Saturday, March 13, 2010

On the small difference between big n...

On the small difference between big numbers

A comment on BraveNewClimate:
Well if we have trouble talking sensibly about the small difference between two very big numbers then we’re not alone. We get it all the time in discussion of company and national accounts. If you say that revenue doubled then that means a lot, but may not tell you anything about net profit. On the other hand it is meaningless to say “profit doubled” since profit is often close to zero in a bad year (indeed often negative in a bad year: and then doubling would have no meaning that ordinary folk would understand). [A change from 10% profit to 20% profit is best seen as a change from 110 to 120, rather than a doubling.]
This came from a Hansen comment that was (apparently) meant to mean that ice loss from Antarctica had doubled in 5 years. Even if Antartic Ice is constant over a longer term it is likely to go up and down. Get close to the beginning of a down period and get a small number. Then further into the period get a bigger number. Then make an announcement that sounds like we're in an exponential explosion.