Forecasting record temperatures

[Loren Booda]

The Central England Temperature series is the longest continuous weather chronicle, reaching back to 1659. What is the probability that any particular day in 2009 sets a record temperature there, assuming only random temperature fluctuations over the 350+ years?

 

[Adeimantus]

I'm thinking that since the temperature fluctuations are random, then the records, one from each year from 1659 through 2009, are also random and independent. So now you have a list of 351 numbers which are equally likely to be ranked in any order (since these yearly records are random and independent). What is the probability that the 351st is the largest? I think this is the probability that the record for 2009 is the "record of the records." Maybe someone else could confirm this.

 

[HallsofIvy]

In such a situation, if there are n values in a list then the probability that a new value is greater than any of the values already on the list is 1/n.