In assessing our understanding of past ozone changes, and attributing them to particular causes, it is important to determine how much of the record is a forced "signal" and how much is climate "noise". At the same time, the "noise" may itself be changing because of chemical forcing. The basic question is, to what extent should we expect to be able to explain the past ozone record? To address these issues almost certainly requires the use of a climate model; moreover the ability to explain the past underpins predictions of future ozone. Here the Canadian Middle Atmosphere Model (CMAM), a fully coupled chemistry climate model, is used to address the issue of past and future ozone in the northern hemisphere. CMAM, like many other models, has a difficult time accounting (in the mean) for the relatively cold Arctic winters and associated low ozone in the 1990s. While this could be a model deficiency, the possibility is also raised that the recent behaviour reflects anomalously low ozone at the lower end of a range of natural variability enhanced by chlorine loading. This has implications for future ozone. The expected future change in stratospheric circulation (and ozone) arising from GHG increases is also critically examined.