Median and IQR parameter fits for alternative value functions and effort models in the EFF task, with posterior model frequencies and protected exceedance probabilities (pxp).

The winning model is the loss aversion and nonlinear value encoding model, with Bayesian omnibus risk P0 = 4.6e-7.

本次竞赛的获胜模型为损失厌恶（loss aversion）与非线性价值编码模型，其贝叶斯综合风险P0 = 4.6e-7。