I’d like to spark a discussion on median outcomes, edge harvesting. The following scenario applies to sports betting player props & very real world trading challenges that we all encounter:

Let’s say a trader has gone through the grunt work of backtesting and finds an edge where the median returns for companies w/ bad fundies 5 days after a press release spike are strongly negative (maybe -20%).

Assuming the trader has a strong set of brokerages and access to locates, how should she go about trading each *indiviudal* signal that pops up on his radar? We know the median outcome is strongly negative; however, how much weight should be given to the idiosyncrasies?

Since these are generally small caps (excluding TSLA),how much does the uniqueness of the next trading signal matter? Or does it matter at all? Or does sound model-based trading mandate that we take all signals and assume that the law of large numbers will provide a happy ending?