8/2/26
My SQX Authoring Toolkit is now a single installable Claude Code plugin. Describe a custom block, a random group, a strategy template or a whole build project in plain English — sqx-lab reads your own StrategyQuant install, generates the artifact, and validates it before you import. Free in the StrategyQuant Codebase.
Read More7/12/26
Generate 10,000 strategies and the best backtest will look brilliant whether or not any real edge exists. The multiple-testing view of strategy development — and the discipline that follows from it.
Read More7/12/26
An AI demo is a backtest: it looks brilliant on the cases it was tuned on. Evals are the new backtests — the oracle-gated, out-of-sample discipline I borrowed from quant research to test my own agents.
Read More7/12/26
For months I've been building something bigger than any single indicator — a framework of AI agents that turn a plain-English idea into a finished StrategyQuant strategy, and a live oracle that loads every one into a running platform to prove it's real. Here's how it works, the numbers behind it, and where it's going.
Read More7/12/26
Scientists freeze their thresholds before they see the data — because humans rationalise. Inside the pre-registered validation gate every strategy population I generate must pass.
Read More7/12/26
Your best backtest's Sharpe ratio is inflated by every strategy you generated to find it. What the Deflated Sharpe Ratio corrects for — and how I actually run it on StrategyQuant databanks.
Read More7/12/26
The honest version of how I went from a misspelled Google search and a live futures desk to building StrategyQuant's AI — an expanded, up-to-date retelling of the interview StrategyQuant once ran with me.
Read More7/12/26
I'm StrategyQuant's AI architect, so this isn't a neutral review — it's an honest insider's tour of what StrategyQuant X genuinely does well, where it struggles, and who should (and shouldn't) buy it.
Read More7/12/26
A $1,290–$2,900 question. Here's the honest ROI on StrategyQuant — who genuinely gets their money's worth, and who should keep their wallet shut.
Read More7/12/26
Over the years I've published 135 custom indicators, databank columns, Monte Carlo tests and workflow tools to the StrategyQuant Codebase — all free. Here's the complete toolkit, organised by what each part is actually for.
Read More7/12/26
Starter, Professional or Ultimate? A clear, honest breakdown of what each StrategyQuant license includes, what it costs, and which one is actually right for you.
Read More9/16/25
Order blocks, fair value gaps and liquidity sweeps are usually taught as discretionary chart-reading. Here is how I think about Smart Money Concepts — and how to turn the parts that survive into objective, backtestable rules.
Read More8/11/25
A single backtest is just one lucky path through history. Here are the five Monte Carlo stress tests I use to separate a genuinely robust trading strategy from one that only looks good on paper.
Read More6/17/25
A fixed-period moving average treats a violent market and a sleepy one exactly the same. Adaptive indicators change their speed with conditions — a practical guide to VIDYA, VMA, VST and HalfTrend, and where each earns its place.
Read More
4/15/25
Meet StrategyQuant's divergence comparison blocks: what divergences are, how the blocks detect them, the key parameters to tune, and how to trade them in practice.
Read More4/10/25
Strategies rarely break at random — the market regime shifts underneath them. Four statistical tools (Kolmogorov–Smirnov, Wasserstein distance, CUSUM, and dynamic time warping) for detecting the change before your drawdown does.
Read More
1/20/25
We put adaptive blocks to the test—generating thousands of strategies with adaptive versus traditional blocks and comparing their performance statistically across several markets.
Read More
12/20/24
The IsGreater/IsLower Adaptive comparison block checks whether one indicator is greater than another—but only fires a signal if similar situations were profitable in the past.
Read More
11/15/24
Adaptive “Crosses Above/Below” blocks bring historical signal performance into your rules, confirming a crossover only when similar past signals proved statistically reliable.
Read More5/4/23
A fixed-distance stop ignores how much the market is actually moving. The ATR trailing stop scales your exit to volatility — how it works, how to set the period and multiplier, and when it earns its keep.
Read More11/22/22
The classic profit factor is easily flattered by a handful of lucky trades. Timothy Masters' internal profit factor measures a strategy's edge bar-by-bar, inside each position — a far harder metric to game.
Read More8/4/22
David Aronson's Evidence-Based Technical Analysis explains why so many backtested edges are illusions. Data-mining bias is the trap — and it matters enormously if, like me, you generate strategies in bulk.
Read More3/15/22
A research project across roughly 1.2 million generated FX strategies, and the four factors that actually separated the ones that survived true out-of-sample from the ones that only looked good in the backtest.
Read More2/28/22
Every edge decays eventually. The hard part is telling an ordinary drawdown from a strategy that has genuinely died — and deciding the rule that separates them before your money is on the line.
Read More8/25/21
The E-ratio strips a strategy down to its raw directional edge, before exits and costs. I tested whether it predicts true out-of-sample performance on EURUSD and the S&P 500 — here is what the data said.
Read More8/2/21
A genuine edge is rarely confined to a single symbol. Testing a strategy across many markets is one of the simplest and most brutal defences against overfitting — and why I look at the median, not the average.
Read More3/22/21
A point-and-click strategy generator gets you a long way. Where it ends, Python begins — databank analysis, decorrelation, machine learning and command-line automation in a real quant workflow.
Read More