--- - |2+ Abstract Objective Psychotherapy patients who experience large, stable symptomatic improvement between sessions are more likely than those without such sudden gains to benefit from treatment. However, there is limited empirical basis for the definition of sudden gains, and it is unclear how they may affect symptomatic change at other points in treatment. Method In a psychotherapy training clinic, 149 adults completed a distress measure at each session. For each criterion in the definition of sudden gains, we evaluated the prediction of outcome, possible moderating variables, local score instability, and multi‐session change before and after score shifts. Results Large intersession improvement did not lead to increased rate of change later in treatment, but it did predict outcome, regardless of whether nearby session scores were stable. Early improvement during the first five treatment sessions was an independent predictor of outcome. Large intersession improvement generally occurred in the context of local score instability. Conclusion Sudden gains appear to predict outcome because of improvement inherent in the gains themselves. Early overall improvement predicts outcome almost as effectively as do sudden gains. There may be advantages to redefining sudden gains as large intersession improvement, regardless of local score stability. - Clinical Psychology & Psychotherapy, Volume 0, Issue ja, -Not available-.