Alternative Data Rules

Alternative data rules allow you to incorporate non-traditional data sources into your trading strategies. While conventional strategies rely solely on price and volume data, alternative data rules tap into sentiment analysis, AI-generated signals, and economic indicators to provide an informational edge that is not reflected in standard technical analysis.

Alternative data rules are available exclusively on the Live Trading plan. If you are on the Backtesting plan, you can view alternative data rules in the Rules Library but cannot add them to your strategies. View plan details to learn more about upgrading.

What Is Alternative Data?

Alternative data refers to any data source used for investment decisions that falls outside the standard categories of price, volume, and fundamental analysis. In the context of algorithmic trading, alternative data can provide signals that lead or complement traditional price action, offering opportunities that purely technical strategies might miss.

Arconomy integrates three categories of alternative data into its rules engine:

Sentiment Data — Aggregated market sentiment derived from news articles, social media, analyst reports, and institutional positioning data.
AI Signals — Machine learning-generated predictions and pattern recognition outputs from Arconomy's proprietary models.
Economic Indicators — Real-time and scheduled macroeconomic data releases that drive fundamental market movements.

Rules Library panel showing the Alternative Data category with available data sources

Available Data Sources

Sentiment Analysis

The Sentiment Analysis rule evaluates the overall market mood toward an instrument by aggregating data from multiple textual and positioning sources. The rule produces a normalised sentiment score that ranges from -100 (extremely bearish) to +100 (extremely bullish), updated at regular intervals throughout the trading session.

Data Sources Included

News Sentiment — Natural language processing of financial news articles from major wire services and financial publications. Headlines and article bodies are analysed for bullish or bearish language specific to each instrument.
Social Media Sentiment — Aggregated sentiment from financial discussion platforms, filtered and weighted to reduce noise from non-credible sources. Includes volume-weighted sentiment to differentiate high-activity topics from low-activity chatter.
Institutional Positioning — Commitment of Traders (COT) report data and equivalent positioning data for non-futures markets. This data reveals how large speculators and commercial hedgers are positioned, providing a view into institutional sentiment.

Parameter	Description	Default	Range
Sentiment Source	News, Social Media, Institutional, or Composite (all combined)	Composite	—
Bullish Threshold	Minimum sentiment score to allow long entries	20	-100 – 100
Bearish Threshold	Maximum sentiment score to allow short entries	-20	-100 – 100
Lookback Window	Time period over which sentiment is aggregated	24 hours	1 hour – 7 days
Update Frequency	How often the sentiment score is recalculated	1 hour	15 min – 24 hours
Use As	Entry signal, Filter, or Confirmation (weighting factor for position sizing)	Filter	—

Alternative Data Rule: Sentiment Analysis
  Source: Composite
  Bullish Threshold: 30
  Bearish Threshold: -30
  Lookback: 24 hours
  Update: Every 1 hour
  Use As: Filter

  Behaviour:
  - Sentiment >= 30: Long entries allowed
  - Sentiment <= -30: Short entries allowed
  - Between -30 and 30: All entries blocked (neutral zone)

AI Signals

AI Signal rules incorporate machine learning predictions into your strategy. Arconomy's AI models are trained on extensive historical data to identify patterns and regime changes that may be difficult to capture with traditional indicator-based rules. These signals are generated by ensemble models that combine multiple machine learning techniques for more robust predictions.

AI Signal rule configuration showing model selection and confidence threshold

Available AI Models

Trend Prediction — Predicts the probability that the current price trend will continue or reverse over a specified forecast horizon. Outputs a directional probability (e.g., 72% chance of upward continuation).
Regime Detection — Classifies the current market into one of several states: trending, ranging, volatile, or transitional. Each regime has different implications for strategy selection and parameter tuning.
Anomaly Detection — Identifies unusual market behaviour that deviates from historical norms, such as abnormal price movements, unexpected volatility changes, or atypical volume patterns. These anomalies can signal the beginning of a significant move.

