Skip to Main Content
Table 3: 

Average expert’s payoff over 1000 simulations against different DMs. The table is split into five sections, from top to bottom: Our model (AE), static rules, dynamic rules, algorithms, and the results in the numerical communication setup, which are not directly comparable to the above, text-based communication results. For each condition, we report the average expert payoff over our 1000 simulations, as well as 95% CI (in brackets, using bootstrap re-sampling with 1000 re-samples of our original 1000 simulations; see Dror et al., 2018). The human experts in the experiments ofApel et al. (2020) achieve an average payoff of 7.36.

Expert/DMHC-LSTMBERT-LSTMHC-LSTM+0.1HC-LSTM+0.2HC-LSTM-0.1HC-LSTM-0.2AVG
AE 7.12 [7.02, 7.22] 7.04 [7.03, 7.29] 8.10 [8.02, 8.19] 8.77 [8.70, 8.84] 6.04 [5.93, 6.20] 5.02 [4.90, 5.13] 7.02 
 
RAND 6.54 [6.49, 6.70] 6.67 [6.56, 6.77] 7.56 [7.47, 7.65] 8.31 [8.24, 8.38] 5.58 [5.47, 5.68] 4.49 [4.38, 4.60] 6.53 
MEDIAN 6.46 [6.37, 6.54] 6.85 [6.76, 6.96] 7.24 [7.16, 7.33] 8.02 [7.96, 8.11] 5.45 [5.37, 5.54] 4.66 [4.56, 4.76] 6.45 
HIGHEST 6.77 [6.65, 6.89] 7.82 [7.73, 7.92] 7.94 [7.84, 8.04] 8.82 [8.74, 8.89] 5.55 [5.42, 5.68] 4.46 [4.33, 4.58] 6.89 
EXTREMIST 6.21 [6.11, 6.32] 6.86 [6.76, 6.96] 7.24 [7.14,7.34] 7.99 [7.92, 8.09] 5.14 [5.04, 5.26] 4.08 [3.97, 4.19] 6.25 
A-LIAR 6.54 [6.42, 6.65] 7.14 [7.06, 7.28] 7.15 [7.06, 7.28] 8.69 [8.61, 8.77] 5.40 [5.28, 5.51] 4.35 [4.24, 4.47] 6.55 
PTD-HC 6.88 [6.78, 6.99] 7.03 [6.86, 7.06] 7.68 [7.63, 7.80] 8.49 [8.43, 8.57] 5.83 [5.72, 5.95] 4.92 [4.79, 5.02] 6.83 
PTD-BERT 6.79 [6.67, 6.88] 6.59 [6.51, 6.73] 7.72 [7.63, 7.82] 8.46 [8.38, 8.54] 5.77 [5.64, 5.88] 4.82 [4.71, 4.93] 6.69 
VM-SM 6.58 [6.50, 6.71] 7.00 [6.91, 7.12] 7.70 [7.60, 7.77] 8.34 [8.26, 8.41] 5.65 [5.58, 5.79] 4.67 [4.57, 4.80] 6.66 
AE-DM2 7.05 [6.93, 7.14] 7.23 [7.13, 7.33] 7.94 [7.86, 8.02] 8.66 [8.58, 8.73] 5.92 [5.84, 6.07] 4.97 [4.89, 5.10] 6.96 
AE-VM2 7.03 [6.93, 7.13] 7.05 [6.96, 7.17] 8.00 [7.93, 8.09] 8.76 [8.72, 8.85] 5.98 [5.88, 6.09] 4.98 [4.90, 5.12] 6.97 
AE-SG 7.53 [7.39, 7.64] – 8.63 [8.48, 8.65] 9.10 [9.09, 9.22] 6.02 [5.93, 6.23] 4.85 [4.61, 4.93] 7.23 
Expert/DMHC-LSTMBERT-LSTMHC-LSTM+0.1HC-LSTM+0.2HC-LSTM-0.1HC-LSTM-0.2AVG
AE 7.12 [7.02, 7.22] 7.04 [7.03, 7.29] 8.10 [8.02, 8.19] 8.77 [8.70, 8.84] 6.04 [5.93, 6.20] 5.02 [4.90, 5.13] 7.02 
 
RAND 6.54 [6.49, 6.70] 6.67 [6.56, 6.77] 7.56 [7.47, 7.65] 8.31 [8.24, 8.38] 5.58 [5.47, 5.68] 4.49 [4.38, 4.60] 6.53 
MEDIAN 6.46 [6.37, 6.54] 6.85 [6.76, 6.96] 7.24 [7.16, 7.33] 8.02 [7.96, 8.11] 5.45 [5.37, 5.54] 4.66 [4.56, 4.76] 6.45 
HIGHEST 6.77 [6.65, 6.89] 7.82 [7.73, 7.92] 7.94 [7.84, 8.04] 8.82 [8.74, 8.89] 5.55 [5.42, 5.68] 4.46 [4.33, 4.58] 6.89 
EXTREMIST 6.21 [6.11, 6.32] 6.86 [6.76, 6.96] 7.24 [7.14,7.34] 7.99 [7.92, 8.09] 5.14 [5.04, 5.26] 4.08 [3.97, 4.19] 6.25 
A-LIAR 6.54 [6.42, 6.65] 7.14 [7.06, 7.28] 7.15 [7.06, 7.28] 8.69 [8.61, 8.77] 5.40 [5.28, 5.51] 4.35 [4.24, 4.47] 6.55 
PTD-HC 6.88 [6.78, 6.99] 7.03 [6.86, 7.06] 7.68 [7.63, 7.80] 8.49 [8.43, 8.57] 5.83 [5.72, 5.95] 4.92 [4.79, 5.02] 6.83 
PTD-BERT 6.79 [6.67, 6.88] 6.59 [6.51, 6.73] 7.72 [7.63, 7.82] 8.46 [8.38, 8.54] 5.77 [5.64, 5.88] 4.82 [4.71, 4.93] 6.69 
VM-SM 6.58 [6.50, 6.71] 7.00 [6.91, 7.12] 7.70 [7.60, 7.77] 8.34 [8.26, 8.41] 5.65 [5.58, 5.79] 4.67 [4.57, 4.80] 6.66 
AE-DM2 7.05 [6.93, 7.14] 7.23 [7.13, 7.33] 7.94 [7.86, 8.02] 8.66 [8.58, 8.73] 5.92 [5.84, 6.07] 4.97 [4.89, 5.10] 6.96 
AE-VM2 7.03 [6.93, 7.13] 7.05 [6.96, 7.17] 8.00 [7.93, 8.09] 8.76 [8.72, 8.85] 5.98 [5.88, 6.09] 4.98 [4.90, 5.12] 6.97 
AE-SG 7.53 [7.39, 7.64] – 8.63 [8.48, 8.65] 9.10 [9.09, 9.22] 6.02 [5.93, 6.23] 4.85 [4.61, 4.93] 7.23 
Close Modal

or Create an Account

Close Modal
Close Modal