= Prediction markets monitored during and after a political debate shown on national TV, and compared to the talking heads’ sentiments.
—
Caveat Bettor liked the section of Emile Servan-Schreiber’s blog post about the NewsFutures prediction markets plebisciting French Republican presidential candidate Nicolas Sarkozy, whereas the spin doctors and journalos’s sentiment was 50/50 or favoring his Socialist opponent (the woman).
It’s true. The TV debate was on Wednesday. And on Friday morning, the first post-debate polls were given to the journalos. I listened to some French radios that Friday morning. It was like the sky was falling onto those leftist journalos’ heads. They were stunned to see that, post-debate, it was a 55/45 race —in spite of all their efforts to cheer up the Socialist candidate after the debate.
UPDATE: Mike Giberson has a comment below and on his blog, Knowledge Problem.
Servan-Schreiber’s blog post was a bit unfair to the pundits, no? There is no contradiction between the debate being won by Royal and yet it increasing the likelihood that Sarkozy would win the election.
I’m no expert on French politics, but from what I understand Sarkozy had a strong lead going into the debate. All he had to do in the debate was not blow up, and as the debate continued and he didn’t blow up, the likelihood he would win the election increased.
Now the post-debate polling suggested that, contrary to many pundits, the majority of the people polled felt that Sarkozy won the debate. But that is a “polls vs. the pundits” posting, and not a “prediction markets vs. the pundits” posting.
And by the way, aren’t these conclusions, by pundits and by people polled, generally so afflicted with confirmatory bias as to be generally suspect as indicators of success?
I’ve also posted related remarks over at the Knowledge Problem blog as Prediction Markets vs. the Pundits in the French Election.
“Servan-Schreiber’s blog post was a bit unfair to the pundits, no? There is no contradiction between the debate being won by Royal and yet it increasing the likelihood that Sarkozy would win the election.”
–> No.
–> The debate was clearly won by Sarkozy. His opponent was vague. Plus, she flew into a rage on half-wrong data.
“All he had to do in the debate was not blow up, and as the debate continued and he didn’t blow up, the likelihood he would win the election increased.”
–> Turned out *she* blew it up. She got into a small rage at one point, and it turned out that it is not all clear that she was right.
“But that is a “polls vs. the pundits†posting, and not a “prediction markets vs. the pundits†posting.”
–> That is a POLLS + PREDICTION MARKETS versus the SOCIALIST PUNDITS discussion. What is interesting with the prediction markets (BetFair, InTrade, NewsFutures) is that you don’t have to wait two days to get the inkling.
Thanks for your input, Mike.
Michael is absolutely right that the NewsFutures contracts priced “winning the election” rather than “winning the debate”. However, one must ask what exactly does it mean to “win a debate” if it doesn’t increase your likelihood of winning the election even just a little bit? That’s why I think this episode showed a true disconnect between the pundits and the market, which was resolved in favor of the market.
Emile, I do agree with your general theme about the disconnect between pundits and the market, and your post illustrated the point in a somewhat general sense. My reaction was that the logic underlying the post wasn’t quite as direct as suggested.
On the contrary, I think the logic is as direct as can be: my point is that it makes no sense to argue that a candidate has “won the debate” if that has not increased his or her probability of winning the election. So the pundits arguing that Royal had won the debate were in direct disagreement with the market increasing the price of the Sarkozy contract.