9. Output Data Structures & Comprehensive Data Dictionary

The Promising Score Analyzer (technosight.analytics.promising_score) compiles the primary candidate export file promising_symbols.csv following a strict 3-part schema:

$$\text{Output Schema} = \text{Common Prefix (8)} + \text{Strategy-Specific Columns} + \text{Common Suffix (12)}$$

1. Mathematical Formulation of the 3-Factor Promising Score

The composite Promising Score evaluates candidate symbols using a 3-factor linear model combining trading performance, strategy signal intensity, and directional sentiment:

$$\text{Promising Score} = w_1 \cdot \text{norm}(\text{Avg Gain \%}) + w_2 \cdot \text{norm}(\text{Score}) + w_3 \cdot \text{norm}(\text{Sentiment})$$

• Normalized Average Gain % ($\text{norm\_avg\_gain}$): Measures average return across scenarios where an entry signal was generated.
• Normalized Strategy Signal Score ($\text{norm\_score}$): Measures strategy factor conviction ($R_{20}$ for Momentum, $\text{Score}_{\Omega}$ for Momentum Omega, $-Z$ for Mean Reversion/Stat Arb, MA spread for Trend Following, channel distance for Volatility Breakout).
• Normalized Side-Aware Sentiment ($\text{norm\_sentiment}$): $\text{Sentiment} \times \text{sign}(\text{Position})$.
• Default Weights: $w_1 = 0.40$ (Gain %), $w_2 = 0.60$ (Signal Score), $w_3 = 0.00$ (Sentiment).

2. Pre-Enrichment Candidate Selection Boundary

To ensure high processing throughput and avoid external API rate limits, candidates are partitioned by direction and pruned to at most 100 symbols:

• top_long: Top 50 Long candidates sorted by promising_score descending.
• top_short: Top 50 Short candidates sorted by promising_score descending.
• External FMP APIs are queried strictly for these $\le 100$ symbols.

promising_symbols.csv strictly adheres to a non-duplicated 3-part column layout:

Part A: Common Prefix Columns (8 Fields)

1. Symbol: Equity ticker symbol (e.g. AAPL).
2. Company Name: Full legal corporate entity name.
3. Position: Trade stance (long or short).
4. Promising Score: 3-factor normalized composite score $\in [0.0, 1.0]$.
5. Average Gain (%): Mean simulated gain across active scenarios.
6. Current Price: Most recent closing price in USD.
7. Price Target Consensus: Mean 12-month Wall Street analyst price target.
8. Price Target Median: Median 12-month Wall Street analyst price target.

Part B: Strategy-Specific Technical Columns

• Mean Reversion: Z-Score, Moving Average (SMA/EMA), Standard Deviation (σ), Half-Life of Mean Reversion, Relative Strength Index (RSI).
• Statistical Arbitrage: Z-Score, Moving Average (SMA/EMA), Standard Deviation (σ), Half-Life of Mean Reversion, Spread Deviation (%).
• Standard Momentum: Average Momentum (252-day cumulative return).
• Momentum Omega: Omega Ratio / Score ($\text{Score}_{\Omega}$).
• Trend Following: Trend Score / Signal Intensity, Fast SMA, Slow SMA, Breakout High, Breakout Low, Autocorrelation, Drift Coefficient, Average True Range (ATR), Volatility-Scaled Position Size, Trailing Stop-Loss Price, Ex-Ante Annualized Volatility.
• Volatility Breakout: Upside Breakout Price, Downside Breakdown Price, Historical Base Range, Stop-Loss Level, Take-Profit Target, Dynamic Trailing Stop Multiplier, Volatility Compression Ratio, Volume Expansion Multiple, ADX Value.

Part C: Common Fundamental Valuation & Consensus Suffix Columns (12 Fields)

1. Revenue Growth: Trailing twelve-month revenue growth rate ($y/y$).
2. Forward PEG: Forward Price/Earnings-to-Growth ratio.
3. FCF Yield: Free Cash Flow Yield ($\text{FCF} / \text{Market Cap}$).
4. Strong Buy: Number of Wall Street analysts issuing Strong Buy ratings.
5. Buy: Number of analysts issuing Buy ratings.
6. Hold: Number of analysts issuing Hold ratings.
7. Sell: Number of analysts issuing Sell ratings.
8. Strong Sell: Number of analysts issuing Strong Sell ratings.
9. Consensus: Aggregate Wall Street consensus recommendation grade.
10. Consensus Buy %: Proportion of analyst recommendations that are Buy or Strong Buy.
11. Consensus Sell %: Proportion of analyst recommendations that are Sell or Strong Sell.
12. EMA(50,200,800)15m: Multi-timeframe 15m trend pattern label (AAA, BBB, NEUTRAL, BYPASS). Displayed for reference only; no candidates are dropped based on this field.

1. Parameter Sweep Summary (sweep_summary.csv / .parquet)

• scenario_id: Evaluated scenario code (e.g. A1 to G5).
• liquidity_tier: Tranche letter (A through G).
• min_adv / max_adv: Dollar volume threshold boundaries in USD.
• total_signals: Total non-neutral signals generated.
• long_signals / short_signals: Directional trigger counts.
• confirmed_rate: Proportion of triggers passing 15m EMA regime alignment.
• mean_score: Average strategy conviction score across active equities.
• exec_time_sec: Execution runtime in seconds.

2. Full Vectorized Signals Dataset (signals.parquet / signals.csv)

• timestamp: UTC datetime of the evaluation bar.
• symbol: Equity ticker symbol.
• close: Split-adjusted closing price.
• signal: Discrete trading directive (+1.0, -1.0, 0.0).
• direction: String directive (LONG, SHORT, NEUTRAL).
• raw_score: Continuous strategy conviction score.
• regime_15m: Intraday 15m EMA regime classification.