About Gecko

Gecko is a personal quantitative trading research project built to investigate whether trading strategies and rules actually demonstrate a measurable historical edge.

It began in 2026 with a simple question: how can I determine whether the trading strategies and ideas I encounter really work?

Gecko uses historical market data to backtest strategies, compare variations and investigate specific hypotheses about market behaviour.

This site documents that research — including successful experiments, failed hypotheses and findings that are interesting even when they don't constitute a tradeable edge.

How a finding gets validated

A backtest result on its own isn't proof of anything. Gecko's process works through progressively stricter stages, each one designed to catch a failure mode the stage before it structurally can't:

In-sample backtest — tune and test a strategy against the same historical data. Fast, and useful for narrowing down what's worth pursuing, but a result here can look good partly because it was shaped to fit that specific history, not because the underlying edge is real.

Out-of-sample / walk-forward backtest — the finalised configuration is tested once, cold, against a later slice of history it was never tuned against (or repeatedly, in a rolling window, for a proper walk-forward test — the technique and its name follow Robert Pardo's The Evaluation and Optimization of Trading Strategies, the standard reference on the subject). Still historical, still simulated, but it checks whether a result survives beyond the exact data it was found on.

Forward test (paper trading) — running the strategy on genuinely live data as it happens, with no money at risk. This is the first stage that can catch what no historical test can: real spread, real slippage, real fill behaviour, and whether the market has actually moved on from anything in the historical record.

Live trading with real capital — the final stage, and the only one that tests execution with genuine consequences attached.

Most of what's published on this site so far sits at the first stage. Working through the second, and eventually the third, is a deliberate, ongoing part of the project rather than a formality.

About me

I'm a software engineer specialising primarily in Java and AWS. I came to trading through investing and quickly found that my software-engineering instincts were difficult to suppress.

Rather than accepting that a particular trading ruleshould work, I wanted to test whether it actually did.

Gecko is the result.

Daniel Butcher
May 2026

Reading list

The books on my shelf (and in my headphones) while building Gecko. Where one has shaped a piece of Gecko's research, it links to the write-up.

Strategy and technical analysis

How to Make Money in Stocks — William J. O'Neil. The source of the cup-and-handle rules Gecko put to the test.
Trading Classic Chart Patterns — Thomas N. Bulkowski. See the chart patterns studies.
Getting Started in Technical Analysis — Jack D. Schwager
The Beginner's Guide to the Stock Exchange — Ross Larter. Defined the structure for the three indicators used behind the TFM strategy as used in the earlier backtests.

Psychology and discipline

Best Loser Wins — Tom Hougaard. Prompted Gecko's look at trending versus ranging markets, and led to testing hisAdvanced School Run strategy.
Trading in the Zone — Mark Douglas

Traders' stories

Reminiscences of a Stock Operator — Edwin Lefèvre
Unknown Market Wizards — Jack D. Schwager
Market Wizards: The Next Generation — Jack D. Schwager and George Coyle

TradingView scripts

Some of the indicators and strategies I've built along the way are published on my TradingView profile.