Get backtests

Get one Research backtest's calibration: the situation every twin read, the study's single-choice questions, which one asks the decision and how much of usual buying each of its options keeps (the member's mapping), the outcome the member recorded, and every run of the study: what it asked (its respondent models and any context its twins read after the situation), how many twins chose each option, the forecast change those answers add up to with its 95% range, how far it stands from the outcome in points, whether both moved the same way and the range holds the outcome, and the share of answers it holds in common with the first run. Also how many twins a rerun would ask. A campaign that is no backtest is refused 404. id is the campaign's id.