I'm still enthralled by the StrategyPage prediction market.

As of writing, there are 950 points of investment piling on behind my bet that Afghanistan is going to get worse. And no contrarians claiming an improvement. OTOH most people think the US casualty rate will fall in Iraq. Are they expecting things to get better? Or the US to pull out? Or maybe Iraqi factions to be too busy killing each other to bother with Brits and Americans?

The majority also think that the Israeli push into Lebanon is going to hurt Hezbollah. Yeah, right!

I'm struck by a couple of other things too.

It's clear that we can't expect the market itself to ajudicate a disagreement.

Or can we? If the market is shown, statistically, to be reliable, does that mean that you can actually appeal to its opinion to "prove" things? Erm ...

Another question of ajudication is "who decides if a prediction comes true?" Some bets are fairly concrete : casualty rates are either up or down. But what do you make of "America will succeed in creating a stable democratic Iraq." which I'm betting against (at least by the close-date of end of 2007) How will "stable and democratic" be defined? Is there currently a civil war in Iraq or not? Is it currently stable and democratic? In the case of StrategyPage it will be the site's owners who have to make the judgement call.

I'm not saying that this detracts from the value of the market. The point is not to win, but to take part (admittedly with a rational, self-aggrandizing behaviour) and see what comes out of the collective intelligence. But it's an interesting question.

Meanwhile, while we're on the subject of accountability of predictions ...