Archive for November, 2004


General Election Prediction time (RECOVERED)

Tuesday, November 30th, 2004

Let’s do some number crunching

With just over five months from the predicted General Election day of 05/05/05 it’s time we started making some forecasts of what we think might happen first in terms of the vote share and secondly the number of Commons seats.

LABOUR VOTES. Our formula is based in the chronic tendency of the pollsters to always over-state Labour. In two of the last three General Election every single poll from every single pollster produced an exaggerated figure. In the other election, 1997, the majority of pollsters did the same. So our formula is to take the average of the bottom two Labour shares in the most recent polls and then DEDUCT 1.5%.

TORY VOTES. If the US automated pollster, Rasmussen, does do UK surveys then we will take their Tory share and make no adjustment. They were the only firm to get it right last time. If not we will take the average of the highest two figures from the latest round of polls and ADD on 1%.

LIB DEM VOTES. The party seems to get about the same or slightly more than the highest polling figure so we’ll take the average of the top two shares from the latest poll round and then ADD on 0.5%.

CONVERTING VOTES TO SEATS. We’ll put the poll shares we’ve determined into the Baxter calculator and then make two adjustments: we’ll take 12 seats from Labour and add them to the LD total because of special targeting. We’ll also take 14 seats from the Labour total and give them to the Tories for “tactical unwind” – a controversial notion but one we believe will happen.

On current figures this produces the following “result”

Lab: 306 (33.5%)
Con: 237 (34.0%)
LD: 72 (23.5%)

The rest is easy – just put your bets on and wait for the money to roll in!

Latest spread prices:- LAB 345-353: CON 200-208: LD 71-75
Comments (77)



Are the bookies scared of Blunkett?

Monday, November 29th, 2004

    Why no WILL BLUNKETT SURVIVE markets?

In recent years a healthy tradition has grown up of betting on whether politicians in trouble will survive.

Within days of the David Kelly case last year you could back or lay on whether or not the main players – Alastair Campbell, Geoff Hoon, Tony Blair, Andrew Gilligan, Gavyn Davies – would hold onto their jobs beyond September 30 2003. In terms of taste the Kelly case was far worse than the Home Secretary’s current predicament.

A week or so later last autumn it was the hapless IDS who became the focus of betting attention and whether he would still be in post in mid-November. Then in the run-up to the publication of the Hutton report the main players came into the frame again and money could be won and lost.

Currently you can bet on whether Peter Mandelson will serve his full term as a Euro Commissioner. So why no market on Blunkett?

    Could it be, we wonder, that the betting industry is worried about the promised Government plans to tighten up the law on gambling, particularly on the internet?

There’s a huge interest in the story and lots of speculation on whether or not he can carry on as an authoritarian Home Secretary laying down the law to all of us with the papers full of his private life. People want to bet but the only Blunkett market about is whether he’ll be the next Labour leader. Even at 20/1 we do not think that this is a value bet.

Meanwhile the affair – political that is – has taken the wind a little out of Labour’s sails on the main General Election market. Although the bookie price on Labour is still 1/7 the Betfair odds have moved a touch to 1.22. They were 1.18 a week ago.


The Politicalbetting Top Ten – November 28 2004

Sunday, November 28th, 2004

    UK General Election back as most popular market

With two new polls both giving totally different pictures it’s not a surprise that the UK General Election is back at the top of the politicalbetting top 10 – which is based on “click-throughs” to online bookmakers from the site.

    If you are thinking of betting in the UK markets it’s worth recalling that at the 2001 General Election every single poll from every single pollster over-estimated Labour.

Our safety first strategy is to base your bets on whatever is the current lowest Labour poll share and then knock one or two points off.

