Skip to main content
Jury’s main feature revolves around its rank aggregation system. As judges view projects, they are asked to provide input through ranking and starring projects. These metrics will be aggregated to form a comprehensive ranking. In the following section, we go more in-depth on how this is done.

​
Ranking

The basis of Jury’s algorithm is the Copeland Counting method. A judge will rank as many projects as they can (ideally as many as possible). See the Judging Interface page for a visual on the judges’ UI. The Copeland method works by breaking down rankings into pairwise comparisons. For example, if a judge has ranked projects in the order A, B, C, it would mean the same as the following comparisons: (A, B), (B, C), (A, C), where we make (winner, loser) pairs. For every time a project is a “winner,” it will get one point; every time it’s a “loser,” it will lose a point. These point values are then aggregated across all judges’ rankings to form a final score for each project. This method was chosen for its simplicity while supporting uneven counts of partial rankings between different judges.

​
Starring

Projects can be starred from the finish judging popup: finish judging popup Or from the ranking menu: ranking menu Stars are simply added up for each judge. We introduced stars as a way for judges to pick their absolute favorites out of the projects that they’ve seen. If they saw a project that they think should win the hackathon, we encourage judges to star those projects. There is no limit to how many projects a judge can star, but we recommend they star at most 3. The star system allows for judges to provide more information to the organizers and allows for organizers to get a better sense of which projects were most loved. We recommend using the score system mainly but considering the number of stars as a secondary metric.

​
Final deliberation

Once all projects are judged, organizers will go into a room to do final deliberation. They will use the scores from Jury to determine the final ranking of projects, as well as the rankings among any tracks. To help in the final decision, it’s a good idea to send organizers out before final scores are tallied and view the top ranked projects on Jury, especially if there are multiple that stand out among the best.