A new tool to help predict the next House of Commons

My biggest ever political gambling loss was on the 2001 General Election because I had not thought through the issue of relating votes to seats – the critical factor in UK General Elections. I bought the Conservatives in the spread betting Commons seat markets because I believed that the opinion polls were wildly overstating the Labour lead and that the election would be much closer than was being predicted.

I was completely right. Instead of the 25-30% leads that the pollsters were recording Labour’s eventual lead was 9% down 4% on the landslide victory of 1997. But all the accumulation of votes in their own areas did was to notch up the overall voting figures without producing what really mattered to them (and my bet!) – more MPs. They were also hit by strong tactical voting.

There were two lessons:- The Westminster seat distribution is totally skewed to Labour: Spread betting can be very expensive!

The biggest pitfall political gamblers are likely to make at the coming General Election is to assume that Tory poll leads will simply translate into a Commons majority. As we’ve been saying here for months the way the seats are distributed gives Labour a huge advantage and whenever we have reported new poll results we have also tried to demonstrate what this would mean in terms of seats at Westminster. Our main tool has been the calculator produced by the Cambridge and now City mathematician, Martin Baxter. This operates using a uniform national swing and it is very simple to work out a given outcome from new poll findings.

Now a new tool that should interest political gamblers considering General Election bets has been produced by the polling commentator, Anthony Wells. He’s created, for download, a spreadsheet that allows simple projection of the House of Commons on any given share of the vote.

All you have to do is enter percentage vote shares for Labour, the Tories and the Lib Dems and up comes a projection of what the new Common make-up would be. All very simple.

You can choose either the current boundaries, or the new Scottish boundaries. The Wells programme differs
slightly from Martin Baxter calculator, which we use a lot on Politicalbetting.com, due to slightly different
methodology when dealing with split wards. Comparing the two the Wells system seems to be less generous to the Lib Dems. Thus taking the May Populus Poll the two calculators produce the following predictions.

Populus Poll: CON 36%, LAB 32%, LIB 22%
In Commons seat terms these vote shares convert to-
WELLS: CON 253 seats; LAB 305 seats; LIB 56 seats
BAXTER: CON 245 seats; LAB 305 seats; LIB 65 seats

A useful function with Wells allows the user to include anti-Labour tactical voting. The figure is the percentage of those voting for a third candidate (who must be significantly behind the second placed candidate) who would instead vote tactically against Labour. Anthony points out that this is very basic at the moment, and unrealistic figures would produce odd results. This is the Wells calculator version of the May Populus Poll with a 10% third party tactical voting element.

WELLS (10% Lab Tactical) CON 262 seats; LAB 295 seats; LIB 57 seats

Another polling expert, Chris Lightfoot has produced his own one-dimensional calculator based mainly on Martin Baxter’s system.

No doubt that at the General Election the BBC and several of the newspapers will be making their own versions of the Wells and Baxter calculators available. Now all we need is a 100% accurate system for determining which opinion poll is the most accurate.

WILLIAM HILL UPDATE: More readers recorded problems yesterday trying to place bets on Norris in the London Mayoral market. It seems you can put as much money as you want on Livingstone – but hardly anything on Norris. This is creating a false market. When last week’s poll came out showing Norris only 5% behind Ken the prices should have got much tighter. This has not happened in any significant way. If you have any more experiences please let us know. What better argument can there be in favour of betting exchanges!