What systems are set up to cross validate ideas in a healthy out of sample? Do we have correlation and covariance matrices across different strategy sleeves and are there guardrails in place that alert us when this is becoming dangerous? Do we have multiple data sources, and are there designated checks and balances to clean, reconcile, and validate that data? Is there designated time allocated for idea generation and performance feedback, backed by direct market observation? Is the lead of the desk intellectually honest about the returns each sleeve is producing? Are there checks and balances in place to ensure actionable processes exist in the event of underperformance? Is there a daily file that flags cautionary market outputs when abnormalities are detected, even if they do not immediately lead to underperformance? What mechanisms are in place to track trade crowding, and can we model market impact and alpha degradation as the strategy scales? And this is the tip of the iceberg. Infrastructure and organizational design takes a lot of time up front, but they pay dividends in the long run.

Oh don’t even get me started on check points and unit testing. A.I can handle all junior level coding. There should be no reason to be sloppy with your code. You can literally have 🤖 write unit tests within the scripts as yet another safety check. I’m the annoying PM that’s skeptical about everything and I beg for unit tests everywhere lol