What are the stylized electricity market games?

We've pre-created six stylized scenarios - a base game and five variants that differ from the base game in only one dimension of market structure so students can easily see the impact. In these stylized games the students play generating companies. In five of the scenarios the only thing players do is to place bids for their generating units at various points in the day, in the sixth scenario they additionally can trade carbon permits.

  1. Base game (uniform-price auction, no transmission constraint, no forward contracts, no carbon cap)
  2. Use Pay-as-bid auction
  3. Add Transmission constraint
  4. Add Forward contracts (the generating companies will be pre-assigned forward contracts but can't trade them, there are four variants with differing levels of contracts)
  5. Add a carbon tax
  6. Add a carbon cap (generating companies will have tradeable carbon permits)

For further information please see games for more info.

Where has the electricity trading game been used?

How long does a game take?

Each game is 4 rounds (plus an extra round of carbon permit trading in the 4th stylized scenario in carbon markets). You control how long the round will be. We've run games where each round is 3-4 minutes so a game could be 16-20 minutes. If you give your students more time a game could stretch to 40-50 minutes. You can also make the students submit bids for all 4 periods at the same time and then increment them all at once which could make a game as short as 10 minutes or so. In one of Frank's undergrad classes we ran through three games each in two two-hour sessions. The scenarios are semi-independent so one wouldn't need to run through all of them. If you have lots of students you could have multiple copies of the same game running at the same time or if they find it worthwhile play the same game again.

What are ways the game has been used in a real class?

Lecture followed by illustrative game or two:

The pre-created stylized scenarios illustrate various elements of electricity markets which feature a variant of a base game that differ in only one dimension of market structure. What Frank and Mark having been doing in their business school class (which are 110 minutes long) and workshops for electricity market regulators is to have a short lecture illustrating a concept and then to have the students play a game in order to have them directly engage with the topic. Sometimes this is preceded by the base game and then the stylized game so they can directly see the difference in market dynamics. They then hold a short discussion with the student to discuss the outcomes, why the market behaves differently, and player strategies and reinforce their understanding of the material.

Background readings and then play multiple games so they get a taste of a breadth of electricity markets:

Instead of exploring one idea in depth in Frank's undergraduate freshman seminar and Law school regulatory economics they've assigned background reading and then in class play 3 games in a two-hour period experiences multiple different configurations of an electricity market.

Battle royale:

In the MBA class they've played 4-7 rounds of the multiple rounds of the same base game (different seed each round) where each player plays a portfolio by themselves and after each round players shuffle to new games against new players.

Conference break activity:

The game has been used in conferences where the game is played on break and each break corresponds to one round. The gamemaster screen has the tab "Toggle enjoy talk" which can disable players from playing the game during talks or lectures.


We've also run experiments where half the players play one version of the game and the other half play the another version. Can test which differences emerge.

How to I create a stylized game?

Once you register you'll be given a link which will allow one to create as many of the "stylized" games as you'd like which will be hosted on one of our servers. Each time you type enter on the form it will give you a link for the "gamemaster" and seven links of various generating companies "gencos". You'll might want create more than one game for each scenario if the class is large.

How many stylized games can I create?

Currently we are not imposing any limit on number of games you can create. We may want to create more than one game for each scenario if the class is large and/or to do practice rounds.

How should I share games with players?

When you create a game a list of links are created that allow each team to logon to their game. You can then directly give those links to players or teams. We've found it useful to directly copy and paste the links directly into Google Docs, Google Sheets, or any e-mail client that supports html possibly adding personalized names.

What is the difference between "automatic" and "manual" trading tabs?

In the "Carbon markets" scenario gencos may also trade carbon permits. The difference between the "automatic" and "manual" trading tabs is how the trading data is provided (automatically refreshing via javascript) or loading once without any javascript. If latency is an issue in your room than "automatic" is better but some players with bad wi-fi connections might have a better experience with the "manual" option.

Can the game be played on a tablet or phone?

Yes, people have done so although a laptop is preferred.

Is Internet Explorer fully supported?

Some features do not work in some versions Internet Explorer, we strongly recommend using a different browser such as Firefox, Chrome, Chromium, Safari, etc.

What are all the gamemaster tabs?

The "gamemaster" tab has a bunch of tabs but the most important one are the "Increment period" which clears a period of electricity production. In the "Carbon markets" you must also additionally increment an additional period in order to close the carbon market and assess any fees for failing to comply with the cap. "View Gencos" lists all the gencos, the underlined "Pin" value of each generating company takes you to that teams page. "Market results" tab has a bunch of interesting output files (many of which can be viewed by the players as well).

See Gamemaster tutorial for more information.

How should we allocate the teams?

The game provides a fun learning experience if the teams are allocated randomly but if you'd like to have a "competition" to determine who did best a random allocation would give some teams an unfair advantage. We've controlled for this in various ways in class, if many games are played one can regress team profits on portfolio fixed effects and team fixed effects - whichever teams fixed effect is highest "did best". You could also before the game auction off the portfolios, one possible procedure is the following:

  1. Setting a reservation value of 0 for all portfolios, each teams places bids on all portfolios
  2. The team with the highest bid receives their preferred portfolio and pays the highest bid from another team, remove all bids for that portfolio and that team
  3. With remaining teams and bids the team with highest bid receives their preferred portfolio and pays highest bid from another team, remove all bids for that portfolio and that team
  4. Repeat until only one team and portfolio is left, give it to them at a price of zero

If you do run an auction you can modify the teams starting cash in the "Gencos" tab in the gamemaster view.

Can we get the source code in order to play the game?

Currently we are not giving out the source code but are hosting some stylized games for interested parties to play for free on our server.

What is the distribution of renewable wind/solar plants (for games that have them)?

You can see the expected solar/wind output by clicking on "View capacity". You can see power plant info by clicking on "Game Conditions" and then click on the "Portfolio spreadsheet". The distribution is always zero when the "mu" for that period of the day is zero otherwise it is a censored normal distribution: we draw a normal with the relevant mu and sd and then censor at 0 and 1 and then multiply with the "capacity" of the plant. In particular for the "High Solar" scenario each of the solar power plants are zero at night and are distributed 1300 * CensoredNormal(mu=0.3076921, sd=0.2, lower=0, higher=1). Draws are perfectly correlated within a location, locations are independent.