Great example in Portugal - election results predicted by the pace of the Reconquista: https://threadreaderapp.com/thread/1181292177130999808.html

Your counterargument proves too much; if "people" don't have agency then you can't blame anyone for anything, including being unable to blame the Mongols for anything about modern China. The concept of blame becomes entirely pointless.

This may, in fact, be true. But in this specific context, it's an argument that proves too much.

Psychohistory is pure fiction. The closest real equivalent is economics and economic predictions are essentially always wrong. I don't remember where exactly, but I recall reading that someone had done a mass study on economic forecasts and discovered that the only time they're correct was when they were predicting a continuation of whatever the current trend line was. Forecasting never managed to predict an actual change to the status quo, making such forecasts useless.

We can't even measure GDP accurately in quarters already passed, let alone predict what it will be in future, and surely predicting the movement of GDP is the most basic task a psychohistorian could possibly set themselves.

Our ability to make predictions based on psychohistory is pure fiction — but that doesn't mean that the rest of the comment in false. I think it'd be a real challenge to accurately predict the weather 168 hours from now, but that doesn't mean that the weather a week from isn't theoretically knowable or that it's not based on a series of knowable actions and reactions between now and then.

If you assume people are just molecular reactions in their brain that could, theoretically, be modelled like any physical system then sure. But otherwise there's no useful way to predict the response of whole societies to arbitrary events beyond trivialities anyone could predict.