Any thoughts and how this relates to marketing management During this period, the market was correct in 11 out of 14 times. Whenthe market missed a forecast, the error was never more than one level of sever-ity away. This success enables health officials to better contain an influenza out-break.SummaryThe above examples show that prediction markets can work in a widerange of industries even in cases where forecasting is challenging. These marketswork because they could effectively aggregate information from a diverse groupof individuals.Potential PitfallsWhile the prediction market can be a powerful tool, it is not a panacea forall forecasting situations. There are many common pitfalls.Prediction markets only aggregate information well if the number of par-ticipants is large. A small pool of participants can limit the amount of liquidity inthe market and hence the market's ability to aggregate information. This runscounter to the principle of having a large crowd, as noted above. The success of aprediction market relies on active participation. If participants are notsufficiently motivated, they will have a low level of trading activities. Conse-quently, prices of the markets will not capture information well. This violates theprinciple of incentives. In combination, these problems lead to low market liq-uidity.A recent application at a computer manufacturer is an illustration of thisliquidity problem. This particular manufacturer experimented using a predictionmarket to forecast sales of computing products. This particular prediction marketsize had a small pool of participants (about 15). Most participants were execu-tives from marketing and finance organizations. Since the size of the stakes wassmall ($50 per person on average) and these executives were busy, they did notpay enough attention to the prediction market. As a result, the market had lowliquidity. Since information aggregation stems from trading activities, this lowlevel of liquidity stopped active participants from executing trades that theywanted and this further exacerbated the problem. Despite these problems, theresults were somewhat encouraging. The generated forecasts slightly beat the icompany's official forecasts in six out of eight times. However, the improve-ments were not substantial enough to justify the cost and time spent by theparticipants. As a consequence, this prediction market failed to be adopted asan ongoing forecasting tool.The principles of crowd and independence imply that people who haverelevant information should be included in the market. Otherwise, the aggregateinformation may be coarse and consequently the forecast may not be accurate.Put differently, the prediction market is a system of "garbage in, garbage out."In 2003, after the U.S. forces took Bagdad, Saddam Hussein was on the run andmanaged to avoid capture for several months. During that time, www.trades-156UNIVERSITY OF CALIFORNIA, BERKELEY VOL 50. NOI FALL 2007ports.com conducted a prediction market on whether Saddam Hussein would becaptured before a certain deadline. The price of the scenario, saying Saddamwould be captured before the deadline, was hovering around 9 cents over thedollar for a long time and only jumped up to 30 cents two days before his actualcapture. This rapid adjustment is typical when new information is assimilatedinto the market. However, even at 30 cents, the market was predicting a strongchance (70%) that Saddam would not have been captured before the deadline.Obviously, the market was inaccurate, predicting only a 9% probability of cap-ture until two days before his capture. This inability of the market to forecastaccurately underscores the importance of the existence of "wisdom" in thecrowd. Therefore, the importance of recruiting individuals who have informa-tion cannot be overemphasized. When the size of the group is limited by cir-cumstances to be small, one can use related methods that are developed forthis specific purpose."Prediction markets can only work well if players do not use other incen-tives to trade. That is, players trade solely to make money in the market. Themarket can fail if players have other incentives. For example, consider a predic-tion market for product demand. A product manager, who will gain a higherlevel of resources if the generated demand forecast is high, may want to misleadthe market by pushing up prices of shares of high demand scenarios. Similarly, aRepublican might want to depress prices of Democratic Party nominee in a polit-ical market in order to influence the ultimate political outcome (e.g.. affect voterturnout). That is the reason why it is important to limit the stake of individuals.in the market. The IEM limits the stake to $500 for the same reason. This sizelimit of the stakes also minimizes incentives for the participants to manipulateultimate outcome. For example, a sales representative, who has bought sharesof low demand scenarios, would not have deliberately reduced his own saleseffort to make the prediction come true.ConclusionsPrediction markets are "smart" markets that are capable of accurately.predicting outcomes of uncertain future events. They have been successfullyused in a wide range of settings and industries. Prediction markets work well ifthe underlying scientific principles (incentive, indicator, improvement, indepen-dence, and crowd) are adhered to. When a prediction market fails to predictfuture outcomes, it is often the case that one of the underlying principles doesnot hold.As long as prediction markets are active, they always contain the mostcurrent "wisdom of crowds." Hence, they are better than a one-time survey inaggregating information from individuals because these same individuals havestrong incentives to learn from each other through the price discovery process.Thus, the participants become smarter over time. In effect, a well-functionedmarket will eventually contain a large group of experts who freely share knowl-edge through their trading behaviors.CALIFORNIA MANAGEMENT REVIEW VOL 50. NOI FALL 2007157

