Tickers-Dual-Rank — Fact-Based Equity Scanner

Broker-agnostic Python scanner that ranks US equities every scan cycle
No opinions, no hype — every ranked field is backed by a computed metric
Powers the live leaderboards on tickersrank.ai

Overview

Tickers-Dual-Rank scans hundreds of US equities daily and sorts them into groups — not by predicting price, but by observing trend strength and mean-reversion setups:

• Leaders — momentum + trend confirmed candidates.
• Reversion — oversold / undervalued names showing early stabilization.
• Decliners — weakening names, ranked for context.

Every ranked field is backed by a computed metric — RSI, ADX, CCI, EMA structure, multi-timeframe slope, and IV percentile. There are no discretionary opinions in the ranking.

How it ranks

• Slope analysis against the S&P (relative trend strength).
• RSI / CCI trend staging and mean-reversion detection.
• Triple-timeframe scan (30m / 1h / 4h) for multi-timeframe confirmation.
• EMA crossovers (8 / 21 / 50 / 200) for trend structure.
• ADX for momentum quality.
• IV percentile — true percentile-of-history, not min-max IV Rank.

Price and average volume are the only hard universe filters. PE and PEG are display-only — shown per stock for your own fundamental read, but they do not feed the composite score or decide which names make the universe (a genuine trend/momentum leader often carries a "bad" PEG by design).

Data

Real Schwab market-data integration is wired in — daily/intraday bars, option-chain IV, and fundamentals. A Yahoo Finance path (--broker yfinance) works standalone for anyone without broker API access. Broker access goes through the same BrokerRouter pattern used across the bot suite (Schwab primary, yfinance fallback).

Outputs

• Discord — compact monospace leaderboard tables (Leaders, Reversion, Decliners): rank, symbol, price, RSI, ADX, PE, PEG, IV, and trend.
• Live tables — the full scan is persisted to the ingestion service, which keeps the tickersrank.ai leaderboards current.
• Ticker-list export — the leaders/reversion lists can be exported and fed directly into the other bots (e.g. wheeling-alerts) as their scan universe.

Design notes

• Modular by design — universe handling, scan logic, scoring, and broker access are separate modules, which makes the model easy to tune.
• IV percentile is a genuine historical store — it needs ~20 daily samples before returning a reading, so new symbols show null for the first few weeks by design.
• 60m / 4h bars are synthesized from native 30m bars (session-anchored), since the data provider rejects native 60-minute minute-granularity requests.
• GEX / gamma exposure is intentionally out of scope here — it belongs to a separate gamma library used by short-DTE strategies.

Disclaimer

For educational and research use. Not financial advice. Rankings describe observed metrics, not predictions.