Behavioural and neural interactions between objective and subjective performance in a Matching Pennies game

Abstract

To examine the behavioural and neural interactions between objective and subjective performance during competitive decision-making, participants completed a Matching Pennies game where win-rates were fixed within three conditions (win > lose, win = lose, win < lose) and outcomes were predicted at each trial. Using random behaviour as the hallmark of optimal performance, we observed item (heads), contingency (winstay, lose-shift) and combinatorial (HH, HT, TH, TT) biases across all conditions.Higher-quality behaviour represented by a reduction in combinatorial bias was observed during high win-rate exposure. In contrast, over-optimism biases were observed only in conditions where win rates were equal to, or less than, loss rates. At a group level, a neural measure of outcome evaluation (feedback-related negativity; FRN) indexed the binary distinction between positive and negative outcome. At an individual level, increased belief in successful performance accentuated FRN amplitude differences between wins and losses. Taken together, the data suggest that objective experiences of, or, subjective beliefs in, the predominance of positive outcomes are mutual attempts to self-regulate performance during competition. In this way, increased exposure to positive outcomes (real or imagined) help to weight the output of the more diligent and analytic System 2, relative to the impulsive and intuitive System 1.

Switching competitors reduces win-stay but not lose-shift behaviour: The role of outcome-action association strength on reinforcement learning

“…This leads to the possibility that the use of 3+ response games naturally offers more response switch options (at least 2) relative to response repeat option (only 1). If participants are biased towards shift behavior solely as the number of response options increases, then using binary games contexts (e.g., [27,28]) should decrease any lose-shift bias and perhaps even reverse to a win-stay bias. While such accounts seem reasonable, there are counter examples in binary game contexts where lose-shift is dominant [14] It is interesting to note that when comparing unexploitable with exploitable computer opponents, participants report an increased sense of co-presence for an automated program that played randomly [46].…”

A micro-genesis account of longer-form reinforcement learning in structured and unstructured environments

“…Furthermore, RPS represents an intriguing, non-binary paradigm in terms of its relation with traditionally dominant forces of behavioural modification. Specifically, reward mechanisms tend to shape behaviour to a greater degree than punishment mechanisms 12,15 , such that win-stay selections are more frequent than lose-shift within certain simple games (see 16 in the context of cooperative games, see 17 in the context of Matching Pennies). However, RPS can yield an over-use of lose-shift relative to win-stay behaviour [18][19][20][21] .…”

Switching Competitors Reduces Win-Stay but Not Lose-Shift Behaviour: The Role of Outcome-Action Association Strength on Reinforcement Learning

“…This can also give rise to more predictable responding, as seen in the increases in lose-shift behavior relative to win-stay behavior, compared against the values expected by MS performance (e.g., [25,26]). While there are also contexts in which win-stay behavior is more in evidence than lose-shift behavior (e.g., [27,28]), the present concern is the degree to which individuals can regulate the expression of win-stay and lose-shift, and what environmental features might trigger them into doing so.…”

“…One question also remains regarding the extent to which the relative success of win-stay modulation and the relative failure of lose-shift modulation might be due to the initial weighting of these behaviors generated by the specific game context. For example, a typical pattern for performance against an MS opponent in a three-response zero-sum game such as Rock, Paper, Scissors is that the percentage of win-stay behavior is roughly equivalent to that predicted by MS (i.e., = 33.3%) whereas the proportion of lose-shift behavior tends to be larger than that predicted by MS (i.e., >66.6%; e.g., [26,27]). Alternatively, when, as in the present study, both win-stay and lose-shift deviate from expected MS values, lose-shift deviation is greater than win-stay deviation.…”

“…This leads to the possibility that the use of 3+ response games naturally offers more response switch options (at least 2) relative to response repeat option (only 1). If participants are biased towards shift behavior solely as the number of response options increases, then using binary games contexts (e.g., [27,28]) should decrease any lose-shift bias and perhaps even reverse to a win-stay bias. While such accounts seem reasonable, there are counter examples in binary game contexts where lose-shift is dominant [14] and 5 It is interesting to note that when comparing unexploitable with exploitable computer opponents, participants report an increased sense of co-presence for an automated program that played randomly [46].…”