like an investor thinks that an asset is undervalued, maybe through a differentiated view, or just pure discovery (harder), that is the view I am trying to approach more my system design.

what could very well be a good trade in the future given today's participant behavior? is not just about did it work on a backtest in the past , who cares for that anyways, but whats the economical reasoning that could continue to work, in the future, in the market that I trade.
all my best trades so far, are trades I didnt think about it like this at the time, I will be honest. but it turns out, they played out quite well , just because that thing, behavior, I was not thinking about at that time. I can't give examples right now as I am still trading them.
for example for AI right now. there's a good trade I do think that will pay off if certain things happen in the future, and if it doesnt, it will be mostly covered imo. is it real ? I am not sure yet, but I am thinking that it might as it makes sense. it makes sense because its something I trade elsewhere and the economical reasoning is there. it will take some work to build out, but I'll make it work.
markets are forward looking machines. anything in the past is risk premium. which is fine, its what I do. but the best opportunities lies where no one is looking, or there's very little appetite for participation as they are riskier and historical precedence is not there.
at least thats how I think about it.