 sigma      learner  reaches optimum  mean reward  ends near 6
   0.5  independent            0.093         6.54        0.897
   0.5   sequential            0.977        11.86        0.023
   0.5        joint            1.000        12.00        0.000
   1.0  independent            0.170         6.97        0.807
   1.0   sequential            0.817        10.87        0.173
   1.0        joint            0.943        11.65        0.050
   2.0  independent            0.210         7.12        0.727
   2.0   sequential            0.637         9.69        0.320
   2.0        joint            0.777        10.51        0.160
