Recommendation systems often create echo chambers by repeatedly exposing users to a narrow set of content This phenomenon, referred to as a content trap, undermines content diversity and user agency. In this paper, we introduce a novel framework to measure the trap intensity of structural groups in recommendation networks by analyzing how easily users are drawn to and retained within these groups. We model user navigation using Uniform Random Walk (URW) and Degree-Biased Random Walk (DBRW) simulations across one-, two-, and three-hop neighborhoods. To evaluate both attraction and retention capacity, we compute first-hit metrics and visit frequencies respectively, then adjust for structural size to derive a normalized trap intensity score. Our method identifies structural regions with disproportionate influence on user pathways, offering insights into how recommendation algorithms can lead to semantic confinement. We validate our approach on YouTube’s recommendation network, highlighting critical zones that act as both attractors and retainers of user attention.