# Direct Reciprocity and Model-Predictive Strategy Update Explain the Network Reciprocity Observed in Socioeconomic Networks

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. The Model

#### 2.1. Tit-for-tat vs. D

- her own strategy, T or D, to be used at round t (in Table 1, second column, Tn and Dn denote the fact that the agent changed strategy after round $t-1$); and
- for each neighbor j, the strategy, T or D, that i expects j to use at round t and the last actions, C or D, played by i and j at round $t-1$.

#### 2.2. Tit-for-two-tats vs. D

#### 2.3. New-tit-for-two-tats vs. D

## 3. Results

#### 3.1. Tit-for-tat vs. D

**Theorem**

**1.**

**Corollary**

**1.**

**Theorem**

**2.**

**Corollary**

**2.**

#### 3.2. Tit-for-two-tats vs. D

**Theorem**

**3.**

**Corollary**

**3.**

**Theorem**

**4.**

**Corollary**

**4.**

**Theorem**

**5.**

**Corollary**

**5.**

#### 3.3. New-tit-for-two-tats vs. D

**Theorem**

**6.**

**Corollary**

**6.**

**Theorem**

**7.**

**Corollary**

**7.**

## 4. Discussion and Conclusions

#### 4.1. Three Versions of the T-strategy

#### 4.2. Comparison with the T-strategy in Dercole et al. (2019)

#### 4.3. Impact on Applied Behavioral Science

#### 4.4. Future Directions

## Author Contributions

## Funding

## Conflicts of Interest

## Appendix A. Theorems and Proofs of Results in Section 3

#### Appendix A.1. Tit-for-tat vs. D

**Proof of Theorem**

**1.**

**Proof of Theorem**

**2.**

#### Appendix A.2. Tit-for-two-tats vs. D

**Proof of Theorem**

**3.**

**Proof of Theorem**

**4.**

**Proof of Theorem**

**5.**

#### Appendix A.3. New-tit-for-two-tats vs. D

**Proof of Theorem**

**6.**

**Proof of Theorem**

**7.**

## References

- Trivers, R.L. The Evolution of Reciprocal Altruism. Q. Rev. Biol.
**1971**, 46, 35–37. [Google Scholar] [CrossRef] - Axelrod, R.; Hamilton, W.D. The Evolution of Cooperation. Science
**1981**, 211, 1390–1396. [Google Scholar] [CrossRef] [PubMed] - Axelrod, R. The Evolution of Cooperation; American Association for the Advancement of Science: Washington, DC, USA, 2006. [Google Scholar]
- Sigmund, K. The Calculus of Selfishness; Princeton University Press: Princeton, NJ, USA, 2010. [Google Scholar]
- Nowak, M.; Highfield, R. SuperCooperators: Altruism, Evolution, and Why We Need Each Other to Succeed; Simon and Schuster: New York, NY, USA, 2011. [Google Scholar]
- Ocampo-Martinez, C.; Quijano, N. Game-Theoretical Methods in Control of Engineering Systems: An Introduction to the Special Issue. IEEE Control Syst.
**2017**, 37, 30–32. [Google Scholar] - Perc, M.; Jordan, J.J.; Rand, D.G.; Wang, Z.; Boccaletti, S.; Szolnoki, A. Statistical physics of human cooperation. Phys. Rep.
**2017**, 687, 1–51. [Google Scholar] [CrossRef][Green Version] - Hofbauer, J.; Sigmund, K. Evolutionary Games and Population Dynamics; Cambridge University Press: Cambridge, UK, 1998. [Google Scholar]
- Perc, M.; Gómez-Gardeñes, J.; Szolnoki, A.; Floría, L.M.; Moreno, Y. Evolutionary dynamics of group interactions on structured populations: A review. J. R. Soc. Interface
**2013**, 10, 20120997. [Google Scholar] [CrossRef] [PubMed][Green Version] - Nowak, M.A. Evolutionary Dynamics; Harvard University Press: Cambridge, MA, USA, 2006. [Google Scholar]
- Szabó, G.; Fath, G. Evolutionary games on graphs. Phys. Rep.
**2007**, 446, 97–216. [Google Scholar] [CrossRef][Green Version] - Nowak, M.A. Five rules for the evolution of cooperation. Science
**2006**, 314, 1560–1563. [Google Scholar] [CrossRef][Green Version] - Nowak, M.A.; May, R. Evolutionary games and spatial chaos. Nature
**1992**, 359, 826–829. [Google Scholar] [CrossRef] - Ohtsuki, H.; Hauert, C.; Lieberman, E.; Nowak, M.A. A simple rule for the evolution of cooperation on graphs and social networks. Nature
**2006**, 441, 502–505. [Google Scholar] [CrossRef] - Konno, T. A condition for cooperation in a game on complex networks. J. Theor. Biol.
**2011**, 269, 224–233. [Google Scholar] [CrossRef][Green Version] - Allen, B.; Lippner, G.; Chen, Y.T.; Fotouhi, B.; Momeni, N.; Yau, S.T.; Nowak, M.A. Evolutionary dynamics on any population structure. Nature
**2017**, 544, 227–230. [Google Scholar] [CrossRef] [PubMed][Green Version] - Roca, C.P.; Cuesta, J.A.; Sánchez, A. Promotion of cooperation on networks? The myopic best response case. Eur. Phys. J. B
**2009**, 71, 587. [Google Scholar] [CrossRef] - Hauert, C.; Doebeli, M. Spatial structure often inhibits the evolution of cooperation in the snowdrift game. Nature
**2004**, 428, 643–646. [Google Scholar] [CrossRef] [PubMed] - Lieberman, E.; Hauert, C.; Nowak, M.A. Evolutionary dynamics on graphs. Nature
**2005**, 433, 312–316. [Google Scholar] [CrossRef] - Santos, F.C.; Pacheco, J.M. Scale-free networks provide a unifying framework for the emergence of cooperation. Phys. Rev. Lett.
