What a Plain English Trading Strategy Spec Should Contain

Most backtests that disappoint did not fail because of the market. They failed because the strategy was never written down precisely. A good spec removes guesswork before any code runs.

1. Universe

Which assets can the strategy trade? A list of tickers, an index, or a rule for choosing them. If the universe changes over time, say how.

2. Signals and ranking

What triggers an entry, and if several assets qualify at once, how are they ranked? "RSI below 30" is a signal. "Pick the five lowest RSI readings" is a ranking.

3. Rebalancing and allocation

How often does the strategy check for changes, and how much capital goes into each position? Equal weight, fixed dollar amounts and volatility based sizing all give different results.

4. Risk controls

Stop losses, maximum position size, and limits on total exposure. Write these down before you see results, not after.

5. Execution

What order types does the strategy use, and at what price does it assume it gets filled? Assuming a fill at the close you just observed is a common source of inflated results.

6. Assumed costs

Commissions, spreads and slippage. A strategy with no stated costs is not ready to test.

In Nvestiq, each strategy gets a plain English spec covering these parts before it is built, and the same compiled logic is then used in the backtest. The product is in early access. Nvestiq publishes this page, so weigh it accordingly.