Big Data Series

The  is an MCM-sanctioned programme to develop a scenario that sees a huge amount of simulated participants pitted against each other, with each participant "learning" how to beat their opponents and getting better over time. The series was originally called the Machine Learning Series, but was renamed as it does not strictly fit that criteria, and Big Data is cooler.

250 participants were initially generated with a bunch of constant and variable traits on July 23, 2017. Notable traits include interest (how often the participant will want to participate), natural skill (representing inate ability) and experience (which increases as the participant participates but deteriorates over time). When the simulations were determined to work as intended and matchmaking ratings (MMR) were introduced, on July 25, a system was developed wherein participants join and leave twice-daily 3-hour sessions, during which they queue to be matched with an opponent. The matchmaker tries to match participants with an opponent of similar MMR but does not necessarily have to, and when an approximately-five-minute match is complete, they may requeue if the 3 hour session has not ended. An influx of new players and slow change of each players' "interest" is intended to keep things fresh long-term.

These twice-daily, three hour sessions of being able to queue for matchmaking – provisionally entitled BDS Rated – are the backbone of the Big Data Series, providing a constant battlefield for participants to test themselves against those of similar ability and acting as an insight in to the best participating players – around 2000 matches are simulated each day. On July the 27th–28th, the first tournament took place, entitled Tournytest, pitting the top 16 players against each other. Another larger tournament, the Elite Series, will take place on the weekend of August 5, consisting of 50 players.

Future planned developments include the development of playstyles (with rock-paper-scissors–esque relationships with each other and simulation of AI deciding when participants should use each playstyle), team games (doubles, triples, perhaps even huge team games), and the simulation of friendships/rivaly between players.