This paper examines practical and theoretical problems and issues that arose during evaluation of an adventure program for young offenders. During 1989-93, the Sail Training Association and the Humberside (England) Probation Service collaborated on a project in which probation clients aged 17-25 made sailing voyages across the North Sea or around the coast of Britain. Although the underlying "treatment theory" was never fully articulated, it revolved around the notion of the personal developmental benefits to be derived by young adult offenders from a challenging physical experience. Evaluation included five elements: (1) reconviction rates after voyage completion; (2) changes in offender self-perceptions based on an adapted problem checklist administered before and after voyage; (3) ship captains' reports; (4) postvoyage client self-evaluations; and (5) probation officers' survey. Reconviction rates of voyagers were similar to those of matched probation controls. Administration of the checklist encountered many practical difficulties and was abandoned. Nevertheless, ship captains' reports, client self-evaluations, and the probation officer survey were very positive and supportive of the program. Questions to consider when conducting such evaluations include distinguishing between individual-level and system-level objectives, linking evaluation closely to declared project aims, and the need for parallel evaluations of outcomes and processes. (SV)