DPK and Colb both bring up good points.
Stoke implies the destruction (usually necrosis initially, and apoptosis in the penumbra later) of neural tissue, whereas cerebrovascular accident is more broad and refers to any compromise to the blood supply to the brain. CVA, therefore, could include a temporary condition, such as a transient ischemia attack, whereas a TIA is not necessary a stroke unless tissue infarct occurs. (Note: TIA typically precedes a more profound CVA within 2 weeks, so it is also NOT benign.)
Anyway, there are two broard classifications of stoke, and these are hemorrhagic and ischemic (embolitic, thrombotic, and hypoperfusion). Your question is about hemorrhagic strokes: subarachnoid and intracerebral. It is probably important to note that thrombotic ischemia secondary to athersclerosis is the most common type of cerebrovascular accident, second to embolism, which usually occurs secondary to a cardiac disorder.
Subarachnoid strokes a characterized by rupture of berry aneurysms that form at the bifurcation of Circle of Willis vessels non-traumatically or through trauma to bridging veins that span the pia to arachnoid mater space, the subarachnoid space. The classical presentation of SAH is "the worst headache I have ever had", and these patients MUST be admitted and treated with neurosurgical occlusion of the aneurysm, because the risk of repeat hemorrhage is about 40% within 6 months, and obtundation and meningeal irritation present significant risk to the patient acutely. The upside, though, is that the bleeding is directly into the subarachnoid space, which is occupied by CSF, and this can ultimately be reabsorbed into the arachnoid vili of the lateral and superior surfaces of the cerebral hemispheres. But the larger vessels that are compromised in SAH gives it greater acute-phase mortality as compared to the intracerebral hemorrhage. So if the patient is managed acutely, then they are better off than having had an ICH (for reasons explained below).
ICH (not intracranial) is due to rupture of the small vessels embedded in the cerebral parenchyma, or lacunar strokes. These are bad, because they typically occur near vital neural tissue of the basal ganglia, and thus give rise to motor deficit. Also, depositing blood into the parenchyma poses a problem for the limited microglia around able to debride this unrequested cellular material. Therefore, then risk of tissue damage in the sub-acute period following ICH is greater than with the properly-managed SAH patient. Also, remember that this is the brain, so you could much more easily access clots or other material for removal on the surface of the pia mater then from within the brain parenchyma. However, with either of these serious conditions, then ultimate cause of death is increased intracranial pressure, transtentorial uncal and foramen magnum herniation, and compression of brainstem respiratory centers (or, more rarely, the compromise of vital brainstem function in ICH), so you have to also consider that ICH will produce a large space-occupying lesion much more slowly than SAH.
Personally, I would rather have a well-managed SAH than an ICH, because there is less chance of post-morbid mental deficit, but SAH possess a greater risk of sudden death acutely.
