Speedruns | WikiMapping

Choose a start and target article to begin.

Under the hood

How bidirectional BFS finds the fastest path

Start
Follow children →MeetTarget
← Follow parents

Breadth-first search (BFS) explores articles in layers: first one link away, then two, then three. Every link has the same cost, so a shortest path means the fewest links, not the least reading time.

Bidirectional BFS keeps two searches. One follows outgoing links from the start; the other works backward through incoming links from the target. At each step, we expand a layer from the smaller frontier. When the searches meet, their saved steps join into a shortest directed path.

This often explores fewer articles than searching only from the start. Already visited articles are skipped, so cycles do not trap the search. If a frontier runs out before the searches meet, no route exists in the stored graph. Searches here have no fixed depth cutoff.

For a roughly balanced graph with branching factor b and path length d, the intuition is two searches of about b^d/2 instead of one of about b^d. Real Wikipedia links are uneven; worst-case work is still proportional to the reachable articles and links.

Find the shortest way there.

Fastest path

You made it!

Your route