Our new seat predictor uses local betting knowledge to get beneath the polls

By Anthony Painter

How on earth can we make sense of the changes that seem to be occurring in the campaign since last Thursday? The day after the debate I went live with a new seat predictor that has been developed by the very clever people at tmg.co.uk. We’ve called it the ‘tmg.co.uk anthonypainter.co.uk seat predictor’ and it uses an entirely different methodology to the universal swing (UNS) calculators that generate the sort of seat projections you have been hearing on the news for the last few days (which have been projecting Labour as the largest party while in third place in the vote – which I’m very sceptical about).

The new predictor shows that there has been a solid shift to the Liberal Democrats over the last few days – in the main from the Tories in terms of votes, and Labour in terms of seats. It forecasts that the Lib Dems have gained 16 seats since last Thursday.

How does this model work? It’s based on 100,000s of bets placed on mainly local betting exchanges. Here’s a description from the developers at tmg.co.uk:

“The system collects live (in play) betting pattern information from a selection of different sources and rates the current crowd sourced information against previous points in time and uses various methods to judge if the current leader is falling back or a challenger is catching up. This is turned into a trend (a percentage win chance matrix) and is projected forward to the finishing line and a seat is called.”

Now for the caveats. This is experimental. The concept, methodology and results have a plausibility. The slower pace of change in this model compared to UNS calculators is not necessarily a downside. Remember, the fuel protest in 2000 and the impact on opinion polls then. Well, it was short-lived and evaporated almost as quickly as it happened. And UNS calculation then of the likely outcome in the 2001 election would have been way off. And here is some analysis by Political Betting about the need for caution about UNS calculations.

It seems that betting markets are not yet convinced that the change in this election has been seismic. And remember, these are forecasts about the outcome of individual seats on May 6th not a snapshot of opinion now.

So my advice would be to keep an eye on this as it develops. A little local knowledge goes a long way in a first past the post system. The dynamics of our political system may be fundamentally changing. No-one has a monopoly of understanding about how it is changing. This is one way – amongst others – of assessing the change. Check www.anthonypainter.co.uk every morning or follow me on Twitter to get updates.