1. General Election Winning Party Labour still 1/7 best with the conventional bookies but has moved out from 1.18 to 1.21 (better than 1/5) on Betfair. Is there just a sense of weakness? Maybe – but there is no value here either way.
2. Republican Candidate 2008 Quite a bit of interest with money going on Rudolph Guiliani – former New York mayor now at 11/4. It’s a long way to go and the evangelical right will want to ensure that they continue to have their person in the White House.
3. Winning Party 2008 Given it’s evens on both and you have to wait nearly four years to pick up your winnings we are amazed that people are betting. Remember to factor in the cost of locking up the money.
4. Conservative Seats at General Election. The 10/11 on the Tories getting less than 220 seems a popular bet. We are not so sure because the polls are giving conflicting messages.
5. US President 2008 Hilary Clinton is still the 5/1 favourite. No comment.
6.General Election Date. Given the Sun has now said that its 05/05/05 why not get on now while there’s value at 1/5? One thing’s for sure – Tony Blair’s not going to upset the Sun’s owner Rupert Murdoch.
7. Tony Blair’s departure date 2006 is the 6/4 favourite. We are not so certain.
8 Labour Party Seats A seat total of 360+ is the 11/10 favourite. Much less risky than a spread bet but much reduced returns.
9. Liberal Democrats seats A total of 61+ remains the 1/3 favourite. Probably good value.
10 Democratic Candidate 2008 Hilary Clinton is 5/4 favourite. Not good value given the time your money is locked up


Labour back at 2001 General Election level – Indy poll

Sunday, November 28th, 2004

    Has Blair really not been hurt by Iraq, Kelly, and Hutton?

The Independent on Sunday’s new pollster, Communicate Research, has Labour back at their 2001 General Election share according to their latest survey out today.

The figures with changes on the same poll last month are CON 31 (-2), LAB 42(+2), LD 20(+3). At the last election Labour got 42%, the Tories 32.7% and the Lib Dems 18.8%. So the only change on 2001, according to CR, is a small switch the Tories to the LDs.

We have talked a lot on the site about weighting results according to people’s recall of what they did in 2001 because the difficulty that phone pollsters have of finding a representative sample. CR do not do this.

What we have are figures unweighted by recalled past vote of people who have landlines and happened to be in, ready to answer an unsolicited phone call from a stranger and be willing to talk when the computer generated dialler got through to them.

For a range of reasons, as other pollsters have found, such an approach tends to produce disproportionately more Labour supporters than those of other parties. Unlike other pollsters, as well, CR do no set out in their interviews the different party options – an approach that has been shown to favour Labour and the Tories.

    At the 2001 General Election the Indy’s then pollster, Rasmussen, was one of the most accurate and was the only firm to predict precisely the Tory share. Why they should have been replaced by CR and their methodologies beggars belief.

Elsewhere in the Indy today there is sceptical piece on pollsters by the veteran political commentator, Alan Watkins Everyone seems to agree that, come May or whenever it is, this government is going to get in again. The polls say so. The journalists at Westminster have few doubts. Even the Tories admit ruefully that their time has not yet come: if, indeed, it ever does come. But the strange thing is this. Of all the people I know – friends, relations, acquaintances whether close or slight – not one of them intends to vote Labour at the election. Some do not propose to vote at all. Others promise to support the Respect coalition, if there is a candidate available. Most are going for the Liberal Democrats, about half of them (I would estimate) in a switch from Labour. A few, most eccentric of all, say they will vote Conservative. But then, that is because they have always voted this way and, if they were prepared to put up with Mr John Major and then Mr William Hague, they see no reason to forsake Mr Michael Howard now. This, I realise, is what the sociologists used scornfully to call anecdotal evidence, though what other kind of evidence there is I do not know.

The CR survey also opens up a massive 10% gap between the various polling organisations on the Labour lead. They cannot all be right and our money would certainly not be on the Indy survey.

The spread markets are unchanged at LAB 344-352: CON 202-210: LD 70-74. The Labour price is still two seats below where it was at the end of July when Labour’s lead was just 1%.


