Sentiment Analysis Algorithm

Entry Logic

• Long entry is triggered when a sentiment analysis model detects a significant positive shift in sentiment from news and social media sources.
• Short entry is triggered by a significant negative shift in sentiment.
• Confirmation requires a corresponding price action signal, such as a breakout or reversal pattern.
• Timeframe is typically intraday (e.g., 30-minute chart).
• Location context is not the primary driver, but trades taken at key support/resistance levels have a higher probability.
• Market condition is a high-sentiment environment.

Exit Logic

• Profit target is a fixed percentage gain or when sentiment returns to neutral.
• Scale out at predefined profit levels.
• Trailing stop is used to protect profits.
• Exit on signal failure if the price does not follow the sentiment shift.
• Exit on opposite signal if sentiment reverses.
• Exit on time expiration after a set holding period.
• Exit on momentum loss if the price stalls.

Stop Loss Structure

• Hard stop is placed at a level that invalidates the entry signal.
• Soft stop is not used.
• Maximum dollar loss is defined per trade.
• Maximum percent loss is a set percentage of the account.
• Structural stop is based on the price action pattern.

Risk Management Framework

• Risk per trade is a fixed percentage of the account.
• Maximum daily and weekly loss limits are enforced.
• Maximum drawdown is monitored.
• Risk-reward ratio is based on historical performance.

Position Sizing Model

• Sizing can be fixed fractional or adjusted based on the strength of the sentiment signal.
• Volatility adjustment can be used.
• Conviction sizing is based on the magnitude of the sentiment shift.
• Scaling in is not typically used.
• Scaling out is performed at profit targets.

Trade Filtering

• Filter out weak sentiment signals.
• Avoid trading in low-liquidity markets.
• Instrument selection is based on which assets are most sensitive to sentiment.
• Time of day restrictions may apply depending on when sentiment-driving news is released.
• Avoid trading around scheduled news events that can override the sentiment signal.

Context Framework

• The sentiment signal provides the primary context.
• The relationship to VWAP and moving averages can be used as a secondary filter.
• Higher timeframe sentiment can be used to confirm the signal.

Trade Management Rules

• Move stop to breakeven after a certain profit is achieved.
• Scale out at predefined levels.
• Do not add to positions.
• Be prepared for fast moves driven by sentiment.

Time Rules

• Optimal trading window is when sentiment-driving news is most active.
• Avoid trading during periods of low news flow.
• Session-specific sentiment patterns can be identified.

Setup Classification

• A+ setup: Strong sentiment signal with confirming price action and high volume.
• A setup: Strong sentiment signal.
• B setup: Moderate sentiment signal.
• C setup: Weak sentiment signal (avoid).

Market Selection Criteria

• Instruments are those that are heavily discussed in the media, such as popular stocks and cryptocurrencies.
• High liquidity is essential.
• The asset should have a history of reacting to sentiment.

Statistical Edge Metrics

• Metrics are derived from backtesting the sentiment analysis model.

Failure Conditions

• The sentiment signal can be a false positive.
• The market may not react to the sentiment shift as expected.
• The sentiment data may be noisy or inaccurate.

Psychological Rules

• Be aware of the risk of trading on sentiment, which can be fickle.
• Stick to the trading plan and avoid being swayed by anecdotal evidence.
• Trust the model but be prepared for it to be wrong.

Advanced Components

• Natural Language Processing (NLP) techniques are used to analyze the sentiment of text data.
• Machine learning models can be trained to predict price movements based on sentiment.
• The sentiment data can be combined with other data sources to improve performance.

Location

• The strategy is most effective in markets with a high level of retail participation and media coverage.
• It may be less effective in more institutionalized markets.
• The source and quality of the sentiment data are critical.