**2005**, 95, 098104. [Google Scholar] [CrossRef][Green Version] - Santos, F.C.; Pacheco, J.M. A new route to the evolution of cooperation. J. Evol. Biol.
**2006**, 19, 726–733. [Google Scholar] [CrossRef] - Santos, F.C.; Pacheco, J.M.; Tom, L. Evolutionary dynamics of social dilemmas in structured heterogeneous populations. Proc. Natl. Acad. Sci. USA
**2006**, 103, 3490–3494. [Google Scholar] [CrossRef][Green Version] - Santos, F.C.; Rodrigues, J.F.; Pacheco, J.M. Graph topology plays a determinant role in the evolution of cooperation. Proc. R. Soc. B
**2006**, 273, 51–55. [Google Scholar] [CrossRef][Green Version] - Gómez-Gardeñes, J.; Campillo, M.; Floría, L.M.; Moreno, Y. Dynamical Organization of Cooperation in Complex Topologies. Phys. Rev. Lett.
**2007**, 98, 108103. [Google Scholar] [CrossRef][Green Version] - Ohtsuki, H.; Nowak, M.A. Direct reciprocity on graphs. J. Theor. Biol.
**2007**, 247, 462–470. [Google Scholar] [CrossRef][Green Version] - Poncela, J.; Gómez-Gardeñes, J.; Floria, L.M.; Moreno, Y. Robustness of cooperation in the evolutionary prisoner’s dilemma on complex networks. New J. Phys.
**2007**, 9, 184. [Google Scholar] [CrossRef][Green Version] - Assenza, S.; Gómez-Gardeñes, J.; Latora, V. Enhancement of cooperation in highly clustered scale-free networks. Phys. Rev. E
**2008**, 78, 017101. [Google Scholar] [CrossRef] [PubMed][Green Version] - Gómez-Gardeñes, J.; Poncela, J.; Floría, L.M.; Moreno, Y. Natural selection of cooperation and degree hierarchy in heterogeneous populations. J. Theor. Biol.
**2008**, 253, 296–301. [Google Scholar] [CrossRef] [PubMed] - Pusch, A.; Weber, S.; Porto, M. Impact of topology on the dynamical organization of cooperation in the prisoner’s dilemma game. Phys. Rev. E
**2008**, 77, 036120. [Google Scholar] [CrossRef] - Santos, F.C.; Santos, M.D.; Pacheco, J.M. Social diversity promotes the emergence of cooperation in public goods games. Nature
**2008**, 454, 213–216. [Google Scholar] [CrossRef] - Devlin, S.; Treloar, T. Evolution of cooperation through the heterogeneity of random networks. Phys. Rev. E
**2009**, 79, 016107. [Google Scholar] [CrossRef][Green Version] - Devlin, S.; Treloar, T. Cooperation in an evolutionary prisoner’s dilemma on networks with degree-degree correlations. Phys. Rev. E
**2009**, 80, 026105. [Google Scholar] [CrossRef][Green Version] - Floría, L.; Gracia-Lázaro, C.; Gómez-Gardenes, J.; Moreno, Y. Social network reciprocity as a phase transition in evolutionary cooperation. Phys. Rev. E
**2009**, 79, 026106. [Google Scholar] [CrossRef][Green Version] - Perc, M. Evolution of cooperation on scale-free networks subject to error and attack. New J. Phys.
**2009**, 11, 033027. [Google Scholar] [CrossRef] - Cardillo, A.; Gómez-Gardeñes, J.; Vilone, D.; Sánchez, A. Co-evolution of strategies and update rules in the prisoner’s dilemma game on complex networks. New J. Phys.
**2010**, 12, 103034. [Google Scholar] [CrossRef] - Débarre, F.; Hauert, C.; Doebeli, M. Social evolution in structured populations. Nat. Commun.
**2014**, 5, 3409. [Google Scholar] [CrossRef] [PubMed] - Nowak, M.A.; Sigmund, K. Evolution of indirect reciprocity by image scoring. Nature
**1998**, 393, 573–577. [Google Scholar] [CrossRef] [PubMed] - Hamilton, W.D. The genetical evolution of social behaviour. J. Theor. Biol.
**1964**, 7, 1–16. [Google Scholar] [CrossRef] - Wilson, D.S. A general theory of group selection. Proc. Natl. Acad. Sci. USA
**1975**, 72, 143–146. [Google Scholar] [CrossRef] [PubMed][Green Version] - Jansen, V.A.; van Baalen, M. Altruism through beard chromodynamics. Nature
**2006**, 440, 663–666. [Google Scholar] [CrossRef] [PubMed] - Fletcher, J.A.; Doebeli, M. A simple and general explanation for the evolution of altruism. Proc. R. Soc. Lond. B
**2009**, 276, 13–19. [Google Scholar] [CrossRef][Green Version] - Bolton, G.E.; Ockenfels, A. A theory of equity, reciprocity and competition. Am. Econ. Rev.
**2000**, 90, 166–193. [Google Scholar] [CrossRef][Green Version] - Capraro, V. A model of human cooperation in social dilemmas. PLoS ONE
**2013**, 8, e72427. [Google Scholar] [CrossRef][Green Version] - Capraro, V.; Rand, D.G. Do the Right Thing: Experimental evidence that preferences for moral behavior, rather than equity or efficiency per se, drive human prosociality. Judgm. Decis. Mak.