Welcome to the new world of polling transparency

Saturday, November 27th, 2004

    Revealed – How the pollsters produce their numbers

People who like to probe behind the headline figures that are issued by the polling firms are going to have a field day following the introduction of the British Polling Council rules requiring pollsters to publish the background data that was used.

For the first time the main pollsters are having to publish a specified range of information that went into producing their headline figures so that interested observers can subject them to proper scrutiny. This is a very welcome move and will allow us to see in detail if there is something in the pollsters’ methodology that has caused the inherent over-statement of Labour for nearly two decades.

The internet poll by YouGov in the Telegraph poll yesterday had Lab 35: CON 32: LD 23. But according to the information that YouGov has provided the following weightings based on the surveyed individuals’ recall of how they voted in 2001 were applied – CON 28: LAB 56: LD 13.4 So from this we can calculate:-

  • Labour is polling at 62.5% of its 2001 recall figure
  • The Tories are polling at 114% of their 2001 recall figure
  • The Lib Dems are polling at 172% of their 2001 recall figure
  • When we did a similar calculation on the November ICM figures we observed that “you cannot easily dismiss the fact that such a large proportion of those saying they were Labour last time have apparently deserted the party.”

    We are sure Lib Dem supporters will be delighted that applying the 172% to what actually happened in 2001 produces a projected share of 32.3%. The Tory figure moves to 37.4% and Labour to 26.25%. What’s extraordinary about this calculation is that it very nearly mirrors what actually happened in the local elections on June 10 this year when the votes split – CON 38: LD 29: LAB 26.

      We acknowledge that this is a bit mischievous and we are certainly not saying that this will happen. But the more information you get the less clear things look.

    The polling commentator, Anthony Wells, has got a good analysis and concludes that it is “bizarre” that with the published weightings YouGov has tended to give Labour lower ratings than the other pollsters.

    The only betting conclusion is BE VERY CAUTIOUS. Full odds round-up.


    What are we to make of the White House Race rigging allegations?

    Friday, November 26th, 2004

      Is former Labour minister, Michael Meacher on to something?

    At the risk of appearing a bad loser, a conspiracy theorist or a paranoid we publish extracts from an article by former Labour cabinet minister, Michael Meacher, in this week’s New Statesman.

    He raises questions that do seem worthy of further investigation about the exit polls on November 2nd in states where there was electronic voting (above) and the rest. Meacher’s claim is that the polls were reasonably accurate in the latter but in states where electronic systems were in place they were not – and the “errors” were all in the President’s favour.

    Unfortunately the article does not contain enough detailed statistical analysis to make the case fully but we thought that Politicalbetting users might find it interesting.

    Meacher writes:-Now allegations are surfacing that the use of electronic voting systems and optical scanning devices may have had a significant influence on the result. Computer security experts insist that such sys- tems are not secure and not tamper-proof, yet they were used to count a third of the votes across 37 states. Though the Democrats remain strangely coy about the whole subject, academics and political analysts are now drawing comparisons between areas that used paper ballots and areas that used electronic systems. Is it possible that results in the latter were rigged?

    An analysis of the poll by different states points up inconsistencies that cannot be explained by random variation. In Arizona, Colorado, Louisiana, Michigan, Iowa, New Mexico, Maine, Nevada, Arkansas and Missouri, where a variety of different voting systems were used, including paper ballots in many cases, the four companies carrying out exit polls were almost exactly right and their results were certainly within the margin of error. In Wisconsin, Pennsylvania, Ohio, Florida, Minnesota, New Hampshire and North Carolina, however, where electronic or optical scanning machines were used (though not exclusively), the tracking polls were seriously discrepant from the published result.

    Two aspects of this are immediately striking. One is the large size of the variance, and the other is that in every case it favoured Bush. In Wisconsin and Ohio, the discrepancy favoured Bush by 4 per cent, in Pennsylvania by 5 per cent, in Florida and Minnesota by 7 per cent, in North Carolina by 9 per cent and in New Hampshire by an astonishing 15 per cent.

