General player behaviour

We have analyzed the structure of Wikipedia’s hyperlinks graph. Let’s now see if it reflects the players behaviour.

The following plot shows an histogram of the number of clicks per article. Notice that the y-axis is logarithmic.

If players play smartly, they should click on articles that are hubs. We can see if there is any correlation between the number of clicks and the number of incoming links. We expect to see the top HITS articles being clicked more often. Both because they are more useful to get quickly to the goal and because they appear often.

Let us have a look at the most clicked articles:

This matches most of our top 10 HITS Authortities as hypothetised.

However, wikispeedia games have a low chance of beginning or finishing at a top authority. Players must thus “home out” and go from their initial page to hubs and then “home in” to get to their target.

To confirm this, we can look at the size of the hubs (in term of number of incoming links for example) players click througout the game.

This plot shows the average number of incoming and outgoing links across all articles at each relative position along paths. The relative position of a link is the position of the link divided by the total length of the path.

Players, in average, follow a similar strategy. And their clicking is clearly not random.

If would want to replicate a player behaviour, random picks would not be sufficient. We will come to that later…

We only did look at the articles players clicked on in term of “hubness” but do player also click on certain articles based on their position?

Are players lazy and only select the first links they choose? Let’s find out!