Learn TessTrade

Appearance

Example script ​

A minimal script that buys whenever the price rises relative to the previous bar, with no indicators. Uses:

  • Confirm that the Code Editor is working.
  • Observe signals on the chart.
  • Understand the main() dispatcher in a real run.

This is not a profitable strategy - it is only a working skeleton.

1. Open the Code Editor ​

In Backtest or Chart Trading, click Editor. A code window appears with a scaffold.

2. Erase the scaffold and paste the code below ​

python
DECLARATION = {
    "type": "strategy",
    "inputs": [
        {
            "name": "qty",
            "label": "Quantity",
            "type": "float",
            "default": 1.0,
            "min": 0.001,
            "max": 1000.0,
            "step": 0.001,
        },
    ],
}


def on_bar_strategy(sdk, params):
    qty = float((params or {}).get("qty", 1.0))

    # Needs at least 2 candles to compare the previous one with the current one.
    if len(sdk.candles) < 2:
        return

    last_close = sdk.candles[-1]["close"]
    prev_close = sdk.candles[-2]["close"]
    went_up = last_close > prev_close

    if sdk.position == 0 and went_up:
        sdk.buy(
            action="buy_to_open",
            qty=qty,
            order_type="market",
        )
    elif sdk.position > 0 and not went_up:
        sdk.sell(
            action="sell_to_close",
            qty=abs(sdk.position),
            order_type="market",
        )


def main(df=None, sdk=None, params={}):
    params = params or {}
    if sdk is not None:
        return on_bar_strategy(sdk, params)
    # This script does not draw plots, so the `df=` branch returns empty.
    if df is not None:
        return {"plots": [], "series": {}}
    return DECLARATION

3. Click Run (or Backtest) ​

  • In Backtest: choose the symbol, period, and click start. In 1-2 min the results panel appears.
  • In Chart Trading: start a paper trading bot (see paper bots).

4. What to expect ​

Dozens of trades. Win rate close to 50% (coin flip of the previous bar). Net profit close to zero, visible drawdown. This behavior is expected: it is a noise generator, not a strategy.

Checkpoints:

  • The parameter panel appeared with the editable "Quantity" field.
  • Orders were emitted (markers on the chart).
  • Equity evolved candle by candle.
  • The script compiled and executed without error.

5. Variations ​

Modifications in order of difficulty:

Only buy when it rises 2 bars in a row ​

python
prev_close = sdk.candles[-2]["close"]
prev_prev_close = sdk.candles[-3]["close"]
went_up_twice = sdk.candles[-1]["close"] > prev_close > prev_prev_close

Add an indicator and plot it on the chart ​

See SMA Crossover, which contains a full template commented line by line.

Store state between bars ​

python
if not isinstance(sdk.state, dict):
    sdk.state = {}
sdk.state["trades_taken"] = sdk.state.get("trades_taken", 0) + 1

Details in persistent state.

6. Error handling ​

"Strict Mode" error ​

The main() function was not defined at the root level. Check that it is not indented.

"explicit action" error ​

sdk.buy() was called without action=. All calls require that argument - the canonical actions documentation lists the 7.

"Import not allowed" ​

Import outside the whitelist. See sandbox limits. There are no imports in this script; if the error appears, remove any import that is not numpy/pandas/math/json/datetime.

0 trades ​

The DECLARATION declares entry_conditions but the script also defines on_bar_strategy. These modes are mutually exclusive (details in when to use declarative mode). Remove entry_conditions from the DECLARATION.

Empty parameter panel ​

DECLARATION["inputs"] is empty or main() does not return DECLARATION in the no-argument branch. Review the contract in main dispatcher.

Next steps ​