I remain a believer in prediction markets, but it appears that I’ve been quoting the wrong betting odds on the Presidential race. The Intrade market is small and thin compared with the UK gambling site Betfair, and there seems to be some evidence of Romneyites trying to generate false market signals. Right now, Intrade gives Romney about a 40% chance of winning, while Betfair rates his chances no better than 33%.
Footnote Why markets rather than models? Because no model that has a probability number as its outcome can account for the risk that the model itself may be wrong, or that some external event could render it irrelevant. Of course markets can’t tell you about the future, either: all they can do is integrate all the predictive information available in the present, including the efforts of all the model-builders.
I remember something similar happening last cycle, with McCain. Now that intrade is charging monthly ($5) rather than per-transaction, these sorts of price distortions should be ironing out more quickly.
I’d suggest that another advantage of markets is their inclusiveness. Whereas the community developing models is pre-selected from certain professions and backgrounds, anyone with a few hundred dollars and enough drive to send a cashier’s check to the UK/Ireland can contribute their judgement to Intrade/Betfair.
That Nate Silver post is absolutely must read.
As far as the rogue trader’s $18,000 purchase of Romney stock goes:
It fell flat because the rich dolt purchased at Romney’s wave crest.
If the idea is to catapult the momentum (and it obvious was) you’ve got to tweak the sine wave before it gets to pi/2.
Buy at 45 degrees. Sell at 80. The rich dummy (whose nickname should really be “Shorty”) was late to the tit…
By the way: it’s all Obama from here on out: Shorty and Rove have shot their load.
The smart move is to bet Obama on InTrade for $0.60, and Romney on BetFair for $0.33,
and then either way you make $1 back for a return of 7.5% within 2 weeks with
practically no risk …
Anyhow, there’s pretty weird gap between the InTrade odds and the quantitative estimates
of the best polling gurus (e.g. Sam Wang, Nate Silver) - Obama looks to be ahead in
most of the swing states, and reports of the ground game and early voting suggest he
may outperform the polls by 1.5% or so. I think the gamblers must think there’s a
substantial chance of a repeat of 2004, where - for reasons still not satisfactorily
explained - the official vote didn’t match the polls (even the exit polls!).
“…for reasons still not satisfactorily explained – the official vote didn’t match the polls (even the exit polls!).”
Exit polls are the major standard by which international observers assume election fruad (i.e. - when the apparent votes don’t match the exit polls). In any other country, the assumption would have been that the electronic machines (notably in OH) were fixed, given Tthat ten, as now, the companies that supply those had strong Rep ties. Of course, that would never happen in the USA….
Mark, your critique applies to markets, as well.
Where are the arbs? This is a classic “middle”.