Docs / Build a strategy

Strategy recipes

Three complete strategies you can build in the builder step by step, then adapt to your own ideas.

Three complete strategies, each built entirely in the builder. Follow one end to end to see the pieces fit together, then change the market, the thresholds, or the indicators to make it yours.

Each recipe follows the same order every strategy does: create the shell, add indicators, write signals, set risk, then backtest. If any step is unfamiliar, the strategy builder, indicators, signals and risk management pages cover it in full.

TIP

The golden rule for signals: add the indicator first, then open the Line dropdown in a condition to see its exact line name (like RSI.rsi). Reference lines from the dropdown rather than typing them, so you can never misname one.

Recipe 1 — RSI mean reversion#

Buy oversold, sell overbought. BTC/USDT, 1h.

  1. New Strategy — Name RSI Mean Reversion, Ticker BTC/USDT, Timeframe 1h.
  2. Indicators → Add Indicator — RSI, length 14.
  3. Signals → Add Signal, name it rsi. Then Add Rule twice:
    • Enter · long — RSI.rsi goes above 30.
    • Exit · long — RSI.rsi goes above 70.
  4. Risk Management → Set Stop Loss — StopLossSimple, stoploss_pct -0.05 (a 5% stop).
  5. Run Backtest, then read the Reports tab.

The whole idea is on the entry: goes above 30 fires as price leaves oversold, not while it sits there. Swap the levels to 25 / 75 to trade only deeper extremes.

RSI oscillates between 30 and 70; enter long as it crosses up through 30, exit as it crosses up through 70 70 30 overbought oversold enter long RSI goes above 30 exit RSI goes above 70
You hold the position while RSI travels from the oversold band up to the overbought band. The 5% stop is the safety net for the times it keeps falling instead of bouncing.

Recipe 2 — Bollinger Band breakout#

Enter as price breaks the upper band, exit back at the middle, trail the stop. ETH/USDT, 4h.

  1. New Strategy — Name BB Breakout, Ticker ETH/USDT, Timeframe 4h.
  2. Indicators → Add Indicator — BollingerBands, length 20, standard deviation 2.0. This publishes BollingerBands.bband_top, bband_mid and bband_low.
  3. Signals → Add Signal, name it breakout. Add Rule twice:
    • Enter · long — close goes above Threshold line BollingerBands.bband_top.
    • Exit · long — close goes below Threshold line BollingerBands.bband_mid.
  4. Risk Management → Set Stop Loss — StopLossTrailing, trailing_pct 0.03. A trailing stop suits a breakout: it lets a run continue and only closes you out once price gives back 3%.
  5. Run Backtest.

This is a trend-following idea, so expect a lower win rate than Recipe 1 but larger average winners. Read the profit factor, not just the win rate.

Price breaks above the upper Bollinger band to enter, and falls back to the middle band to exit upper band middle band lower band enter — break above upper exit — back to middle
The entry fires as price pushes out through the upper band — the sign of a breakout — and the position rides the move until price falls back to the middle band. The trailing stop protects the run if it reverses hard first.

Recipe 3 — Trend entry gated by news sentiment#

Only take moving-average entries while the news flow is net-bullish. BTC/USDT, 1h. Combines a signal, a bias, and the news feed.

  1. New Strategy — Name Sentiment Trend, Ticker BTC/USDT, Timeframe 1h.

  2. Indicators → Add Indicator twice:

    • EMA, length 50 — publishes EMA.ema.
    • NewsSentiment — publishes NewsSentiment.sentiment_score, NewsSentiment.article_count, and a few more. (There is a one-click banner on the Indicators tab to add it.)

  3. Bias Generators → Add Bias Generator, name it news_bias. Add Rule:

    • Bias long when NewsSentiment.sentiment_score is greater or equal to 0.2 AND NewsSentiment.article_count is greater or equal to 2.

    The second condition stops a single stray article from swinging the bias.

  4. Signals → Add Signal, name it trend. Add Rule:

    • Enter · long — close goes above Threshold line EMA.ema AND news_bias equals 1.

    The moving-average cross is the event that fires the entry; the bias is the state that filters it.

  5. Risk Management — StopLossSimple stoploss_pct -0.04, and a TakeProfitSpread that banks half at +5% and the rest at +10%.

  6. Run Backtest.

The entry fires only when the price crosses above the moving average and the news bias is long at the same time Event — the trigger close goes above EMA.ema State — the filter news_bias equals 1 AND Enter long only when both hold
Two conditions, joined with AND: the moving-average cross is the moment to act, and the persisted news bias is the permission to act. Either one alone does nothing — the entry needs the cross to happen while the news mood is bullish.

This is the platform's core combination in miniature: a signal shaped by sentiment, which a hedge-fund Squad can then review before anything trades. See news and sentiment for what the sentiment lines mean and how they are scored.

Turn a good strategy into a grid#

Once a recipe backtests reasonably, do not stop at one setting — sweep the parameter you are least sure about to see whether the result is robust or lucky. From the strategy, create a grid template, set a range (say RSI length from 10 to 20), check the variation count against your plan limit, and run it. Look for a region of settings that all work, not a lone spike. Full detail in grid search.

Related
The strategy builder How to create a strategy and where every part of it lives — the tabs across the top of the strategy page. Signals The rules that open and close positions — how the condition builder works, and the one distinction that decides whether a rule fires once or constantly. Grid search Backtest a whole range of parameter settings at once to find a robust region rather than a lucky spike.
Was this page helpful?
Previous Entry strategy Next Backtesting
Community Docs About Discord Gateway Sign Up Login
© 2026 riskmanaged.io — All rights reserved.
Trading carries substantial risk of loss. Nothing here is financial advice, and past or simulated performance does not predict future results.