Table 8. Classification of players based on the estimated parameter values. The parameters are defined in Table 2. The 248 players in the 10-round experiments are classified into 16 possible cases. The total number for each type of learning model is given in the last row, and the total number for each type of utility function is given in the last column. The 30 players of the 40/60-round experiments are also classified, and these numbers are given in parentheses.


Cournot
(λ < 0.5; δ > 0.5)
Belief learning
(λ > 0.5; δ > 0.5)
Reinforcement learning
(λ < 0.5; δ < 0.5)
Reinforcement learning with forgetting
(λ > 0.5; δ < 0.5)
Total
(sum of columns)
Inequity aversion
(ρ > 0.1; χ < -0.1)
32
45
11
60 (10)
143 (10)
Social welfare
(ρ > 0.1; χ > 0.1)
27 (1)
32 (12)
3
11 (6)
73 (19)
Competitive
(ρ < -0.1; χ < -0.1)
1
2
1
3
7 (0)
Selfish
(-0.1 < ρ < 0.1; -0.1 χ < 0.1)
6
5 (1)
1
8
20 (1)
Total
(sum of rows)
66 (1)
84 (13)
16 (0)
82 (16)
248 (30)