With the opinion polls deeply divided, essentially between the telephone surveys, which are the favourites of the ‘Remain in the EU’ crowd, and the online polls, favoured by the ‘Leave’ tribe, where can we look for the true state of opinion? Conventional wisdom is to turn to the prediction markets. That is precisely what I intend to do in today’s article. Rather than focus on the relative merits of telephone and online polling, or assessing the likely impact of a differential turnout of younger and older voters, or considering the likely effect of the recent distribution of the Government’s official ‘Stay In’ leaflet, I shall simply highlight the current findings of the big state-of-the-art prediction market and crowd aggregation sites.

I shall exclude the betting exchanges and fixed-odds bookmakers, though in part these are fed in to some crowd wisdom-aggregation algorithms, and seek to derive conclusions from these forecasting methodologies as to the value in the current prices.

The prediction markets and crowd-wisdom aggregation sites I shall consider today are Almanis, Hypermind, Pivit, PredictIt and PredictWise.

Almanis works by allocating points to those who join which can be used to back forecasts about the probability that an event will occur. The more points the forecasters stake, the more they can shift the probability. While no money changes hands, there are prizes for the best forecasters. The wise heads at Almanis are currently predicting the probability of the UK voting to remain a member of the European Union in the forthcoming referendum as 70 per cent.

Hypermind consists of a panel of forecasters selected and rewarded based on past performance. The idea behind it is that forecasting is quite a stable skill, so focusing on the predictions of proven forecasters is a sound basis for estimating event probabilitiies. The collective mind at Hypermind gives a probability of 62 per cent that the ‘Remain’ camp will prevail.

Pivit offers a methodology of combining historical data, real-time information and a traditional ‘wisdom of crowds’ prediction market platform to generate live odds of different events occurring. The Pivit model currently puts the likelihood of the UK voting to stay in the European Union at 68 per cent.

PredictIt is a real-money prediction market, albeit limited to a cap of $850 on individual investments per question to comply with US regulatory policy. Each question is also limited to five thousand traders. The current estimate of the PredictIt community that the UK will vote to remain in the EU is 62 per cent.

PredictWise is a crowd wisdom aggregation site, which offers no markets of its own but essentially derives probabilities based on a combination of other forecasts, notably from prediction markets, betting exchanges and bookmakers. PredictWise currently has ‘Remain’ at 63 per cent.

Taking an average of each of these probability estimates gives us a current probability of 65 per cent that the UK will vote to remain in the European Union on 23rd June. If this is a good estimate, it would mean that we are looking for odds a shade long of 2/1 for ‘Leave’ or better than 1/2 for ‘Stay.’ At the 70 per cent probability of ‘Remain’ estimated by Almanis, on the other hand, or the 62 per cent by Hypermind and PredictIt, the value plays look rather different.

The best value at the moment, however, may lie in the Vote Share markets. Spread betting operator, Sporting Index, are pitching the line for Remain between 53 and 54 per cent, which appears consistent with an estimated probability of a win for ‘Remain’ at a shade long of 2 in 3.

Indeed, taking an estimate anywhere between the 62 per cent and 70 percent derived from each of our five highlighted forecasts, we might expect a vote share for ‘Remain’ somewhere between 50 and 55 per cent.

In conclusion, the best odds of 11/4 that between 50 and 55 per cent of voters will opt to ‘Remain’ in the European Union currently looks very reasonable value.

Follow Leighton on Twitter @leightonvw

50-55% remain vote - 1pt @ 11/4