Brexit, Predictions and Trump

The bookies, until the votes were being counted, were showing greater than 2:1 odds against Brexit in yesterday’s referendum. The subsequent Brexit victory appears to confirm the hypothesis that many Brits were lying to pollsters.

The bookies are showing odds of around 3:1 against a Trump victory in our presidential election. Arguing predictions is a fool’s game, but it may be that our election polls are wrong for the same reason as the Brexit polls apparently were. The Democrats and their media allies have demonized Trump as a racist and misogynist, and it seems likely that many people who intend to vote for him aren’t admitting it. We’ll know soon enough.

10 thoughts on “Brexit, Predictions and Trump”

  1. I was reading awhile back that British bookmakers have been among the most shorted stocks the past couple years because of increased taxation and regulations imposed on them. Widening budget deficits have made the ‘sin taxes’ popular, and they’ve hit gambling operations hard. I wonder if the bookies are skewing the odds to curry favor with the bureaucrats that are tightening the leash on them?

  2. I’ve been blasting every version of “Rule, Britannia!” that I can find this morning. My wife came downstairs asking “what is that?”

    “That, my dear, is the smell of Churchill’s cigar”

  3. Muslims take over London.

    Henry VIII emerges from his grave to vote to save the English church.

    Dies irae is come.

  4. Heh. Reminds me of the way Mike Royko would occasionally use his column to advise people to lie to pollsters in general and to exit pollsters in particular as a way of frustrating their attempts to use their polls to influence the elections. Nice to see that Mike Royko’s advice is being taken. ^_^

  5. I don’t think it’s people lying to pollsters, it’s the pollsters who are lying, albeit somewhat unintentionally.

    Polling has been infected with same pseudo-scientific policy of adjusting the raw polling data to match ‘reality’ as climate ‘science’ and seemingly other disciplines. Back in the day it was relatively laborious to normalize raw data against population characteristics so the data was relatively untouched beyond what was needed to avoid a ‘Dewey defeats Truman’ debacle. Modern computers has expanded the number of characteristics that are checked, the most pernicious among them being normalizing against party affiliation or political philosophy. In essence pollsters are playing with the numbers until the results ‘look right’. Given that the people running polling concerns are, with a few notable exceptions, the same elites that are running everything else their concept of ‘looks right’ often isn’t.

  6. Check the most recent edition of WIRED to see what real data collection looks like for private polling. These are the folks that started working for Obama in 2008 and have been refining their BIg Data results since them. I get queasy when reading about how they track people through their facebook and twitter feeds and were accurate in the 2012 election down to the county level. You no longer have any privacy if you use these services. Their complaint is that the Supervisors of Elections list are not accurate enough to use (not news to me or to anyone who has tried to work on Elections recently). So Presumably the bookies were not willing to pay a data service to correctly set the odds. Several days ago CNBC noted in passing that the bookies were getting a lot of small bets for Leave, but the big money was larger in volume, with fewer bettors for Remain.

Comments are closed.