winter 2010-2011 and its 500+ inches and booked a trip for 2011-2012. They ended up sad skiers and snowboarders as some resorts received less than 180" of snowfall leaving locals feeling compelled to apologize for not producing the exceptional snowfall from the previous year.
We were not being punished by the Weather Maker.
The ignorant gapers who showed up last year expecting a repeat of the previous year's snow totals and who aren't here now despite much better early season conditions are guilty of a planning fallacy and ignoring the statistical concept of "regression to the mean".
Take a look at the average annual snowfall from three of our state's ski resorts.
Breckenridge - 300"
Copper Mountain - 282"
Steamboat - 335"
These are annual averages. Mean is a fancy statistical way of saying average over time. So, if 300" is average then it logically follows that some years we will have more than 300" like in 2010-2011 when we received around 550" depending on where you measured. And in some years you will have less than 300" like last year's underwhelming 160" give or take a few inches.
Regression to the mean is an explanation without a cause where a variable, in this case snowfall, is measured as extreme in the first year tending closer to the average in the second year.
Daniel Kahneman's wonderful book "Thinking Fast and Slow," Farrar, Strauss and Giroux, 2011 does an excellent job of showing the many incorrect ways our minds look at things. In it he writes, "the more extreme the score," (or snowfall) "the more regression we expect." In other words, 550" of snow is bound to be followed by less snow the next year. Conversely, if it snowed 160" last year then this year we should expect. . . well . . . "the regression prediction is reasonable but not guaranteed," writes Kahneman.
People shouldn't be avoiding Colorado after a bad year they should be flocking here.
We're human of course and not machines. We like a good story to provide our causal explanations over "mere statistics". El Nino or La Nina, shifting jet streams or global warming, anything so long as we don't have to hear our local weatherman tell us that the reason we are getting more snow this year is because we didn't get much snow last year. Even if this explanation is at least as true as the La Nina bit, it lacks the compelling narrative that we love so much.
I would like to see that lovely lady in product sales try just once to see if it works.
"Yes Mr. Overstreet, last year was as bad a snow year as I can remember but have you heard about the regression to the mean?"
Jeff McAbee is a former Summit County resident now living on the Front Range. Contact him at firstname.lastname@example.org or via Twitter @Jeff_McAbee.