Fixing the win rate in the During phase reduces optimal behaviour rates, indicating suppression of the previously learned pattern¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? In Experiment 1, OBR was significantly reduced from Pre to During and rose again from During to Post. → Zhang 2026 q3 i? In Experiment 2 (25%), OBR was higher in Pre and Post than During, with no interaction with Learning. → Zhang 2026 (2)

Evidence¶

Zhang 2026¶

Zhang, Y., & Dyson, B. J. (2026). Recover or discover: How outcome feedback shapes response patterns within game spaces. npj Science of Learning. https://doi.org/10.1038/s41539-026-00414-w

q3 · i? · causal · r2

Learning x Phase interaction in Experiment 1's three-way repeated measures ANOVA (n = 76): OBR dropped when the win rate was fixed at 50% and recovered when exploitable structure returned; no effect size printed.

"OBR was significantly reduced from Pre to During phases, consistent with the reduction in win rate from Pre to During phases. OBR also significantly rose from During to Post, consistent with the release of the fixed win rate."

Zhang 2026 (2)¶

Zhang, Y., & Dyson, B. J. (2026). Recover or discover: How outcome feedback shapes response patterns within game spaces. npj Science of Learning. https://doi.org/10.1038/s41539-026-00414-w

q3 · i? · causal · r2

Experiment 2 (25% condition, n = 72 analysed) ANOVA produced significant main effects of Phase and Outcome only; the absence of interaction with Learning supported the claim that severe punishment eliminated the relearning-old advantage.

"OBR was also higher in both Pre- (66.71%) and Post- (68.88%) phases relative to During (56.13%; Tukey's HSD, p < .05)."

Discussion¶

Related Claims¶

• Optimal behaviour is more frequent after wins than after losses, but the after-win advantage is absent in the Post phase and under high reinforcement — related
• Suppression of the Pre response pattern between Pre and During phases positively predicts Post phase win rates when learning a new pattern, mainly after wins — related
• Suppression of the old pattern predicts relearning-old success positively under punishment but negatively under reinforcement — related
• High reinforcement (75%, 88% fixed win rate) maintains a relearning-old advantage in Post phase win rates and OBR — related