In the cover letter (and in the paper), we state that no strategy performs well generally. However for noise and prob_end < 0.1, Evolved FSM 16 Noise 05 does quite well in all three scenarios according to Table 3. Evolved HMM 5 does pretty well in these cases as well. Is the performance of these strategies consistent with the state that no strategy performs well broadly?
Changes LGTM
In the cover letter (and in the paper), we state that no strategy performs well generally. However for noise and prob_end < 0.1, Evolved FSM 16 Noise 05 does quite well in all three scenarios according to Table 3. Evolved HMM 5 does pretty well in these cases as well. Is the performance of these strategies consistent with the state that no strategy performs well broadly?