Scenario 1 (Neutral) Payoff Correlation = 0.00
This is the baseline scenario from which we can evaluate the other four.  In this scenario, the Payoff Correlation was 0.00.  This means that there was no relationship between the payoffs within an encounter matrix cell for the two Agents.
K1
Self-Gain was almost perfectly predictable from Basic Strategy alone.  The rank of strategies was:
  • Best Row showed the most Self-Gain.
  • Minimize Loss
  • Best Cell
  • Assume Selfish
  • Assume Benevolence
  • Assume Persecution
  • Assume Death-Wish
  • Least Surprise
  • Most Surprise showed the least Self-Gain.  This was the one strategy with negative Self-Gain.
Good/Bad Will and Reciprocity had little effect on Self-Gain.  An Agent did slightly better whenever either of these Altruism adjustments was set to Zero.  Random Altruism settings slightly detracted from Self-Gain.
K2
The average Self-Gain across all encounters was greater than zero.  So even when there was no correlation between cell payoffs, the population of Agents, as a whole, was able to obtain positive Self-Gain.  This is a significant result.  Random decisions would have generated an average Self-Gain of zero.
K3
Basic Strategy had no effect on Other-Gain.  In contrast, the Altruism adjustments were the major predictors of Other-Gain.  In this Neutral scenario, Good/Bad Will and Reciprocity contributed equally to overall Altruism.
  • Good Will increased Other-Gain.
  • Bad Will decreased Other-Gain.
  • Responsive (positive Reciprocity) increased Other-Gain.
  • Contrary (negative Reciprocity) decreased Other-Gain.
  • Zero and Random levels of Good/Bad Will or Reciprocity had no effect on Other-Gain.
It is obvious how inherent Good/Bad Will would affect Other-Gain.  The effect of Reciprocity requires explanation.  Recall that the average encounter resulted in a minor positive Self-Gain.  A Responsive Agent would "return the favor" during the next encounter with the same Agent.  In contrast, a Contrary Agent would respond to gain with Antagonism.  So as long as the average encounter had positive results, the average Responsive Agent would make some effort toward Altruism in the next encounter.  If the average Self-Gain of all encounters were negative, we could expect a negative relationship between Reciprocity and Other-Gain.
Commonwealth, the sum of Self-Gain and Other-Gain, was best predicted from an equal combination of Basic Strategy and Altruism.  Basic Strategy strongly affected Commonwealth, but only so far as it contributed to Self-Gain.  Altruism strongly affected Commonwealth, but only so far as it contributed to Other-Gain.  The strategies that best contributed to Commonwealth were the successful Self-Gain strategies (Best Row, Minimize Loss, Best Cell, and Assume Selfish) accompanied by Good Will and/or Responsiveness.  The strategies that detracted from Commonwealth were the unsuccessful Self-Gain strategies accompanied by Bad Will and/or Contrariness.
K4
Equality, the complement to Commonwealth, was affected only by Basic Strategy.  Good/Bad Will and Reciprocity had no effect on the equality of payoffs.  The rank of Basic Strategies (from most equal to most unequal) were:
  • Least Surprise showed the most equality between Agent payoffs.
  • Minimize Loss
  • Assume Persecution
  • Assume Selfish
  • Best Row
  • Assume Death-Wish
  • Assume Benevolence
  • Best Cell
  • Most Surprise showed the most inequality between Agent payoffs.