I seem to have missed out several months of Corbynwatch. In fact my last post appears to have been early Sept, concerning his August forecast. To be honest I was getting a bit bored by then, and was also suspicious of his methodology of only sending out his forecasts a few days into the month, and sometimes not at all.
But he did send out the October forecast, and turned out to get it right overall (amusingly, he predicted a cold first half and milder second half, whereas reality produced the exact opposite - nevertheless, on the monthly figures his overall estimate of -0.2 to -1.2C captured reality's -0.6C, and the very average rainfall was also in the middle of his range). So that got his score up to a reasonable 10 out of 16. November, on the other hand, was a complete bust. His forecast was for very cold (at least -1C compared to the long-term mean) and wet (130-200% of normal) month, but it turned out slightly mild (+0.1C) and slightly dry (90%).
I never saw a December forecast, so his final score for the year score sticks on 10 out of 18. (with only 9 forecasts available). If his accuracy really was the claimed 80%, such a poor result would have a less than 2% probability of occurring.
While he didn't issue a public prediction about the December mean temperature, he did say:
He has also predicted that 2009 will be outside the top 5 years, in contradiction of the UKMO prediction. I don't have any plans to follow his forecasts in detail though - especially as there is no sign of the January prediction.
But he did send out the October forecast, and turned out to get it right overall (amusingly, he predicted a cold first half and milder second half, whereas reality produced the exact opposite - nevertheless, on the monthly figures his overall estimate of -0.2 to -1.2C captured reality's -0.6C, and the very average rainfall was also in the middle of his range). So that got his score up to a reasonable 10 out of 16. November, on the other hand, was a complete bust. His forecast was for very cold (at least -1C compared to the long-term mean) and wet (130-200% of normal) month, but it turned out slightly mild (+0.1C) and slightly dry (90%).
I never saw a December forecast, so his final score for the year score sticks on 10 out of 18. (with only 9 forecasts available). If his accuracy really was the claimed 80%, such a poor result would have a less than 2% probability of occurring.
While he didn't issue a public prediction about the December mean temperature, he did say:
This £1,000 was of course lost.Piers and colleagues have placed bets of about £1,000 with William Hill and Coral for snow at various locations on Christmas Day and stand to win many thousands if they are correct.Piers whose record of weather bets is second to none* said: "We have just made a breakthrough in our 'Solar Weather technique of long range forecasting - called SWT25 - and are 75% sure there will be snow in many but not all locations offered by boookies this Xmas, and we have a fair idea of where the best bets are likely to be".
He has also predicted that 2009 will be outside the top 5 years, in contradiction of the UKMO prediction. I don't have any plans to follow his forecasts in detail though - especially as there is no sign of the January prediction.
2 comments:
You actually think they placed that bet?:)
Well the announcement is all that really matters. Not sure whether it reconciles with his claims to have been banned by various bookies either, but these are mere details...
Post a Comment