    Moreover, extensive voting irregularities have been reported across the US – including intimidation, exclusion of black voters from electoral rolls, touchscreens that consistently registered support for Bush when the name Kerry was touched, and a large number of county precincts (including in Ohio) where the number of votes cast exceeded the total number of registered voters, sometimes by large margins. In Florida, for example, the number of votes reported for all the candidates exceeded the maximum possible voter turnout by 237,522, so that a minimum of 3.1 per cent of the votes must be fraudulent, and possibly considerably more…

    …So can we really be sure that this year’s result was an accurate reflection of the popular will? It has emerged that the Diebold Gems software and optical scan voting machines used in counting a high proportion of the votes may not be tamper-proof from hacking, particularly via remote modems. Two US computer security experts, in their recently published book Black Box Voting, argue that “by entering a two-digit code in a hidden location, a second set of votes is created; and this set of votes can be changed in a matter of seconds, so that it no longer matches the correct votes”. After the Florida fiasco four years earlier, the US Congress voted $3.9bn to improve the quality of voting systems. Perhaps the latest revelations about what happened where electronic systems were used may become known as the “November surprise”.

    It will be interesting to see if this is taken up. It does need the detailed numbers to support the case.


    YouGov shows Labour lead down to 3%

    Friday, November 26th, 2004

      Slightly better news for Michael Howard

    After the NOP and ICM polls showing Labour leads of 8-9% this month’s YouGov poll in the Daily Telegraph has the margin dropping to just 3%. The figures are:-

    LAB 35% (-1): CON 32% (n/c): LD 23% (+1): UKIP 5% (n/c)

    Compared with the YouGov poll at the start of October the Labour lead has been halved. Today’s poll is also good news for the Lib Dems with it returning to its highest share of the year from the pollster.

    YouGov’s methodology tends to produce a more stable share for Labour than the other pollsters and during the whole of 2004 has not moved out of the 34-36% range. Today’s figure is not even at the top of that range. When tested against real results the pollster has proved to be more accurate with the Tory-Labour share than other pollsters although it still over-stated Labour and understated the Tories.

      The move reinforces what we’ve been saying about when in the the month to bet on the General Election. Those polls that report at the turn of the month – YouGov and Populus – generally show smaller Labour shares than those that come out mid-month.

    If you want to bet on Labour or against the Tories do it after the markets have digested the latest information. If you want to bet against Labour or for the Tories do it after ICM.

    Today’s poll shows the impact that UKIP continues to have on the Tories and this is where, we believe the election will be decided. If UKIP remains at 4-5% then the Tories will continue to struggle and Labour will win a substantial majority.

    If Michael Howard and his Austrailian Campaign Director, Lynton Crosby, can find a way of dealing with the UKIP issue then the outcome will be much closer. They might take some comfort in the fact that YouGov, when tested, does over-state UKIP significantly.

    Full round-up of General Election and other political bets – click here.



    Did this man’s tummy bug change the course of British Politics?

    Thursday, November 25th, 2004

    For sometime now we’ve been meaning to draw users attention to the extraordinary internet book by the polling blogger, operator of UK Pollingreport and regular contributor to Politicalbetting, Anthony Wells.

    It’s called “What if Gordon Banks had played” and is a fun-packed counterfactual political history of the 1970s based on the premise that Banks played in the quarter-final of the 1970 World Cup three days before the General Election that ended Harold Wilson’s Government and saw Edward Heath’s Tories win the election. Would Labour have held on if England had won?

    If you’ve got a couple of hours to spare this is well worth reading.

    What’s this got to do with political betting you might ask? Well it shows how little it takes to turn an entire election and it’s a reminder that elections are not always foregone conclusions.

    Latest Labour seat prices:-
    360 + 11/10
    0 – 335 7/4
    352 – 359 11/2
    344 – 351 6/1
    336 – 343 7/1