Repost Rings: How Telegram Channels Launder Each Other's Reach (and a Graph Fix)

Recent investigations reveal that Telegram channels often engage in a deceptive practice to inflate their reach. A single channel might repost content from several other channels, making it appear as though it has a larger audience than it actually does. A study found that even a channel with a healthy view ratio of 15% can still be part of a network of 40 channels reposting the same six items daily, giving the impression that each has active viewership.

The report outlines a method to detect and expose these repost rings using publicly available data. The process involves collecting "repost edges," which are links in the form of t.me/other_channel/post_id. By parsing these links, one can create a directed graph where each edge represents a repost from one channel to another.

To avoid false positives, the study recommends filtering out one-time citations, keeping only the edges that recur frequently within a short time window. This helps to distinguish genuine information relationships from mere reposting. Once the graph is constructed, the researchers identify connected components, which represent clusters of channels that repost each other's content.

A component where every node reposts every other node within minutes, and the component has minimal connections to the outside world, is identified as a closed laundering ring. This indicates that the aggregate reach belongs to a single channel, not multiple ones.

The study also proposes a scoring system to rank Telegram channels based on their involvement in these rings. Channels found to be part of dense, outward-isolated components, where they repost content to many other channels without receiving new content in return, are penalized. Conversely, channels that act as sources of outward reposts, connecting many independent components, are considered valuable and receive a boost in their score.

The key takeaway is that reach is not solely determined by the number of channels reposting content but also by the uniqueness of the audience reached by original content. By understanding the graph of repost relationships, one can separate genuine channels from those artificially inflating their reach through repost rings. This method is publicly available and can be implemented on services like GitHub Actions without the need for paid APIs.