**2018**, 13, 99–111. [Google Scholar] [CrossRef][Green Version] - Vanberg, V.J.; Congleton, R.D. Rationality, Morality, and Exit. Am. Polit. Sci. Rev.
**1992**, 86, 418–431. [Google Scholar] [CrossRef] - Batali, J.; Kitcher, P. Evolution of altriusm in optional and compulsory games. J. Theor. Biol.
**1995**, 175, 161–171. [Google Scholar] [CrossRef] [PubMed][Green Version] - Boyd, R.; Gintis, H.; Bowles, S.; Richerson, P.J. The Evolution of Altruistic Punishment. Proc. Natl. Acad. Sci. USA
**2003**, 100, 3531–3535. [Google Scholar] [CrossRef] [PubMed][Green Version] - Dercole, F.; De Carli, M.; Della Rossa, F.; Papadopoulos, A.V. Overpunishing is not necessary to fix cooperation in voluntary public goods games. J. Theor. Biol.
**2013**, 326, 70–81. [Google Scholar] [CrossRef] [PubMed] - Semmann, D. Conditional cooperation can hinder network reciprocity. Proc. Natl. Acad. Sci. USA
**2012**, 109, 12846–12847. [Google Scholar] [CrossRef] [PubMed][Green Version] - Grujić, J.; Fosco, C.; Araujo, L.; Cuesta, J.A.; Sánchez, A. Social experiments in the mesoscale: Humans playing a spatial prisoner’s dilemma. PLoS ONE
**2010**, 5, e13749. [Google Scholar] [CrossRef][Green Version] - Grujić, J.; Röhl, T.; Semmann, D.; Milinski, M.; Arne, T. Consistent Strategy Updating in Spatial and Non-Spatial Behavioral Experiments Does Not Promote Cooperation in Social Networks. PLoS ONE
**2012**, 7, e47718. [Google Scholar] [CrossRef][Green Version] - Gracia-Lázaro, C.; Ferrer, A.; Ruiz, G.; Tarancón, A.; Cuesta, J.A.; Sánchez, A.; Moreno, Y. Heterogeneous networks do not promote cooperation when humans play a Prisoner’s Dilemma. Proc. Natl. Acad. Sci. USA
**2012**, 109, 12922–12926. [Google Scholar] [CrossRef][Green Version] - Gracia-Lazaro, C.; Cuesta, J.A.; Sanchez, A.; Moreno, Y. Human behavior in Prisoner’s Dilemma experiments suppresses network reciprocity. Sci. Rep.
**2012**, 2, 1–4. [Google Scholar] [CrossRef][Green Version] - Grujić, J.; Gracia-Lazaro, C.; Milinski, M.; Semmann, D.; Traulsen, A.; Cuesta, J.A.; Moreno, Y.; Sanchez, A. A comparative analysis of spatial Prisoner’s Dilemma experiments: Conditional cooperation and payoff irrelevance. Sci. Rep.
**2014**, 4, 4615. [Google Scholar] [CrossRef] - Rand, D.G.; Nowak, M.A.; Fowler, J.H.; Christakis, N.A. Static network structure can stabilize human cooperation. Proc. Natl. Acad. Sci. USA
**2014**, 111, 17093–17098. [Google Scholar] [CrossRef][Green Version] - Li, X.; Jusup, M.; Wang, Z.; Li, H.; Shi, L.; Podobnik, B.; Stanley, H.E.; Havlin, S.; Boccaletti, S. Punishment diminishes the benefits of network reciprocity in social dilemma experiments. Proc. Natl. Acad. Sci. USA
**2018**, 115, 30–35. [Google Scholar] [CrossRef] [PubMed][Green Version] - Dercole, F.; Della Rossa, F.; Piccardi, C. Direct Reciprocity and Model-Predictive Rationality Explain Network Reciprocity over Social Ties; Scientific Report; Nature Publishing Group: London, UK, 2019. [Google Scholar]
- DiMeglio, A.; Dercole, F.; Della Rossa, F. Direct reciprocity and model-predictive rationality: A setup for network reciprocity over social ties. In Proceedings of the 2019 18th European Control Conference (ECC), Naples, Italy, 25–28 June 2019; pp. 1531–1536. [Google Scholar]
- Grilo, C.; Luís, C. Effects of asynchronism on evolutionary games. J. Theor. Biol.
**2011**, 269, 109–122. [Google Scholar] [CrossRef] [PubMed][Green Version] - Boccaletti, S.; Latora, V.; Moreno, Y.; Chavez, M.; Hwang, D.H. Complex networks: Structure and dynamics. Phys. Rev.
**2006**, 424, 175–308. [Google Scholar] [CrossRef] - Traulsen, A.; Semmann, D.; Sommerfeld, R.D.; Krambeck, H.J.; Milinski, M. Human strategy updating in evolutionary games. Proc. Natl. Acad. Sci. USA
**2010**, 107, 2962–2966. [Google Scholar] [CrossRef][Green Version] - Melamed, D.; Harrell, A.; Simpson, B. Cooperation, clustering, and assortative mixing in dynamic networks. Proc. Natl. Acad. Sci. USA
**2018**, 115, 951–956. [Google Scholar] [CrossRef][Green Version] - Perc, M.; Szolnoki, A. Coevolutionary games—A mini review. BioSystems
**2010**, 99, 109–125. [Google Scholar] [CrossRef][Green Version] - Fehl, K.; van der Post, D.J.; Semmann, D. Co-evolution of behaviour and social network structure promotes human cooperation. Ecol. Lett.
**2009**, 14, 546–551. [Google Scholar] [CrossRef]

