Manual trading

The rules panel was written so you could trade it by hand. Here is how.

Every strategy’s Rules tab renders the spec as plain English — universe, entry, exits, sizing — precise enough to act on with nothing but a brokerage account and a routine. A daily golden cross reads like this:

Universe  Trades 5 symbols: SPY, QQQ, AAPL, MSFT, NVDA.
Entry     Go long when (the 50-day average is above the 200-day
          average) AND ((the 50-day average from 1 bar ago) is at
          or below (the 200-day average from 1 bar ago)).
          Fills happen at the open price of the signal bar.
Exits     Exit when the 50-day average is below the 200-day average.
Sizing    Each position uses 20% of equity, up to 5 positions at once.

Nothing here tells you to place a trade. These are rules you built and tested; whether and how to act on them is entirely your call, in your account, at your broker. What this page adds is the discipline layer — because by hand, the strategy is only half the system. You are the other half.

The end-of-day routine

The engine and the forward ledger evaluate rules on closed candles only. Manual trading should copy that discipline exactly: decisions happen once, after the close, on finished bars — never mid-session on a bar that is still moving. A rule that looks true at 2pm can be false by 4pm; acting early means trading a different strategy than the one you tested.

  1. After the close, check exits first

    For each open position, work down the exit lines in order: has the exit rule fired on today’s closed bar? Has the stop level been reached? Has the position hit its maximum holding period? Exits come before entries so a full book never blocks you from seeing that something should have left it.

  2. Check the entry rule across the universe

    Go through every symbol in the universe — not just the ones you like — and check the entry condition against today’s closed bar. Skipping symbols you feel bearish about is an override, and overrides are a different strategy.

  3. Size from the sizing line, against current equity

    Compute the position from the spec’s sizing rule and your account’s current equity — 20% of equity means 20% of what the account is worth tonight, not at the start of the month.

  4. Place orders for the next session

    A spec that fills at the open maps to orders for tomorrow’s open. If the strategy carries a stop, place the stop order with the entry — the simulated version never forgets its stop, so the manual version must not either.

  5. Write the journal entry

    Record what fired, what you did, your fill versus the rule’s price, and — most importantly — anything you did differently and why. Two minutes now; the whole feedback loop later.

The pre-trade checklist

Before any order, derived straight from the rules panel:

  • The entry condition is true on a closed bar — verified against the actual numbers, not eyeballed on a chart.
  • The position count after this trade stays within the spec’s maximum.
  • The size comes from the sizing line and current equity, not from conviction.
  • The stop level is computed and the stop order goes in with the entry.
  • The exit conditions are written down where tomorrow’s you will see them.
  • The journal entry exists before the order does.

If any line fails, the trade is not the strategy’s trade. It might still be a fine trade — but log it as yours, not the system’s.

Sizing: the simulation is honest, your account must be too

The backtest and the ledger size positions as a percentage of a paper account. Your account is a different number, and the translation has sharp corners: percentages of a small account round to odd share counts; some brokers do not do fractional shares; a 20%-of-equity position in a $5,000 account is $1,000, and no amount of wanting the backtest’s dollar P&L makes it more. Scale the strategy to your account — never the other way around. If honest sizing makes the returns feel too small to bother with, that is real information about account size, not a reason to triple the risk.

Slippage and fills

The engine charges a slippage penalty on every simulated fill, always in the unfavorable direction — but that is a model, not a promise. Live, you cross real spreads, opens gap through stop levels, and market orders in thin names cost more than the simulation assumed. Expect your fills to be a little worse than the ledger’s on average, and record the difference in your journal: if the gap is persistently large, the strategy’s simulated results are optimistic for you, in your size, in those names — which is worth knowing before the position count grows.

The ledger is your accountability mirror

If the strategy is deployed to the forward ledger, something unusual is true: a frozen, append-only record of what the rules did is being written every day, by the same engine, with no memory of your feelings. Reconcile your journal against it weekly. Every difference between your account and the ledger is one of two things — slippage, which you log and live with, or an override, which is the thing to fix. Most manual traders never get this mirror; if you have it, use it.

When to stop

Decide before the first trade, in writing, what makes you stop — because the version of you that is down 15% is the least qualified person to decide. The backtest’s maximum drawdown is the reference: if the live drawdown grows meaningfully past what the simulation ever produced, the market may have stopped resembling the data the rules were built on. Stopping is not failure; it is the move back to the lab with better information than you left with. Stop, too, when the journal shows you overriding the rules more weeks than not — at that point the record is no longer testing the strategy, and it deserves an honest test or an honest burial.

Historical simulation for research and education. Not financial advice. Past performance does not predict future results.