Statistics Made Easy

[Alan Robinson]

Just out of interest, here is part of an article this morning from the Danish Meteorological Institute, concerning the use of statistics and probability in forecasting winter weather; my free translation:

â€The two record-cold Danish winters (December, January and February) of 1938-40 and 1962-63 had both a mean temperature of -3.6 degrees. Neither however started with a freezing cold December as we experienced it in 2010.

â€The mean temperature in December 1962 was just -0.6 degrees, and December 1939 as high as +0.6. The trouble came in January and February with -4.4 and -6.8 in 1940, while in 1963 it was -5.3 and -4.3. The long-term mean for these two months is however 0 degrees.

â€If we can conclude anything from these terribly inadequate statistics, it must be that:

1. We have never yet experienced an extremely cold winter with three consecutive cold months

2. Two years out of three, extremely cold Decembers are followed by a cold January

3. Extremely cold Decembers are historically followed by a rather milder February

4. The record cold Danish winters didn’t start with a freezing December

â€Historical data suggests that we are now at least half way through the harshest part of this present cold winter, and current medium range forecasts to some extent indicate milder conditions.

â€The question is, of course, whether or not the weather takes any notice of such historical data. Probably not, but it will be interesting to see what develops…………

************************

Well really. If this is all state-funded meteorologists can dish up for us, I think I’ll start a subscription to Old Moore’s Almanac. After all, it is cheap at under three quid and better entertainment than this sort of psuedo science.

Edited January 8, 2011 by Alan Robinson

[Terminal Moraine]

Just out of interest, here is part of an article this morning from the Danish Meteorological Institute, concerning the use of statistics and probability in forecasting winter weather; my free translation:

â€The two record-cold Danish winters (December, January and February) of 1938-40 and 1962-63 had both a mean temperature of -3.6 degrees. Neither however started with a freezing cold December as we experienced it in 2010.

â€The mean temperature in December 1962 was just -0.6 degrees, and December 1939 as high as +0.6. The trouble came in January and February with -4.4 and -6.8 in 1940, while in 1963 it was -5.3 and -4.3. The long-term mean for these two months is however 0 degrees.

â€If we can conclude anything from these terribly inadequate statistics, it must be that:

Hmm, that must be statistics at its most basic, I'm surprised the Danish Met' Office had the temerity to publish it.

It really tells us nothing more than would be gleaned by the cursory glance of a layman presented with such a pitifully small data set; most of it is just stating the obvious.

1. We have never yet experienced an extremely cold winter with three consecutive cold months

2. Two years out of three, extremely cold Decembers are followed by a cold January

3. Extremely cold Decembers are historically followed by a rather milder February

4. The record cold Danish winters didn’t start with a freezing December

â€Historical data suggests that we are now at least half way through the harshest part of this present cold winter, and current medium range forecasts to some extent indicate milder conditions.

â€The question is, of course, whether or not the weather takes any notice of such historical data. Probably not, but it will be interesting to see what develops…………

************************

Well really. If this is all state-funded meteorologists can dish up for us, I think I’ll start a subscription to Old Moore’s Almanac. After all, it is cheap at under three quid and better entertainment than this sort of psuedo science.

Hmm that must be statistics at its most basic, I'm surprised the Danish Met' Office had the temerity to publish it. It really tells us nothing more than would be gleaned by the cursory glance of layman presented with the same data; most of it is just stating the obvious.

Edited February 20, 2011 by Terminal Moraine

[frogesque]

There are lies, damned lies and statistics!