**Figure 1.**Invasion and fixation of the tit-for-tat strategy against a population of unconditional defectors. Panels show the fraction of T-agents reached in ${10}^{4}$ game rounds starting from a single initial T on WS and BA networks (left and right panels) as a function of the PD return r (average degree $\langle k\rangle =4$ and 8 in top and bottoms panels). Solid lines show the average fraction over 100 random initializations (network generation and random placement of the initial T). Dashed lines show the average fraction of cooperative actions, saturating at fixation to 0.5 because of the alternate exploitation triggered by new T-agents with their T-neighbors. Dots show the outcomes of single simulations (T-fraction); transparency is used to show dots accumulation. Colors code the predictive horizon h, from 3 to 5, and the corresponding invasion threshold $\langle {\overline{r}}_{\mathrm{T}}^{h}\left({k}_{\mathrm{n}}^{max}\right)\rangle $, averaged over the 100 networks, is reported; the invasion threshold for degree-4 and -8 regular networks are indicated by the vertical dotted lines. The results were obtained for strategy update rate $\delta =0.05$.

**Figure 2.**Invasion and fixation of the tit-for-two-tats strategy against a population of unconditional defectors. Panels show the fraction of T-agents reached in ${10}^{4}$ game rounds starting from a single initial T in WS and BA networks (left and right panels) as a function of the PD game return r (average degree $\langle {k}_{i}\rangle =4$ and 8 in top and bottom panels). Solid lines show the average fraction over 100 random initializations (network generation and random placement of the initial T). Dots show the outcomes of single simulations; transparency is used to show dots accumulation. Colors code the predictive horizon h, from 3 to 5, and the corresponding invasion threshold $\langle {\overline{r}}_{{\mathrm{T}}_{2}}^{h}\rangle $, averaged over the 100 networks, is reported; the invasion threshold for degree-4 and -8 regular networks are indicated by the vertical dotted lines. The results were obtained for strategy update rate $\delta =0.05$.

**Figure 3.**Invasion and fixation of the new-tit-for-two-tats strategy against a population of unconditional defectors. Panels show the fraction of T-agents reached in ${10}^{4}$ game rounds starting from a single initial T in WS and BA networks (left and right panels) as a function of the PD game return r (average degree $\langle {k}_{i}\rangle =4$ and 8 in top and bottom panels). Solid lines show the average fraction over 100 random initializations (network generation and random placement of the initial T). Dots show the outcomes of single simulations; transparency is used to show dots accumulation. Colors code the predictive horizon h, from 2 to 5, and the corresponding invasion threshold $\langle {\overline{r}}_{{\mathrm{nT}}_{2}}^{h}\rangle $, averaged over the 100 networks, is reported; the invasion threshold for degree-4 and -8 regular networks are indicated by the vertical dotted lines. The results were obtained for strategy update rate $\delta =0.05$.

