Animal models of cerebral ischaemia mimic at best less than 25% of all strokes. Compounds which prove efficacious in animal models should, therefore, only be expected to improve outcome in a quarter of all strokes. If trials for acute stroke are to succeed, stroke subgroups represented by the animal models should be targeted. For the other subgroups, e.g. lacunar stroke, appropriate animal models need to be developed. Moreover, thrombolysis should be included in animal models because it is likely to be used as a first line treatment for ischaemic stroke and any future therapeutics will need to be compatible with it.