Quantitative Research

A diversified portfolio, backtested

The trade journal stakes a fixed $200 on every signal, one at a time — most of a $100,000 bankroll sits in cash. This section tests the actual fix: instead of a token stake, the bankroll is split evenly across every buy signal open on a given day (about two, on average) and redeployed the same way the next day. No leverage, no compounding of the running balance — each day risks a fixed share of the original bankroll, exactly like the fixed-stake journal, just spread across more names at once.

Diversified portfolio SPY buy-and-hold QQQ buy-and-hold

Diversification, not a bigger edge, is what was missing. Nothing about the underlying trade changed — same signals, same window, same frictions. Fully deploying the bankroll across every simultaneous signal, instead of parking most of it in cash behind a token stake, is what turns a statistically real but economically thin per-trade edge into a return that competes with a buy-and-hold benchmark over the same period. This does not erase the caveats above — real transaction costs, capacity limits in illiquid names, and no sector or size controls still apply at this scale — but it shows the earlier "too small to matter" conclusion was about position sizing, not about the signal itself.