**Table 1.**

**Tit-for-tat vs. D.**Neighborhood classification update for agent i when playing the game round t. For each neighbor j, the class label includes the strategy, T or D, that j is expected to use at round t and the last actions, C or D, played by i and j, respectively. The strategy of agent i (second column) takes values in $\{\mathrm{T},\mathrm{Tn},\mathrm{D},\mathrm{Dn}\}$, where the ‘n’ stands for ‘new’ and indicates that the agent is new to the strategy. The first four classes in the first column are those that can occur after the first game round. The other classes are considered, top-to-bottom, in order of appearance in the last column.

At Round t | Class of $\mathit{j}$ after Round t | ||||
---|---|---|---|---|---|

Class of $\mathit{j}$ | Strat. of $\mathit{i}$ | Act. of $\mathit{i}$ | Act. of $\mathit{j}$ | Act. of $\mathit{j}$ | |

(Expected) | |||||

TCC | T | C | C | C | TCC |

${}^{1,2}$ | D | DCD | |||

Dn | D | C | TDC | ||

D | DDD | ||||

DCD | T,Dn | D | D | C | TDC |

${}^{1,2}$ | D | DDD | |||

TDC | T,Tn | C | D | D ${}^{3}$ | TCD |

${}^{1}$ | D,Dn | D | D ${}^{3}$ | TDD | |

DDD | T | D | D | C | TDC |

${}^{1}$ | D | DDD | |||

Tn | C | C | TCC | ||

D | DCD | ||||

D,Dn | D | C | TDC | ||

D | DDD | ||||

TCD | T,Dn | D | C | C | TDC |

${}^{2}$ | D | DDD | |||

TDD | Tn | C | D | C | TCC |

${}^{4}$ | D | TCD | |||

D | D | C | TDC | ||

D | TDD |

When Computing | Class of j Updated after Round t | Contribution per Neighbor per Round | Number of Neighbors | ||||
---|---|---|---|---|---|---|---|

1 | 2 | $\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}\mathit{\dots}\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}$ | $2\mathit{l}+1$ | $2\mathit{l}$ | |||

${\pi}_{\mathrm{TT},i}^{h}$ | TCC | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | … | $r-1$ | $r-1$ | ${k}_{i}^{\mathrm{TCC}}$ |

TCD | r | $-1$ | … | r | $-1$ | ${k}_{i}^{\mathrm{TCD}}$ | |

TDC | $-1$ | r | … | $-1$ | r | ${k}_{i}^{\mathrm{TDC}}$ | |

DCD, DDD | 0 | 0 | … | 0 | 0 | – | |

TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{TD},i}^{h}$ | TCC, TCD | r | 0 | … | 0 | 0 | ${k}_{i}^{\mathrm{TCC}}+{k}_{i}^{\mathrm{TCD}}$ |

TDC, DCD, DDD | 0 | 0 | … | 0 | 0 | – | |

TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{DD},i}^{h}$ | TDC, TDD, DDD | 0 | 0 | … | 0 | 0 | – |

TCC, TCD, DCD | Revising Ds have no such neighbors | 0 | |||||

${\pi}_{\mathrm{DT},i}^{h}$ | TDC, TDD | $-1$ | r | … | $-1$ | r | ${k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}}$ |

