Simulation · deterministic model

PageRank simulator and calculator

Type out a set of pages and the links between them, choose a damping factor, and watch authority flow through the graph until the ranking converges.

Workspace/pagerank
RUNS LOCALLY

PageRank simulator

Say it the way you would say it to a person. Your words are matched to one of 18 tools by readable rules, and the tool runs right here in your browser. No upload, no account.

/ tools⌘K search↑ last message

PageRank simulatorRuns on this device · nothing uploaded

This tool runs in your browser and needs JavaScript turned on.

How to use the PageRank simulator

  1. Describe your pages and links in the settings form, one page per line: Home -> Blog, Pricing. An example graph is filled in to start from.
  2. Run it, or type “run pagerank with damping 0.7”.
  3. Watch the scores converge, iteration by iteration, until they settle.
  4. Read the final ranking to see which pages your link structure favours and which it starves.

Things you can type: run pagerank run pagerank with damping 0.5

The algorithm that is still worth understanding

PageRank is no longer the whole of how search works, and it is still the clearest model of an idea that has not gone anywhere: a link is a vote, votes from important pages count more, and importance is defined recursively by the same rule.

For anyone planning an internal linking structure, that recursion is the practical part. It explains why a link from your busiest page is worth more than five from orphans, and why funnelling every link into one hub can starve the pages you actually want to rank.

What you can do here

  • Describe a link graph in one line per page: Home -> Blog, Pricing.
  • Set the damping factor and watch how a lower value flattens the distribution.
  • Run to convergence and see the delta shrink each iteration.
  • Read the final ranking, normalised so the scores sum to one.

Dangling nodes are handled

A page with no outbound links is not an error — it is most of the web, and mishandling it is the classic way a naive implementation leaks rank until the scores stop summing to one. Here their authority is redistributed properly, and the test suite asserts the normalisation holds.

Try it on your own site

Sketch your top twenty pages and the links between them. The ranking you get will not match anything Google reports, and it will show you which pages your own structure is starving — which is the actionable half.

Questions

PageRank simulator: common questions

What damping factor should I use?

0.85 is the value from the original paper and a sensible default. Lower it and rank distributes more evenly; raise it and the graph structure dominates more sharply.

Is this how Google ranks pages today?

No. Modern ranking involves hundreds of signals and machine-learned models. PageRank remains a good model of how link authority propagates, which is why it is still worth being able to reason about.

Can I import my site’s real link graph?

Not directly. Export your internal links from a crawler and paste them as source -> target lines; the extractor on this site will pull every link out of a page’s HTML if you need a starting point.