A fantasy league for competitive League of Legends, built on Riot's public match data. We had the data and we had the formulas, and never got to run one against the other.
Side venture, four founders · Gaming domain expert, co-founder · H2 2018
Fantasy sports work because performance is public, structured, and scored the same way for everyone. Esports produces more raw data than any traditional sport and Riot Games publishes a large part of it through a public API, but there is no agreed way to turn a match record into a player's score. A player's contribution to a win is spread across kills and assists, objectives, vision control, gold efficiency and tempo, and none of those is a fantasy point on its own. The bet was that the scoring model, not the app around it, was the hard and defensible part.
Gaming domain expert in a team of four friends building it after hours. I was the one who played and followed competitive League of Legends, so my job was to say what actually counts in a match: which of the dozens of statistics the API exposes reflect a player's contribution to a win, and which are noise that looks like signal. The scoring formulas were built on that judgement.
Riot's public API as the only data source, with no scraping and no private feeds. Bulk collection of match and player statistics, definition of the player KPIs, and the aggregation formulas converting a match record into a comparable performance value. Product concept and the rules of the game around it.
The four founders, and the competitive League of Legends player base as the intended audience.
Riot Games public API, bulk match and player statistics, scoring model design.
Several gigabytes of match data collected, and the aggregation formulas defined. The two were never brought together: the step from raw data to a running score is exactly where the venture stopped. The development lead became a father, his spare time went to zero, and with four of us working on it in the evenings there was no way to replace him without starting over. We closed it rather than keep a project alive that existed only on paper.
The failure was neither technical nor conceptual. Both halves were there and nobody had the capacity to close the last mile. It is the same pattern worth watching at work: a delivery is not at risk when the design is hard, it is at risk when one person holds the only route to production.