The Risk Factor

If you asked me where I was last Tuesday, I'd have to think about it for awhile. However, I can tell you where I was 20 years ago today: the Bayswater section of London where I was living at the time. I remember that because today and tomorrow mark the 20th anniversary of England's Great Storm of 1987.

At the time, the storm was considered the worst to hit England since 1703, devastated forests throughout southern England, killed 18 people, and caused over £1 billion in damages. I remember walking around London after the storm hit and being awed by the damage done. I had already seen the aftermath of hurricanes along the US Gulf coast, but this left a greater impression, probably since it was totally unexpected.

The UK Met Office had forecast heavy rains for the overnight of 15 - 16 October, but not the intensely strong winds. Weather forecasters thought a less powerful storm would stay south in the English Channel or hit northern France.

What made the forecasting mistake worse in the public eye was that BBC weatherman Michael Fish said on his 2130 forecast, "Earlier on today apparently a woman rang the BBC and said she'd heard there was a hurricane on the way. Well if you are watching don't worry, there isn't." Fish was actually talking about a different storm, but everyone assumed he was talking about the one that indeed did hit England.

The Drudge Report has a link to a Orlando Florida television news story that tells of Central Florida's most famous hotel owner, Harris Rosen, who is threatening to sue hurricane expert Dr. William Gray of Colorado State University for his hurricane storm predictions saying they have damaged state tourism.

According to the story, Rosen rhetorically asks Gray:

"Look, doctor, you've made these forecasts and you were wrong once. You made the forecast and you were wrong twice. Are you going to continue to make these forecasts?"

Rosen said he believes Florida lost billions of dollars in business because of Gray's outlook, and claims that surveys show 70 percent of guests not returning to his hotels cited hurricane fears as the reason why.

I don't know why Rosen focused on Gray alone, and not the others who also had less than accurate forecasts the last two years. I also doubt Rosen is going to have much luck in filing a lawsuit, and his real target should probably be the media for over-hyping the accuracy of the forecasts which as one commentator points out "are experimental works in progress." If one could sue for inaccurate weather predictions based on computer generated models, the court system would grind to a halt in about three days.

As a side note, Herbert Saffir, who co-created with Robert Simpson the five-category hurricane-scale, passed away about a week ago.

The Turing Test may be how a machine demonstrates artificial intelligence, but my own test is: can a machine interpret correctly the US Tax Code, and can one predict accurately winter weather in Northern Virginia three hours ahead of time?

Yesterday, an unpredicted ice storm hit Washington D.C. and its suburbs leading to accidents and severe icing that shut-down at rush hour I-95 south at the Springfield Mixing Bowl for over 8 hours. The storm was expected to hit well west and north of the Washington area, even as late as noon. This was the second "surprise" ice-storm to hit the DC metropolitan area this winter.

Those of you who know the location can image the mess - there is a nice photo from the Washington Post here.

Yup, I'll believe in AI when my two tests are met.