Question

Any thoughts and how this relates to marketing management During this period, the market was correct in 11 out of 14 times. Whenthe market missed a forecast, the error was never more than one level of sever-ity away. This success enables health officials to better contain an influenza out-break.SummaryThe above examples show that prediction markets can work in a widerange of industries even in cases where forecasting is challenging. These marketswork because they could effectively aggregate information from a diverse groupof individuals.Potential PitfallsWhile the prediction market can be a powerful tool, it is not a panacea forall forecasting situations. There are many common pitfalls.Prediction markets only aggregate information well if the number of par-ticipants is large. A small pool of participants can limit the amount of liquidity inthe market and hence the market's ability to aggregate information. This runscounter to the principle of having a large crowd, as noted above. The success of aprediction market relies on active participation. If participants are notsufficiently motivated, they will have a low level of trading activities. Conse-quently, prices of the markets will not capture information well. This violates theprinciple of incentives. In combination, these problems lead to low market liq-uidity.A recent application at a computer manufacturer is an illustration of thisliquidity problem. This particular manufacturer experimented using a predictionmarket to forecast sales of computing products. This particular prediction marketsize had a small pool of participants (about 15). Most participants were execu-tives from marketing and finance organizations. Since the size of the stakes wassmall ($50 per person on average) and these executives were busy, they did notpay enough attention to the prediction market. As a result, the market had lowliquidity. Since information aggregation stems from trading activities, this lowlevel of liquidity stopped active participants from executing trades that theywanted and this further exacerbated the problem. Despite these problems, theresults were somewhat encouraging. The generated forecasts slightly beat the icompany's official forecasts in six out of eight times. However, the improve-ments were not substantial enough to justify the cost and time spent by theparticipants. As a consequence, this prediction market failed to be adopted asan ongoing forecasting tool.The principles of crowd and independence imply that people who haverelevant information should be included in the market. Otherwise, the aggregateinformation may be coarse and consequently the forecast may not be accurate.Put differently, the prediction market is a system of &#34;garbage in, garbage out.&#34;In 2003, after the U.S. forces took Bagdad, Saddam Hussein was on the run andmanaged to avoid capture for several months. During that time, www.trades-156UNIVERSITY OF CALIFORNIA, BERKELEY VOL 50. NOI FALL 2007ports.com conducted a prediction market on whether Saddam Hussein would becaptured before a certain deadline. The price of the scenario, saying Saddamwould be captured before the deadline, was hovering around 9 cents over thedollar for a long time and only jumped up to 30 cents two days before his actualcapture. This rapid adjustment is typical when new information is assimilatedinto the market. However, even at 30 cents, the market was predicting a strongchance (70%) that Saddam would not have been captured before the deadline.Obviously, the market was inaccurate, predicting only a 9% probability of cap-ture until two days before his capture. This inability of the market to forecastaccurately underscores the importance of the existence of &#34;wisdom&#34; in thecrowd. Therefore, the importance of recruiting individuals who have informa-tion cannot be overemphasized. When the size of the group is limited by cir-cumstances to be small, one can use related methods that are developed forthis specific purpose.&#34;Prediction markets can only work well if players do not use other incen-tives to trade. That is, players trade solely to make money in the market. Themarket can fail if players have other incentives. For example, consider a predic-tion market for product demand. A product manager, who will gain a higherlevel of resources if the generated demand forecast is high, may want to misleadthe market by pushing up prices of shares of high demand scenarios. Similarly, aRepublican might want to depress prices of Democratic Party nominee in a polit-ical market in order to influence the ultimate political outcome (e.g.. affect voterturnout). That is the reason why it is important to limit the stake of individuals.in the market. The IEM limits the stake to $500 for the same reason. This sizelimit of the stakes also minimizes incentives for the participants to manipulateultimate outcome. For example, a sales representative, who has bought sharesof low demand scenarios, would not have deliberately reduced his own saleseffort to make the prediction come true.ConclusionsPrediction markets are &#34;smart&#34; markets that are capable of accurately.predicting outcomes of uncertain future events. They have been successfullyused in a wide range of settings and industries. Prediction markets work well ifthe underlying scientific principles (incentive, indicator, improvement, indepen-dence, and crowd) are adhered to. When a prediction market fails to predictfuture outcomes, it is often the case that one of the underlying principles doesnot hold.As long as prediction markets are active, they always contain the mostcurrent &#34;wisdom of crowds.&#34; Hence, they are better than a one-time survey inaggregating information from individuals because these same individuals havestrong incentives to learn from each other through the price discovery process.Thus, the participants become smarter over time. In effect, a well-functionedmarket will eventually contain a large group of experts who freely share knowl-edge through their trading behaviors.CALIFORNIA MANAGEMENT REVIEW VOL 50. NOI FALL 2007157

Accepted Answer

Prediction markets can be a valuable tool for marketing management in various ways. For instance,…