The baseline your trading strategy has to beat

Before trusting any entry, run this. It's a deliberately uninformative entry with a sensible exit.

1. The rule

# BUY — enters on almost nothing, on purpose if HhmmSs >= 100000 and HhmmSs <= 140000 and DayAmount > 5000000: if RisingStreak(5): Buy()# SELL if ProfitPct >= 2.0 or ProfitPct <= -1.0 or HoldTime >= 1200: Sell()Whatever this produces on your universe is the bar. If your clever entry doesn't clearly beat it after costs, your edge is in the exit rules, not the entry. Which is useful to know and much cheaper to find out now.

2. Why a deliberately weak baseline is useful

This entry is close to random on purpose. It waits for a mild rising streak inside the quiet part of the day on a name that trades enough to exit. It should not have an edge. That is what makes it useful. When your real strategy produces a result, the first question is whether the result comes from your idea or from the market simply moving. Running this baseline over the same period answers it. If your strategy and this one produce similar numbers, the market did the work.

3. When this baseline looks good

In a strongly trending sample, this will look profitable — and so will almost anything else that buys. That is the signal to widen your test period rather than to trade it.

4. What to change first

Nothing. Changing the baseline defeats its purpose. Keep it exactly as written and run it alongside every strategy you test.

5. What the round trip asks of it

A baseline has to include the toll, or it is not a baseline. The 2.0% / 1.0% structure needs 33% ignoring cost and 58% with the measured 0.75% round trip. Any strategy you write later is competing against that second number, not the first. This is the most common way a promising idea dies: it beats the naive baseline and loses to the honest one. Before you compare anything, price both at the same cost — we measured what one tick of slippage does to a real model, and it removed 44% of the edge (slippage vs backtest edge).