Parameter	Description	Default	Range
Model	Trend Prediction, Regime Detection, or Anomaly Detection	Trend Prediction	—
Confidence Threshold	Minimum model confidence required to generate a signal	70%	50% – 99%
Forecast Horizon	The time period the model predicts over	4 hours	1 hour – 7 days
Regime Filter	Only generate signals during specified market regimes (for Regime Detection model)	All Regimes	—
Use As	Entry signal, Filter, or Confirmation	Confirmation	—

AI models are retrained periodically using the latest market data. The models use walk-forward training methodology, meaning they are never trained on future data that would create look-ahead bias. Historical backtests use the model version that was available at each point in time.

Economic Indicators

Economic Indicator rules allow your strategy to respond to macroeconomic data releases and their relationship to consensus expectations. Rather than simply avoiding news events (as the News Event Filter does), these rules use the actual data values to make trading decisions based on whether the release was better or worse than expected.

Supported Indicators

Interest Rate Decisions — Central bank rate decisions and forward guidance statements from the RBA, Fed, ECB, BoE, BoJ, and other major central banks.
Employment Data — Non-farm payrolls (US), employment change (AU, EU), unemployment rate, and wage growth figures.
Inflation Data — Consumer Price Index (CPI), Producer Price Index (PPI), and core inflation measures.
GDP and Growth — Gross Domestic Product preliminary and final readings, PMI surveys, and retail sales figures.
Trade Balance — Import/export data and current account balances, particularly relevant for forex strategies.

Parameter	Description	Default	Range
Indicator	The economic indicator to monitor	—	—
Currencies	Which economies' data to monitor (auto-detected from the traded instrument)	Auto-detect	—
Signal Type	Beat Consensus (bullish), Miss Consensus (bearish), or Surprise Magnitude	Surprise Magnitude	—
Min Surprise	Minimum deviation from consensus required to trigger a signal (in standard deviations)	1.0	0.1 – 5.0
Signal Duration	How long the signal remains active after the data release	4 hours	30 min – 48 hours
Use As	Entry signal, Filter, or Confirmation	Entry	—

Alternative Data Rule: Economic Indicator
  Indicator: CPI (Consumer Price Index)
  Currency: USD
  Signal: Surprise Magnitude
  Min Surprise: 1.5 std deviations
  Duration: 4 hours
  Use As: Entry Signal

  Behaviour:
  - CPI beats consensus by 1.5+ std dev: Bullish USD signal
  - CPI misses consensus by 1.5+ std dev: Bearish USD signal
  - Surprise < 1.5 std dev: No signal

Adding Alternative Data to Your Canvas

Alternative data rules are added to the canvas the same way as any other rule. Open the Rules Library panel, expand the Alternative Data category, and drag the desired rule onto your canvas. The rule will appear with a distinct visual indicator (a purple border) to differentiate it from standard technical rules.

Dragging a Sentiment Analysis rule from the Rules Library onto the strategy canvas

Once on the canvas, connect the alternative data rule to your strategy flow based on its intended role:

As an Entry Signal — Connect it alongside your other entry rules. It will function as an additional entry condition that must be satisfied.
As a Filter — Connect it upstream of your entry rules. It will gate entries based on the alternative data condition, similar to a Trend Filter or Volatility Filter.
As a Confirmation — Connect it to the Position Sizing rule. When the alternative data confirms the entry direction, position size is increased; when it does not, position size is reduced to the minimum.

Start by using alternative data as a filter rather than a primary entry signal. This allows you to evaluate the data source's contribution to your strategy without fundamentally changing your entry logic. If the filter consistently improves performance, you can then experiment with using it as an entry signal.

Configuration Best Practices

Alternative data requires thoughtful configuration to be effective. Here are guidelines for getting the most out of these rules:

Avoid Over-Reliance on Any Single Source

Alternative data sources can be noisy and may have varying levels of reliability across different instruments and market conditions. Never build a strategy that relies exclusively on alternative data for entry signals. Instead, use alternative data to confirm or filter signals generated by proven technical analysis rules.

