How can we rank when we can't measure the competition?
There is a lot of research of how to rank players who participate in head-to-head, zero-sum games (think of chess), but far less exists in multi-competitor games. In these games, many competitors are competing against each other under the same conditions toward the same goal.
Obviously sports with these sorts of competition (sailing, golf, horse racing, etc.) care a lot about how to rank players in these games. But by simply reframing the problem, the applications of ranking in multi-competitor games in business settings.
For example, consider a firm with several retail outlets in a geographic area. Each month, it looks at which stores performed best on various metrics, but over time, it can become difficult to have a clear idea of which stores are the best. On a small scale, we can make a judgement call about how stores A, B, and C rank relative to one another. But in the context of a larger organization, an algorithmic approach is necessary.
This comes from the fact that as the series of competitons becomes more and more complex, it is challenging for us to capture the difficulty of each competition. To get around this issue, we create an optimization based approach which measures the skill of each player off their performance and the difficulty of each competition by its attendents.