DDD | $-1$ | 0 | … | 0 | 0 | ${k}_{i}^{\mathrm{DDD}}$ | |

TCC, TCD, DCD | Revising Ds have no such neighbors | 0 | |||||

$\begin{array}{l}{\pi}_{\mathrm{TT},i}^{h}=\phantom{\rule{-0.56905pt}{0ex}}\left(\right)open="\{"\; close>\phantom{\rule{-4.83694pt}{0ex}}\begin{array}{ll}h{k}_{i}^{\mathrm{TCC}}(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)+\frac{{\scriptstyle h}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TCD}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}})r-\frac{{\scriptstyle h}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TCD}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}})& \mathrm{if}\phantom{\rule{0.277778em}{0ex}}h\phantom{\rule{0.277778em}{0ex}}\mathrm{is}\phantom{\rule{4.pt}{0ex}}\mathrm{even}\\ h{k}_{i}^{\mathrm{TCC}}(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)+\frac{{\scriptstyle h-1}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TDC}}r-{k}_{i}^{\mathrm{TCD}})+\frac{{\scriptstyle h+1}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TCD}}r-{k}_{i}^{\mathrm{TDC}})& \mathrm{if}\phantom{\rule{0.277778em}{0ex}}h\phantom{\rule{0.277778em}{0ex}}\mathrm{is}\phantom{\rule{4.pt}{0ex}}\mathrm{odd}\end{array}\\ {\pi}_{\mathrm{TD},i}^{h}=({k}_{i}^{\mathrm{TCC}}+{k}_{i}^{\mathrm{TCD}})r\end{array}{\pi}_{\mathrm{DT},i}^{h}=\phantom{\rule{-0.56905pt}{0ex}}\left(\right)open="\{"\; close>\phantom{\rule{-5.406pt}{0ex}}\begin{array}{ll}\phantom{\rule{5.69054pt}{0ex}}\frac{{\scriptstyle h}}{{\scriptstyle 2}}\phantom{\rule{4.2679pt}{0ex}}({k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}})r\phantom{\rule{0.28453pt}{0ex}}-\phantom{\rule{4.83694pt}{0ex}}\frac{{\scriptstyle h}}{{\scriptstyle 2}}\phantom{\rule{4.2679pt}{0ex}}({k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}})-{k}_{i}^{\mathrm{DDD}}& \phantom{\rule{29.30634pt}{0ex}}\mathrm{if}\phantom{\rule{0.277778em}{0ex}}h\phantom{\rule{0.277778em}{0ex}}\mathrm{is}\phantom{\rule{4.pt}{0ex}}\mathrm{even}\\ \frac{{\scriptstyle h-1}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}})r-\frac{{\scriptstyle h+1}}{{\scriptstyle 2}}({k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}})-{k}_{i}^{\mathrm{DDD}}& \phantom{\rule{29.30634pt}{0ex}}\mathrm{if}\phantom{\rule{0.277778em}{0ex}}h\phantom{\rule{0.277778em}{0ex}}\mathrm{is}\phantom{\rule{4.pt}{0ex}}\mathrm{odd}\end{array}\\ {\pi}_{\mathrm{DD},i}^{h}0$ |

**Table 3.**

**Tit-for-two-tats vs. D.**Neighborhood classification update for agent i when playing the game round t. For each neighbor j, the class label includes the strategy, T or D, that j is expected to use at round t and the last actions, C or D, played by i and j, respectively. The actions labels CD1/DC1 means that i/j has been exploited by j/i at round $t-1$ but the same did not happen at round $t-2$. The strategy of agent i (second column) takes values in $\{\mathrm{T},\mathrm{Tn},\mathrm{D},\mathrm{Dn}\}$, where the ‘n’ stands for ‘new’ and indicates that the agent is new to the strategy. The first four classes in the first column are those that can occur after the first game round. The other classes are considered, top-to-bottom, in order of appearance in the last column.

At Round t | Class of $\mathit{j}$ after Round t | ||||
---|---|---|---|---|---|

Class of $\mathit{j}$ | Strat. of $\mathit{i}$ | Act. of $\mathit{i}$ | Act. of $\mathit{j}$ | Act. of $\mathit{j}$ | |

(Expected) | |||||

TCC | T | C | C | C | TCC |

${}^{1,2}$ | D | DCD1 | |||

Dn | D | C | TDC1 | ||

D | DDD | ||||

DCD1 | T | C | D | C | TCC |

${}^{1,2}$ | D | DCD | |||

Dn | D | C | TDC1 | ||

D | DDD | ||||

TDC1 | T,Tn | C | C | C | TCC |

${}^{1}$ | D | DCD1 | |||

D,Dn | D | C | TDC | ||

D | DDD | ||||

DDD | T,D,Dn | D | D | C | TDC1 |

${}^{1}$ | D | DDD | |||

Tn | C | C | TCC | ||

D | DCD1 | ||||

DCD | T,Dn | D | D | C | TDC1 |

${}^{1,2}$ | D | DDD | |||

TDC | Tn | C | D | D ${}^{4}$ | TCD1 |

${}^{1,3}$ | D,Dn | D | D ${}^{4}$ | TDD | |

TCD1 | T | C | C | C | TCC |

${}^{2}$ | D | DCD | |||

Dn | D | C | TDC1 | ||

D | DDD | ||||

TDD | Tn | C | D | C | TCC |

${}^{5}$ | D | TCD1 | |||

D | D | C | TDC1 | ||

D | TDD |

When Computing | Class of j updated after Round t | Contribution per Neighbor per Round | Number of Neighbors | ||||
---|---|---|---|---|---|---|---|

1 | 2 | 3 | $\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}\mathit{\dots}\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}$ | h | |||

${\pi}_{\mathrm{TT},i}^{h}$ | TCC, TCD1, TDC1 | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | … | $r-1$ | ${k}_{i}^{\mathrm{TCC}}+{k}_{i}^{\mathrm{TDC}1}+{k}_{i}^{\mathrm{TCD}1}$ |

DCD1 | $-1$ | 0 | 0 | … | 0 | ${k}_{i}^{\mathrm{DCD}1}$ | |

DCD, DDD | 0 | 0 | 0 | … | 0 | – | |

TDC, TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{TD},i}^{h}$ | TCC, TCD1 | r | r | 0 | … | 0 | ${k}_{i}^{\mathrm{TCC}}+{k}_{i}^{\mathrm{TCD}1}$ |

