One of the most dangerous, risky but potentially most profitable political gambles is to bet against the opinion polls. I’ve done that twice this year – first by laying Howard Dean for the Democratic Nomination and second by betting against Ken Livingston being re-elected as Mayor of London.

The Howard Dean call, made publicly on the Betfair Forum, was when Dean was 20% ahead and I got an average price of 1.7. At that stage I felt I could dismiss the polls because many respondents had not really considered the matter and were just repeating back what they had read in the media. I did not believe that any candidate could be priced at such odds when not a vote had been cast and I was influenced by US commentators who were saying that Dean did not have the character to be a credible candidate.

My reading of the Ken polls is the same. They got it disasterously wrong last time, grossly over-estimating his position, and my doubts were underlined in the following YouGov poll at Christmas when his plan to re-join Labour was in the news. The answers to three questions were inconsistent and showed that, when probed, Ken’s vote was soft.

The first was How would you vote with Ken as an Independent?
37%- Ken Livingstone (Independent)
19%- Steve Norris (Conservative)
12%- Simon Hughes (Liberal Democrat)
4%- Nicky Gavron (Labour)
3%- Others

The second was how would you vote with Ken as official Labour candidate?
33%-Ken Livingstone (Labour)
20%- Steve Norris (Conservative)
15%- Simon Hughes (Liberal Democrat)
5%- Others

The third question was the revealing one – Taking everything into account, would you be more likely or less likely to vote for Mr Livingstone if he stood as Labour’s official candidate, rather than as an independent?
9%- More likely
19%- Less likely
26%- Would make no difference – I would vote for him in any case
35%- Would make no difference – I would NOT vote for him in any case
11%- Don’t know

The anti-Ken shares rise as people think more about the issues. With the first it was 38%; with the second it was 40%; and with the third (taking the “less likelys and the wonts) it reaches 54%. The pro-Ken total is 35%. These are not figures that would cause me to back Ken at 1.2 or lay Norris at 11 or more.

Before Ken’s first election in 2000 every one of the six opinion polls had had him in the 50-68% range with Norris at 11-17%. The eventual Ken-Norris split was 39-27.2 %, an appalling outcome for the pollsters – a great outcome for me!