October 28th, 2005, 8:25 pm
QuoteOriginally posted by: StochastixI guess to start you would need to estimate, using historical data, the probability matrix of team A winning if it was losing by x points with y minutes still to play. If you assume all teams to be equal that would be it, but since teams are inequal, you should be considering additional parameters such as the teams relative rankings, their usual scoring pattern in a game and so on. Once you have built a good model for this distribution, pricing derivatives should be easy.No, no, no. First what you estimate are physical measure parameters. On top of that your market is as incomplete as it can be, you cannot do any even approximate hadging. You need some preferences based theory for pricing. You will price differently as your holdings change and other complexities need to be taken into account. It is in fact as complex as it can get.