TDC1 | r | 0 | 0 | … | 0 | ${k}_{i}^{\mathrm{TDC}1}$ | |

DCD, DCD1, DDD | 0 | 0 | 0 | … | 0 | – | |

TDC, TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{DD},i}^{h}$ | TDC1 | r | 0 | … | 0 | 0 | ${k}_{i}^{\mathrm{TDC}1}$ |

TDC, TDD, DDD | 0 | 0 | … | 0 | 0 | – | |

TCC, TCD1, DCD, DCD1 | Revising Ds have no such neighbors | 0 | |||||

${\pi}_{\mathrm{DT},i}^{h}$ | TDC1 | $r-1$ | $r-1$ | $r-1$ | … | $r-1$ | ${k}_{i}^{\mathrm{TDC}1}$ |

TDC, TDD | $-1$ | $r-1$ | $r-1$ | … | $r-1$ | ${k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}}$ | |

DDD | $-1$ | $-1$ | 0 | … | 0 | ${k}_{i}^{\mathrm{DDD}}$ | |

TCC, DCD1, DCD, TCD1 | Revising Ds have no such neighbors | 0 | |||||

$\begin{array}{ll}{\pi}_{\mathrm{TT},i}^{h}=h({k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1})(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)-{k}_{i}^{\mathrm{DCD}1}& {\pi}_{\mathrm{TD},i}^{h}=({k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1})2r\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1}r\\ {\pi}_{\mathrm{DT},i}^{h}=h{k}_{i}^{\mathrm{TDC}1}(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)\phantom{\rule{-1.70717pt}{0ex}}+\phantom{\rule{-1.70717pt}{0ex}}(h\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}})(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)-({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}})\phantom{\rule{-1.99168pt}{0ex}}-\phantom{\rule{-0.56905pt}{0ex}}2{k}_{i}^{\mathrm{DDD}}& {\pi}_{\mathrm{DD},i}^{h}={k}_{i}^{\mathrm{TDC}1}r\end{array}\phantom{\rule{0ex}{0ex}}\begin{array}{ll}\Delta {\pi}_{\mathrm{T},i}^{1}={\pi}_{\mathrm{TD},i}^{1}\phantom{\rule{-0.56905pt}{0ex}}-{\pi}_{\mathrm{TT},i}^{1}={k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{DCD}1}& \Delta {\pi}_{\mathrm{D},i}^{1}={\pi}_{\mathrm{DT},i}^{1}\phantom{\rule{-0.56905pt}{0ex}}-{\pi}_{\mathrm{DD},i}^{1}=-({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{DDD}})\end{array}$ |

**Table 5.**

**New-tit-for-two-tats vs. D.**Neighbors’ classification update for agent i when playing the game round t. For each neighbor j, the class label includes the strategy, T or D, that j is expected to use at round t and the last actions, C or D, played by i and j, respectively. The action labels CD1/DC1 mean that i/j changed to T after round $t-2$ and she has been exploited by j/i at round $t-1$. The strategy of agent i (second column) takes values in $\{\mathrm{T},\mathrm{Tn},\mathrm{D},\mathrm{Dn}\}$, where the ‘n’ stands for ‘new’ and indicates that the agent is new to the strategy. The first four classes in the first column are those that can occur after the first game round. The other classes are considered, top-to-bottom, in order of appearance in the last column.

At Round t | Class of $\mathit{j}$ after Round t | ||||
---|---|---|---|---|---|

Class of $\mathit{j}$ | Strat. of $\mathit{i}$ | Act. of $\mathit{i}$ | Act. of $\mathit{j}$ | Act. of $\mathit{j}$ | |

(Expected) | |||||

TCC | T | C | C | C | TCC |

${}^{1,2}$ | D | DCD | |||

Dn | D | C | TDC | ||

D | DDD | ||||

DCD | T,Dn | D | D | C | TDC1 |

D | DDD | ||||

TDC | Tn | C | D | D ${}^{4}$ | TCD1 |

${}^{1,3}$ | D,Dn | D | D ${}^{4}$ | TDD | |

DDD | T,D,Dn | D | D | C | TDC1 |

${}^{1}$ | D | DDD | |||

Tn | C | C | TCC | ||

D | DCD1 | ||||

TDC1 | T,Tn | C | C ${}^{5}$ | C | TCC |

${}^{1}$ | D | DCD1 | |||

Tn | D ${}^{6}$ | D ${}^{7}$ | TCD1 | ||

D,Dn | D | C ${}^{8}$ | C | TDC | |

D | DDD | ||||

D | D | D | TDD | ||

TCD1 | T | C | C | C | TCC |

${}^{2}$ | D | DCD | |||

Dn | D | C | TDC | ||

D | DDD | ||||

TDD | Tn | C | D | C | TCC |

${}^{2,9}$ | D | TCD1 | |||

D | D | C | TDC1 | ||

D | TDD | ||||

DCD1 ${}^{1,2}\phantom{\rule{-0.166667em}{0ex}}$ | T,Dn | D | D | C | TDC1 |

D | DDD |

When Computing | Class of j Updated after Round t | Contribution per Neighbor per Round | Number of Neighbors | ||||
---|---|---|---|---|---|---|---|

