I Built a Python Tool That Automatically Discovers Trading Strategies For You

Stop building strategies by hand. Let the machine find the combinations worth testing.

Kryptera

~5 min read · May 19, 2026 (Updated: May 25, 2026) · Free: No

Get the Bundle: Simple Indicators Strategy Generator — Python Code Bundle

Free Version Here: Simple Indicators Strategy Generator — Python Code Bundle (Lite / Free Version)

I want to tell you about a small Python prototype I built that has quietly become one of the most useful tools in my trading workflow.

The idea started from a simple frustration.

Every time I wanted to test a new strategy idea, I had to go through the same slow process. Pick an indicator. Decide what condition to use. Write the signal logic. Run the backtest. Look at the result. Adjust something. Run it again. It worked, but it was slow — and I was always making the decisions based on what I already thought I knew. There was no real exploration happening.

So I asked myself: what if the tool just tried combinations on its own?

What It Does

The Simple Indicator Strategy Generator is a Python script that automatically assembles random indicator conditions into trading strategies and tests whether they work — without you having to decide anything upfront.

Here's the basic flow.

You give it a ticker and a date range. The script downloads the data, computes 295 boolean indicator conditions across 34 different indicator families, then starts randomly drawing combinations for entry and exit signals. For each combination it runs a backtest, checks if the result is any good, and either keeps it or throws it away.

The part that makes it actually useful is the validation step. The script doesn't just test a strategy on one stretch of data. It splits the historical data into three separate windows — in-sample, out-of-sample, and a final holdout — and the strategy has to beat buy-and-hold on all 3 independently before it gets returned to you. If it fails any one of them, the loop discards it and keeps searching automatically.

When a strategy finally passes all three phases, the script prints it out as complete, ready-to-run Python code. Imports, indicator functions, entry and exit signals, backtest block — all of it. You copy it, save it as its own file, and it runs on its own.

The Indicator Library

The 295 conditions come from 34 indicator families. This is what the generator has to work with:

Bollinger Bands, Hull Moving Average, KAMA, Kaufman Efficiency Ratio, Laguerre RSI, MACD, Momentum, SMA, WMA, EMA, OsMA, QQE, Reflex, Rate of Change, RSI, Schaff Trend Cycle, SR Percent Rank, Standard Deviation, Ulcer Index, CMO, Elder Impulse System, DPO, QStick, TRIX, Williams Alligator, Chande Trend Score, TEMA, DEMA, McGinley Dynamic, VIDYA, Coppock Curve, TSI, Historical Volatility, and classic candlestick patterns.

Each family contributes multiple specific conditions — not just "RSI is high" but things like RSI exiting overbought, RSI in a bullish regime, RSI accelerating upward, RSI flat. Same for every other indicator. The generator picks from all of these when building each candidate strategy.

What You Do With the Output

This is the part I want to be clear about: the strategy the tool generates is a prototype, not a finished system.

Think of it as the first draft. The tool did the exploration work — it searched through thousands of random combinations and surfaced one that showed consistent results across 3 separate data windows. Now you take over.

After running it, the tool prints ready-to-copy-paste Python code — in the same format as the example systems featured in each of my articles:

ALL-Elder-RSI Mean Reversion Strategy

Beat Buy and Hold with 2 Simple Indicators

medium.com

AKAM-QQE Deceleration & EMA Filter Strategy

When Momentum Sleeps, the Trade Wakes Up

medium.com

You can look at the conditions it found and ask whether they make sense together. You can run the generated code through a more rigorous stress test — walk-forward optimization, Monte Carlo simulation, testing on different tickers or timeframes. You can add filters, adjust the exit logic, or use the discovered conditions as a starting point for a more refined version.

The point is that you're starting from something that already passed a basic validation rather than starting from a blank page. That changes the workflow significantly. Instead of spending time building strategies from scratch and hoping they hold up, you're spending time developing and stress-testing candidates that have already shown some promise.

How to Run It

The script runs in Google Colab or any standard Python environment. You need vectorbt, pandas, numpy, and yfinance installed. Setup takes a few minutes.

Once it's running, you set your ticker, your date range, and a minimum trade count. The generator loops until something passes all three validation phases and then prints the result. On most liquid stocks with long daily price histories, this usually takes anywhere from a few seconds to a few minutes.

There's one parameter worth knowing about: size, which controls how many conditions are drawn per entry and exit signal. At the default of 1, each entry is a single condition and searches are fast. Set it to 2 or 3 and the combinations become more complex — harder to pass, but sometimes more interesting when they do.

Why I'm Sharing This as a Prototype

I'm releasing this as a code bundle rather than a polished product because that's genuinely what it is.

It's a working prototype that I use in my own research pipeline. The validation logic is real. The indicator library is real. The code output is clean and usable. But it's not a finished trading system — it's a research tool that accelerates the early stage of strategy development.

If you're a developer or a quant trader who already has a process for testing and developing strategies, this fits naturally into that workflow. You get the exploration and initial filtering for free. Everything after that — the rigorous testing, the refinement, the deployment decisions — is still yours to do.

Get the Bundle

The Simple Indicator Strategy Generator is available on Gumroad as a single Python file. Everything is included: the full indicator library, the 3-phase validation engine, and the strategy code printer.

If you want to spend less time searching for strategy ideas and more time developing the ones worth pursuing, this is built for exactly that.