Watts in a game? Electricity trading for fun and (imaginary) profit forms basis for weeklong online game during MIT’s IAP

Tom Witkin is a recently hired official at the Massachusetts Department of Energy Resources, but on Monday, January 11, he was sitting in a classroom at MIT learning how to play a game. “My job is to help facilitate the transformation of energy markets,” he says, and the game was actually right on target. He was learning about how today’s wholesale electricity markets actually work.

Energy trading is the kind of thing that made possible the rise—and the calamitous fall—of a company like Enron. More mundanely, it’s the system that ultimately governs the prices we all pay for electricity.

Usually it is a process that takes place well out of the view of ordinary citizens. But at the MIT event, 14 teams of students, industry and government energy specialists, and at least one high school student who hopes to attend MIT, were playing the Fantasy Electricity Strategy Game. They were learning firsthand the ins and outs of how a utility makes its money in the day-by-day, hour-by-hour process of predicting electricity usage, deciding what power plants to turn on, and setting the price for the kilowatts they produce. In this case it’s only a game, but it contains the essence of a process that hits everyone directly in the pocketbook.

Much like the way games such as fantasy baseball try to mimic the decision-making that guides the real business of sports, the Fantasy Electricity Strategy Game is designed to be as faithful as possible to the way the real electricity market works. The one-week game began on Monday with an introductory lecture that outlined how the process works, and then students broke up into teams to play the game. “I thought it was informative and lucid,” Witkin says of the first day’s introduction to the game—and to the way electricity trading works.

At the conclusion of that first day, each team had to make their first set of decisions: which of seven portfolios of power plants of different types the team would “buy” and how much they were willing to pay. (Because of the number of participants, the seven portfolios were duplicated, with teams divided among two different “worlds” that each had the same array of plants.)

The portfolios, which included plants powered by natural gas, coal, oil, hydropower, and nuclear energy, were sold in an open auction format. While the seven portfolios ranged from 2300 megawatts to 3900 megawatts in generating capacity, right off the bat big differences emerged between the teams. One team paid a high of $150,000 for its lineup of plants, while another paid just $35,000.

“My team surprisingly won one of the best portfolios for the lowest bidding price,” says Fernando Jose de Sisternes Jimenez, a graduate student in the Sloan Automotive Laboratory in the Technology and Policy Program. That portfolio included a hydro plant, a coal plant, and two natural gas combined-cycle plants.

“Normally, they would have more time to think about it,” says game co-organizer Mar Reguant-Rido, a doctoral student in economics and an ABB-MIT Energy Fellow. The game was originally developed by a professor at the University of California, Berkeley, and has also been played at Yale, and in both cases is usually spread over a series of several weeks, with lectures along the way filling in details about how the electricity markets work.

The auction here last week yielded “very low prices,” Reguant-Rido says, “so people will be making very positive profits.” If the game were repeated a second or third time, people would have had a chance to learn more about the real values of the plant portfolios, and the prices paid might be more realistic. “Most of the teams were quite concerned about paying too much” in this initial game, she says.

During the week, each team (working independently, with no contact with the other teams after the initial session) made a daily bid, offering their price for electricity hour-by-hour for the next day, based on that day’s projected demand. Then, they each got feedback about what the “actual” prices set for that day turned out to be (which were set by the highest bid needed to meet each hour’s projected demand, the way prices are in fact set on the electricity markets).

Witkin says that his three-person team decided to divide up the bidding, with each member preparing the bids for one of the days. For the first day’s bid, one team member, who works in the software industry, devised a detailed computer algorithm to evaluate the possibilities and generate a bid. For his own contribution on the second day, Witkin says, “I did it much more by gut feeling.” He spent about an hour working on the bid in the evening in his home office and then emailed the results to his teammates, as well as to the game’s organizers.

Then, the next morning, each team got an email about the outcome of that day’s bidding—but they only saw the final bid and how it affected their own performance. Not until the end of the game on Friday did they get to see exactly what each of the other teams did.

“We received the results already from the first session,” de Sisternes said during the week, “and we were not surprised that we had already recovered our initial costs, as the bidding price for our portfolio was very low.” As the game went on, the teams learned more, he added: “With more sessions run, perhaps we would be able to have a sense of the strategies of other participants and try to bid close to the marginal price, to optimize our gains.”  

Just to make it more realistic, there was one surprise thrown in. After the third day, without prior notice, a carbon tax was added to the costs. “There’s always uncertainty in the market” in real life, Reguant-Rido explains, so the organizers wanted to introduce an extra element of uncertainty into the game.

Witkin thinks that it would be useful to have even more detailed information provided, but given the time constraints he and others faced, “I think it was a worthwhile exercise. It was interesting and fun to actually put some of my policy and economics education back into play,” after having spent years in the energy consulting and software businesses, he says.

“I was surprised to see the level of engagement” of the participants, Reguant-Rido says. “It’s just a game, but people take it very seriously. They wanted to get a lot out of it.”

— David L. Chandler, MIT News Office

Posted 12 February 2010