Online game proves 80-year-old psychological theory

Updated Emily Banks 0 492 Views

The study of interactions in online game recently proved an 80-year-old psychological theory, confirming the idea that "a friend of my enemy, is my enemy." The theory is known as Structural Balance theory, which means "individuals tend to avoid stress-causing relationships when they develop a society resulting in more stable social networks," according to an Imperial College London .

The study analyzed the relationships between 300,000 players in the online game Pardus , an open-ended virtual universe game. In this simulated society, players make friends and enemies, fight and trade.

"I find it fascinating to understand how we all interact with one another to form complex social networks. I think it is astounding that I'm this tiny point in such an enormous network of people. Our new study reveals in more detail than ever before the key ingredients that make these networks stable," said Renaud Lamblotte, one authors of the study from the Institute for Mathematical Sciences at Imperial College London.

The researchers, who were from Imperial College, the Medical University of Vienna and the Santa Fe Institute, analyzed data from the individual networks and from the larger network perspective. They found that positive relationships were more likely to manifest themselves and be reciprocated than negative sentiments.

"For example, if player A declares player B to be their friend, player B is likely to do the same. If player A declares player B to be their enemy, however, player B is not likely to reciprocate," according to the study.

How does that translate to a virtual game? Some of the connections between players are positive, like friendship, communication and trade, while others are more negative and might include hostility, aggression and punishment.

Players who were "friends" were more likely to have overlapping networks and communicate with one another, while hostility and trade didn't overlap -- meaning enemies tended to avoid each other and choose not to trade with one another.

"This may seem like an obvious finding, as we would all prefer to communicate more with people we like. However, nobody has shown the evidence for this theory on such a large scale before."

The study's researchers are working to apply their mathematical tools to larger networks so that they can study the communications between million sof people using cell phone data, according to the press release.

The paper regarding the study can be found in the current issue of Proceedings of the National Academy of Sciences .

Picture_8.png 102.89 KB

Dani AI

Generated

As summarized, this thread points to an important empirical test of structural‑balance ideas using game logs. Structural balance goes back to Fritz Heider’s 1946 formulation of cognitive balance and was given a graph‑theory treatment by Cartwright and Harary in 1956 — those two works are the foundation for the “balanced vs unbalanced triad” idea. (psychclassics.yorku.ca)

Concrete test (short): represent friendly ties as +1 and hostile ties as −1, then a triangle (three people) is “balanced” when the product of its three edge signs is +1 — equivalently the triangle has 0 or 2 negative edges. That algebraic rule is what lets researchers turn psychology into a repeatable network test rather than a loose verbal claim. (philpapers.org)

Practical checklist for anyone trying the same analysis on game or social‑media logs:

  • decide how events map to signed ties (time window, direction, weight).
  • for balance tests use an undirected signed graph; directed data often follow different mechanisms (Leskovec et al. show classical balance can fail on directed social media and propose status‑based models instead).
  • count triangles efficiently (use library functions or triangle/triadic‑census algorithms), then classify each triangle by multiplying the three sign values. Example (Python + NetworkX): 
import networkx as nx
from itertools import combinations

# G = nx.Graph(); add edges with G.add_edge(u,v,sign=1 or -1)
def count_balanced_triads(G):
    balanced = total = 0
    for a,b,c in combinations(G.nodes(),3):
        if G.has_edge(a,b) and G.has_edge(b,c) and G.has_edge(c,a):
            total += 1
            prod = (G[a][b]['sign'] * G[b][c]['sign'] * G[c][a]['sign'])
            if prod == 1:
                balanced += 1
    return balanced, total

NetworkX and related libraries provide triad/triangle functions and subquadratic algorithms for large sparse graphs. Use those rather than naive O(n^3) loops for big data. (networkx.org)

Caveats: pick sign thresholds and time windows thoughtfully; test results against null models (randomized signs or edge permutations); and remember game mechanics shape behavior — the “control/power/manipulation” appeal noted is real, so interpret findings as behavior in a designed environment that nevertheless offers useful insight into broader social dynamics.

Member Avatar for Member #949455
Member #949455

The study analyzed the relationships between 300,000 players in the online game Pardus , an open-ended virtual universe game. In this simulated society, players make friends and enemies, fight and trade.

Nice article.

This game remind me of a game called Romance of the Three Kingdoms.

I can see why it is addicted having control, power and being manipulate.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.

Edit Preview