Using Cody: Your AI Trading Assistant

You: "Write a strategy that buys Bitcoin when the
price drops 5% in a day and sells when it recovers 3%"

Cody: Here's a mean reversion strategy based on daily
price drops...

[Cody generates complete, runnable code]

# Good prompts - specific and clear
"Buy ETH when RSI drops below 25, sell when it rises above 65"

"Create a momentum strategy for AAPL using 20 and 50 day
moving averages with a 2% stop loss"

"Write a breakout strategy that enters when price breaks
above the 20-day high with volume 50% above average"

# Vague prompts - Cody will ask for clarification
"Make me money" → Too vague
"Trade stocks" → Which ones? What signals?
"Good strategy" → Define "good"

You: "Explain what this line does:
entries = (fast_ma > slow_ma) & (rsi < 50)"

Cody: This creates entry signals where BOTH conditions
are true:

1. fast_ma > slow_ma: The fast moving average is above
   the slow one, indicating an uptrend

2. rsi < 50: RSI is below 50, meaning momentum hasn't
   peaked yet

The & operator requires both conditions. You'll only
enter trades when the trend is up AND there's still
room for the move to continue.

You: "My strategy has zero trades. Here's the code:
[paste code]"

Cody: I see the issue. Your entry condition requires
RSI < 20, but looking at the data, RSI never drops
that low for this asset.

Try RSI < 30 instead, or check if your data range
includes any significant drawdowns.

You: "Why is my Sharpe ratio negative?"

Cody: A negative Sharpe ratio means your strategy
lost money on a risk-adjusted basis. Looking at your
backtest:

1. Total return: -12%
2. You're trading too frequently (847 trades)
3. Fees are eating your profits (0.1% × 847 = 85% drag)

Consider: fewer trades, lower fee broker, or wider
entry thresholds.

You: "Add a stop loss to my strategy and explain
how stop losses work"

Cody: A stop loss automatically exits your position
if the price moves against you by a certain amount.
It limits your downside risk.

Here's your strategy with a 5% stop loss:

pf = vbt.Portfolio.from_signals(
    close=closes,
    entries=entries,
    exits=exits,
    sl_stop=0.05,  # Exit if price drops 5% from entry
    init_cash=10000
)

With this stop loss, no single trade can lose more
than 5% of its entry value. The tradeoff: you might
get stopped out of trades that would have recovered.

You: "My strategy has a 1.2 Sharpe ratio. How can
I improve it?"

Cody: A 1.2 Sharpe is decent but there's room to
improve. Based on your backtest results:

1. Your win rate is 45% but winners are 2x losers -
   that's good, but try tighter stops to improve win rate

2. You're holding positions 12 days average - consider
   adding a time-based exit for trades that go nowhere

3. Your entries cluster around market opens - try
   filtering for time of day

Want me to implement any of these changes?