1 | 2 | 3 | $\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}\mathit{\dots}\phantom{\rule{0.277778em}{0ex}}\phantom{\rule{0.277778em}{0ex}}$ | h | |||

${\pi}_{\mathrm{TT},i}^{h}$ | TCC, TCD1, TDC1 | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | $\phantom{\rule{0.166667em}{0ex}}r-1\phantom{\rule{0.166667em}{0ex}}$ | … | $\mathbf{r}-\mathbf{1}$ | ${\mathbf{k}}_{\mathbf{i}}^{\mathrm{TCC}}+{\mathbf{k}}_{\mathbf{i}}^{\mathrm{TDC}1}+{\mathbf{k}}_{\mathbf{i}}^{\mathrm{TCD}1}$ |

DCD, DCD1, DDD | 0 | 0 | 0 | … | 0 | – | |

TDC, TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{TD},i}^{h}$ | TCC,TCD1,TDC1 | r | 0 | 0 | … | 0 | ${k}_{i}^{\mathrm{TCC}}+{k}_{i}^{\mathrm{TDC}1}+{k}_{i}^{\mathrm{TCD}1}$ |

DCD, DCD1, DDD | 0 | 0 | 0 | … | 0 | – | |

TDC, TDD | Revising Ts have no TDD-neigbors | 0 | |||||

${\pi}_{\mathrm{DD},i}^{h}$ | TDC1 | r | 0 | … | 0 | 0 | ${k}_{i}^{\mathrm{TDC}1}$ |

TDC, TDD, DDD | 0 | 0 | … | 0 | 0 | – | |

TCC, TCD1, DCD, DCD1 | Revising Ds have no such neighbors | 0 | |||||

${\pi}_{\mathrm{DT},i}^{h}$ | TDC1 | $r-1$ | $r-1$ | $r-1$ | … | $r-1$ | ${k}_{i}^{\mathrm{TDC}1}$ |

TDC, TDD | $-1$ | $r-1$ | $r-1$ | … | $r-1$ | ${k}_{i}^{\mathrm{TDC}}+{k}_{i}^{\mathrm{TDD}}$ | |

DDD | $-1$ | 0 | 0 | … | 0 | ${k}_{i}^{\mathrm{DDD}}$ | |

TCC, TCD1, DCD, DCD1 | Revising Ds have no such neighbors | 0 | |||||

$\begin{array}{ll}{\pi}_{\mathrm{TT},i}^{h}=h({k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1})r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)& {\pi}_{\mathrm{TD},i}^{h}=({k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1})r\\ {\pi}_{\mathrm{DT},i}^{h}=h{k}_{i}^{\mathrm{TDC}1}(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)+(h\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}})(r\phantom{\rule{-1.42262pt}{0ex}}-\phantom{\rule{-1.70717pt}{0ex}}1)-({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{DDD}})& {\pi}_{\mathrm{DD},i}^{h}={k}_{i}^{\mathrm{TDC}1}r\end{array}\phantom{\rule{0ex}{0ex}}\begin{array}{ll}\Delta {\pi}_{\mathrm{T},i}^{1}={\pi}_{\mathrm{TD},i}^{1}\phantom{\rule{-0.56905pt}{0ex}}-{\pi}_{\mathrm{TT},i}^{1}={k}_{i}^{\mathrm{TCC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TCD}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1}& \Delta {\pi}_{\mathrm{D},i}^{1}={\pi}_{\mathrm{DT},i}^{1}\phantom{\rule{-0.56905pt}{0ex}}-{\pi}_{\mathrm{DD},i}^{1}=-({k}_{i}^{\mathrm{TDC}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDC}1}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{TDD}}\phantom{\rule{-1.99168pt}{0ex}}+\phantom{\rule{-0.56905pt}{0ex}}{k}_{i}^{\mathrm{DDD}})\end{array}$ |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Della Rossa, F.; Dercole, F.; Di Meglio, A.
Direct Reciprocity and Model-Predictive Strategy Update Explain the Network Reciprocity Observed in Socioeconomic Networks. *Games* **2020**, *11*, 16.
https://doi.org/10.3390/g11010016

**AMA Style**

Della Rossa F, Dercole F, Di Meglio A.
Direct Reciprocity and Model-Predictive Strategy Update Explain the Network Reciprocity Observed in Socioeconomic Networks. *Games*. 2020; 11(1):16.
https://doi.org/10.3390/g11010016

**Chicago/Turabian Style**

Della Rossa, Fabio, Fabio Dercole, and Anna Di Meglio.
2020. "Direct Reciprocity and Model-Predictive Strategy Update Explain the Network Reciprocity Observed in Socioeconomic Networks" *Games* 11, no. 1: 16.
https://doi.org/10.3390/g11010016