Now I'm even more confused about scoring systems.

I have two thoughts, though.

1) I've read material whereby patients over 75 have their anticoagulants stopped because of their supposedly significantly elevated fall risk (which would lead to cerebral bleeding). (Some interventional EPs will OK a watchman on account of age alone.)

2) I am mystified about why no scoring system seems to
incorporate the risk of hemorrhagic stroke (one EP surgeon told me about 15% strokes are hemorrhagic. )

That being so, I'd think the guidelines for blood thinners would be more multifactorial and complex. Also that patient recommendations would be more individualized.

Apparently (read on cardiology website) AI is increased used to make diagnoses and predictions - for instance re severity of an aortic valve leakage. I can easily imagine AI doing a better job than mere humans providing recommendations about whether or not to take blood thinners .

Calling this discussion to the attention of @larrryg333 whose experience-based decision about taking anticoagulants was very interesting to me,

One thing is clear - as things stand, no model or clinician can completely eliminate stroke risk (especially since the two kinds of stroke carry opposite recommendations re prophylaxis and treatment.)

Larry is absolutely right about one aspect to recovery from either type stroke - how intense and immediate rehab is makes an enormous difference in how well and indeed if, a patient overcomes deficits. We have so little control of that, though. Care of the aged is already so terribly sub-par!