Game theory

Recently, much attention has been paid to the idea of ​​viral content distribution, but according to a recent study published in PNAS , the leading US research journal, this is not the only way that ideas, innovations and technologies can spread.

Two researchers conducted interesting theoretical work, tested with the help of focus groups and concluded that the changes are spreading quickly, not only because they are open to a large number of people. On the contrary, such a distribution often passes by the game rules; and the players decide whether to accept something new, just on the grounds that everyone around has already done so.

The growing popularity of web pages or gadgets is often described in terms of the “epidemic”: a social network, with a huge number of connections between people, increasing the impact, and the adoption of something new. It is the intersection of interests (links) between different, in its component, groups. At that moment, when a trend reaches a participant (hereinafter: a node) with a huge number of connections (a well-known personality) - its popularity explodes.
')
A similar development of events, indeed, takes place in some cases; in others, this theory is too simplified - a simple impact (“look at my new iPad!”) on an individual does not guarantee that the trend will be accepted and transferred further along the chain.

Here is what Amin Saberi, one of the authors of the research, says about this: “It’s not only the internal qualities and values ​​of a new technology (or another type of innovation) that make it attractive to people. A huge role is played by the environment that has already taken it. In situations where there is a benefit to make the same decisions as the people around, the spread of such innovations follows the rules of the game theory, which is characterized by much wider waves, compared to viral trends or epidemics. ”

In order to show how this happens, the researchers created a theoretical scenario in which several people or nodes connected to the network took part as friends in social groups. All of them participated in the game, where in each round, each node had to decide whether to accept a new trend only on the basis of data on the current behavior of the “neighbor”.

For example, a node looks around to understand how many friends are participating in a trend, say, Farmville. If there is no such friend, the likelihood that the node will start playing in Farmville is low, if everything is, the probability is extremely high. At the same time, the game is made in such a way that the imitation of the behavior of neighbors contributed to obtaining higher marks, in contrast to their ignoring.

Based only on these rules, the social enclave, where each node receives complete and absolutely accurate information about what everyone else is doing, will never accept any trend. If people could only make decisions on such a basis, then none of the nodes would choose “change” as the best development strategy.

In order to correct this misunderstanding, the researchers added “noise” to this information field, as a result of which many nodes began to receive incomplete or erroneous data. Decisions were weighed so that the node with 0 information about its neighbors would prefer to accept the trend, whatever it was. Compare this with the reality in which a person does not give a damn what others think or do, and he weighs the value of innovation based on some internal factors.

In the course of the game with such a structure and according to such rules, the researchers discovered a pattern: nodes with local connections, as opposed to “far-reaching” nodes (which spread epidemics), spread innovation faster at times.

Nodes “not implanted” into the network structure and possessing a small number of connections (in life they are casual, trained users) transmit information several times faster than nodes with a huge number of connections, which, on the contrary, slow down the chains.

Super-loaded nodes serve as checkpoints, because without having crystal clear information about the opinions or actions of their neighbors, they are still subjected to greater pressure from their connections ( information redundancy ). Strange as it may seem, such a node should not give a damn about its own neighbors to accept a trend, or it should be surrounded by other nodes that have already accepted, without exception, this trend before it. This is the main difference between the game spread and the epidemic.

This model does not work so well in the case of individual content, where a simple “share” is often enough to cause rapid growth. On the other hand, the model perfectly explains over-loyalty to sites that distribute content, such as Digg, Twitter, Reddit, and loyalty to social “genres” and “categories”.

The authors also say that according to the game theory, decisions are also distributed that affect the building of further communication: the choice between democrats and liberals; as well as the adoption of a technological vector: the choice between Apple and Microsoft.

Dr. Sabery gives the following example: “The reason I use Facebook in return for any other social network is not only in its quality, but also in the fact that I have many friends who are already using it”. The same thing happens, as we have already noted, with operating systems, choosing a computer, a place to rest, and so on. While each telecom operator is trying to give us a maximum of reasons why we should connect to it, buns in the form of free calls or SMS can influence the decisions of entire groups of nodes to migrate from one to the other. Carrying the rest.

In this game theory, networks tend to balance the adoption of new changes - this is the only thing that distinguishes theory from practice (in practice, most individuals and groups tend to extremes). But most importantly, this model shows: trends and innovations can spread quickly, based on factors that exclude massive impact (something like “popularity / loyalty without obvious popularity / loyalty”). The rapid and powerful spread of innovation can occur on other models that contain more complex and subtle mechanisms of social impact, such as: "Friends advised me here ..."

PNAS via ArsTechnica