Monitoring, assessment and reporting of stream restoration projects have historically lagged far behind implementation. However, in recent years, rigorous post-project assessments (PPAs) of modern stream restoration practices have steadily increased. This has helped to stimulate debate and inquiry regarding the effectiveness of restoration techniques and has provided critical feedback to practitioners and planners useful in restoration design and implementation. Nonetheless, few studies exist that track the performance of modern restoration projects over a protracted period. Instead, most are based on a brief snapshot taken during the initial post-construction period, which may not always accurately characterize longer-term project performance. Here, we re-visit a stream restoration project implemented in 2005 on a third-order stream in central New York. By repeating several of our quantitative and qualitative evaluation procedures from the original 2007 PPA we demonstrate that (i) despite several recent large flood events and the fact that the current channel geometry differs from the design/as-built configuration, the project has made substantial progress towards the goals of channel stabilization and habitat enhancement; (ii) this more favourable, mid-term outcome was not necessarily evident during or well-predicted by our 2007 PPA; (iii) although continued deformation of in-stream structures may be a harbinger of future channel instability, riparian vegetation is playing an increasingly important role in maintaining channel stability; and (iv) accurately predicting local scour depths proximal to in-stream structures, performing a detailed sediment budget analysis, and prescribing adequate bank protection are critical to project success, especially during early stages of a project. Copyright © 2013 John Wiley & Sons, Ltd.