Validate with Backtesting

Always backtest strategies that include alternative data rules over a sufficiently long period. Because alternative data can be cyclical (for example, sentiment extremes may cluster during crisis periods), a short backtest window may not capture the full range of conditions. Aim for at least two years of backtest history when using sentiment or AI signal rules.

Understand Data Latency

Different alternative data sources have different update frequencies and latencies. Sentiment data may update hourly, while economic indicators are event-driven. Make sure your strategy's evaluation timeframe aligns with the data's update frequency to avoid stale signals driving trading decisions.

Each alternative data rule displays its last update timestamp on the canvas during live trading. If the data is more than two update cycles old, a warning icon appears to alert you that the information may be stale.

Example Strategies Using Alternative Data

Sentiment-Filtered Trend Following

This example combines a traditional Moving Average Crossover entry with a Sentiment Analysis filter. The strategy only takes long entries when the 20/50 EMA crossover fires and the composite sentiment score is above +25, ensuring that the trade aligns with both the technical trend and the broader market mood.

Strategy: Sentiment-Filtered Trend Following

  Filter Layer:
    - Trend Filter: Price above 200 SMA (Daily)
    - Sentiment Filter: Composite score > 25 (for longs)

  Entry Layer:
    - MA Crossover: 20/50 EMA crossover

  Exit Layer:
    - ATR Stop Loss: 2x ATR(14), trailing
    - Ratio Take Profit: 2:1 R:R

  Risk Layer:
    - Position Sizing: 1% risk per trade
    - Max Open Trades: 3

Economic Data Event Strategy

This example uses Economic Indicator rules as the primary entry signal, entering positions when a high-impact data release significantly surprises the market. The strategy is designed for forex pairs and enters in the direction implied by the data surprise, with tight time-based exits to capture the initial reaction move.

Strategy: Economic Data Event Trading

  Entry Layer:
    - Economic Indicator: CPI, NFP, or Rate Decision
    - Surprise Magnitude: > 2.0 std deviations
    - Direction: Aligned with surprise

  Filter Layer:
    - Time-of-Day: Within 5 minutes of release
    - Volatility Filter: ATR below 80th percentile (pre-event)

  Exit Layer:
    - Fixed Take Profit: 30 pips
    - Fixed Stop Loss: 20 pips
    - Time Exit: 2 hours maximum

  Risk Layer:
    - Position Sizing: 0.5% risk per trade
    - Max Daily Loss: 1.5%

AI Regime-Adaptive Strategy

This advanced example uses the Regime Detection AI model to dynamically select which entry rules are active. During trending regimes, the strategy uses breakout entries. During ranging regimes, it switches to mean-reversion entries. The AI model determines the current regime, and the canvas routes the evaluation flow accordingly.

Strategy: AI Regime-Adaptive

  Regime Detection (AI Signal):
    Model: Regime Detection
    Confidence: 75%+

  Trending Regime Branch:
    - Entry: Breakout Detection (20-bar channel)
    - Exit: Trailing Stop (1.5x ATR)

  Ranging Regime Branch:
    - Entry: RSI Threshold (30/70, 14-period)
    - Exit: Fixed Take Profit (1:1 R:R)

  Common Risk Layer:
    - Position Sizing: 1% per trade
    - Max Drawdown: 10%

Regime-adaptive strategies are powerful but add complexity. Ensure you thoroughly backtest each branch independently before combining them with the AI regime detector. Understand how each branch performs during correctly classified and misclassified regime periods.

Plan Availability and Data Coverage

Alternative data features are tiered based on your subscription plan:

Feature	Backtesting Plan	Live Trading Plan
View alternative data rules in library	Yes	Yes
Add to canvas and backtest	No	Yes
Deploy with alternative data rules	No	Yes
Sentiment data (composite)	—	Major forex pairs, indices
AI Signals (all models)	—	All supported instruments
Economic Indicators	—	G10 economies + AU
Historical data depth	—	3+ years

To upgrade your plan and access alternative data rules, visit the Plans & Pricing page or contact support for assistance.

Previous ← Risk Management Rules Next How Backtesting Works →

Was this helpful? Let us know