Ranking on two normalized factors: the double-low pattern

A ranking pattern worth stealing from fixed-income screening: rank every candidate by the sum of two normalized measures, buy the lowest scores, rebalance periodically. The instrument it came from does not matter. The method is cleaner than what most of us do. The pattern generalizes to: pick two measures you believe matter and that aren't the same thing, normalize each across the candidate set, add them, rank. Then trade the top or bottom slice on a fixed schedule. Why it beats hand-picking. It forces you to state exactly what you're selecting on. It makes the selection reproducible, so you can backtest the ranking itself rather than your judgment. And it stops you quietly overriding the list on the days you don't like what it says. The discipline that comes with it is the fixed rebalance schedule. Rank, take the slice, hold until the next rebalance. Not "rank, then think about it." The hard part is picking two measures that aren't secretly the same measure. That's where the work is, not in the arithmetic.