What is goto_conversion?
The Market Index for Sports that powered $47,000 of prize money, 10+ Gold Medals and 100+ Medals on Kaggle. Kaggle is basically an Olympiad for Data Science
The project currently covers basketball, with one page per top-division league and for multinational competitions such as the EuroLeague. Each page compares its sources, every bookmaker separately and goto_conversion, their average, through stage-of-elimination probabilities, match predictions, bookmaker disagreement, starting line-up changes, player values and prediction accuracy.
What do the league page sections mean?
- The Stage of Elimination Table shows each team's points in the league table (its wins in a league that awards none) and its chance of its season ending at each stage, including the chance of winning the title, by date and source; an archived season without stored bookmaker lines shows where each team's season ended, from its results (a season cancelled before its play-offs, as it was ended).
- Next Round's Match Predictions show each source's win probabilities, expected scores and margin and total-points distributions for the next round's matches. The Match Predictions Generator shows the same for any home and away teams, opening on one of the league's biggest rivalries. A bookmaker predicts a match it has not priced only once its recent priced games are all from this season, past the preseason.
- Bookmaker Disagreement shows how much the bookmakers disagree on the winner of each next-round match that two or more of them predict: goto_conversion's home-win probability and the highest and lowest home-win probabilities.
- The Player Changes Table lists the changes to each team's starting line-up over the dates it states, with the change in Adjusted PER, ten at a time by default.
- The Player Values Table ranks each team's six players with the most minutes a game, and every player who averages at least a typical sixth man's minutes, by PER (Off PER + Def PER), ten at a time by default, with their rank, current team, position and age.
- The Team Value Table ranks teams by PER (Off PER + Def PER) over all their current players, each weighted by his minutes, with their players' age weighted the same way, and sorts by any column.
- Prediction Accuracy charts how much probability each source gave the actual results of this season's played matches, preseason included, as a bar chart of each source's average after its last match or, with a switch, a line chart over every match.