- Start Date:
June 28, 2018
- Event Start Time:
2:00 PM
- Event End Time:
3:00 PM
- Organizers:
Lazaros Gallos
- Seminar Series:
REU Seminar
- Presenter(s):
David Pennock - Microsoft Research
- Event Location:
DIMACS Seminar room
- Abstract:
In one online contest, the average prediction of 2231 people beat all but 0.3% of individual players. This is a striking, yet typical, example of the wisdom of crowds. I will discuss the design of wagering games to elicit probabilities from as many experts as possible for as little cost as possible, to maximize accuracy. Dozens of mechanisms have been proposed, starting from the parimutuel betting system invented in 1867. I will discuss three of our more recent proposals: the Weighted-Score Wagering Mechanism (2008), the No-Arbitrage Wagering Mechanism (2014), and the Double-Clinching Auction (2017). The DCA is the only known wagering mechanism that is truthful and close to Pareto efficient. I will also discuss our analysis of the Proxy Parimutuel mechanism (1959), showing that it's Pareto efficient and close to truthful, both theoretically and empirically. If I have time, I will talk about our effort to crowdsource literally billions or more combinatorial predictions.