HN Showcase
Back to browse
ZcoreAI – Z-score regression channel screener

ZcoreAI – Z-score regression channel screener

by tchantchov·Mar 1, 2026·1 point·0 comments
Visit ProjectView on HN

AI Analysis

●●SolidNiche Gem

Regression channel Z-scores across timeframes, but TradingView and QuantConnect dominate quant screening.

Strengths
  • •Novel quantitative approach: Donchian-weighted regression channels + Z-score matrix uncommon in free scanners
  • •Client-side computation via yfinance avoids data licensing costs; Streamlit deployment is accessible
  • •Free tier, no signup, picks up preloaded watchlist immediately
Weaknesses
  • •Extremely niche: requires understanding of regression channels and Z-scores; no education or UX guidance
  • •No evidence of performance testing or backtest validation; methodology transparency missing
  • •Competes against entrenched platforms (TradingView, QuantConnect) with richer signal libraries
Category
Finance
Target Audience

Quantitative traders and retail investors seeking statistical overbought/oversold signals across watchlists.

Similar To

TradingView screener · QuantConnect · Finviz Elite

Post Description

Hey HN,

I built ZcoreAI, a quant stock scanner that applies Donchian-Weighted regression channels to compute Z-scores across multiple timeframes simultaneously.

The idea: instead of eyeballing charts, you get a matrix of Z-score values per ticker per timeframe in one scan — so you can spot statistically overbought/oversold signals across your watchlist in seconds.

How it works: - Pick timeframes (1m to 1wk) - Pick tickers (or use the preloaded free watchlist) - Hit scan — regression channels + Z-scores are computed client-side via yfinance

Stack: Python, Streamlit, yfinance, NumPy. Deployed on Render.

Free tier available, no signup required to scan.

Happy to discuss the regression channel methodology or any feedback on the approach.

Hey HN,

I built ZcoreAI, a quant stock scanner that applies Donchian-Weighted regression channels to compute Z-scores across multiple timeframes simultaneously.

The idea: instead of eyeballing charts, you get a matrix of Z-score values per ticker per timeframe in one scan — so you can spot statistically overbought/oversold signals across your watchlist in seconds.

How it works: - Pick timeframes (1m to 1wk) - Pick tickers (or use the preloaded free watchlist) - Hit scan — regression channels + Z-scores are computed client-side via yfinance

Stack: Python, Streamlit, yfinance, NumPy. Deployed on Render.

Free tier available, no signup required to scan.

Happy to discuss the regression channel methodology or any feedback on the approach.

https://www.zcoreai.com/

Similar Projects

Finance●Mid

I built a simple quant scanner for mean-reversion setups (ZcoreAI)

Computes a regression‑channel Z‑score across timeframes and presents a compact signal matrix (with an ‘expert’ view that exposes raw Z values) — a useful narrow tool if your workflow is 'find oversold across many symbols fast'. It’s a focused MVP that nails the scanning idea, but crucial details are missing: data source and backtest/validation, alert/export hooks, and a more usable workflow to act on the signals.

Niche GemShip It
tchantchov
113mo ago