The problem with historians is that they know too much.
In particular, they know how things turned out. When you have the benefit of hindsight, it’s hard to avoid simply assembling a narrative leading up to that result.
So I wish somebody would try doing blinded historical research.
Here’s how it would work. You want a historian to analyze the events and the forces at play in some historical context without knowing what comes next.
So you give them access to archives, newspapers, records – but only up to a certain date. They write their analysis of what is going on, and describe what they expect to happen next. Then you reveal what did happen, and compare it to the predictions.
The result? The historian can apply whatever methodology they like – and see how good it is at making predictions. Then you could apply the same methods in situations where you don’t know what will happen (e.g. the present), and have a ballpark idea of how much trust to put in it.