Clean Data No Survivorship Bias Blazing Fast Splits Handled Late Prints Normalized Clean Data No Survivorship Bias Blazing Fast Splits Handled Late Prints Normalized

The no-code backtesting library

Backtest with confidence.

A no-code Python library that gives you clean Polygon data—no survivorship bias, splits handled, late prints normalized. Scan and backtest in minutes, not months.

clean_data.py — backtest complete

Pre-processed • Point-in-time correct • No phantom gaps

From Idea to Backtest

You're only a few simple steps away from backtesting with clean, reliable data.

Step 1

Install

pip install kwants

Step 2

Config

Add your Polygon API key

Step 3

Define

YAML or minimal config for your strategy

Step 4

Run

One command to backtest

Step 5

Ship

Export results, move to live trading

What we fixed for you

Comes with all the essential features to get your backtests off the ground—fast.

Data Integrity

No Survivorship Bias

Point-in-time ticker universe; no future stocks in past backtests.

Splits Handled

Automatic adjustment for splits; fake splits filtered out.

Late Prints Normalized

Intraday spikes smoothed; no phantom gaps in your charts.

Performance

Blazing Fast

Pre-processed data for vectorization; scan thousands of tickers in seconds.

Polygon Native

Built for Polygon flat files; no extra API complexity.

Developer Experience

No-Code Setup

Configure strategies without writing backtest logic.

Reproducible

Same inputs = same outputs; deterministic and auditable.

I've been through it all.

Duplicate entries. Wrong splits. Data you can't trust.

Survivorship bias that makes every backtest a lie.

Late prints that look like alpha until you look closer.

Months of manual backtesting before a single valid result.

Expensive data that still has holes.

Kwants gives you the data quality pros pay for, without the price tag—and backtests that actually run.

Powered by modern tech

Built with the latest technologies for performance, reliability, and developer experience.

Python

Industry-standard for quantitative research

Polygon

Data source and flat file integration

NumPy / Pandas

Vectorized operations for speed

Parquet

Efficient columnar storage

YAML

Config-driven, no-code strategies

Built from real pain.

After 8+ hours of wrestling with Polygon data, wrong splits, and survivorship bias that turned backtests into fiction—Kwants was built to fix it.

The methodology is documented in depth:

📊

Join the kwants community

Connect with traders building with clean data.

Join Discord

Start backtesting today

Get access to clean data and blazing-fast backtests. One-time payment.

$497

One-time payment.

Get Kwants

Refund policy · FAQ