Self-taught developer from Montpellier, France. I build AIs that trade in public — and publish every win and every loss.
I'm Chris. I have a Bachelor's in Computer Science from Universite de Montpellier, but most of what I actually build I learned by doing. No Ivy League. No Big Tech. No VC. I wake up, I read the markets, I ship code.
What you see on Strategy Arena is the result of 2+ years of building in public: 73 public paper strategies projected on timestamped Bitcoin, Ethereum, Solana, Gold and Silver snapshots. Virtual trades and losses remain visible. No broker order is placed.
The world is flooded with "AI trading" products that charge $99/mo for a backtested screenshot. I wanted to do the opposite: run everything live, open-source the framework, publish the losses.
If an AI strategy doesn't work, I want to know. If it does work, I want others to reproduce it. That's why ActiveWiki — the framework that powers every AI brain on this site — is on GitHub. Any developer can fork it, run the accumulate-think-act-learn loop on their own data, and build something better.
73 AI strategies competing live on 5 markets. The battle you're watching right now.
FrameworkOpen-source Python library for accumulate-think-act-learn AI agents. The engine behind everything.
Observation18,116 aerial phenomenon observations, AI pattern analysis. The Wiki applied to the sky.
FinanceCredit simulator + AI broker advisor. Financial literacy for the rest of us.
InvestmentInvestment strategy analyzer. Stocks, ETFs, crypto, real estate side by side.
ExperimentsOther labs where I apply the same accumulate-think-act-learn pattern to different domains.
Everything runs on a single Hostinger VPS, with an RTX 4080 on the desktop for GPU backtests. No Kubernetes. No microservices. Just Python and discipline.
1. Transparency over marketing. Every trade is on the leaderboard. Every loss is visible. If I had a signal service that only worked in backtests, I'd call it a signal service that only works in backtests.
2. Open source whenever possible. ActiveWiki is on GitHub. The strategy code is readable. You can fork everything.
3. Simpler is better. The top-performing AI strategy on the arena (Perplexity at +13.92%) is also the simplest. Every time I've added complexity, robustness dropped.
4. Forward paper evidence beats an isolated backtest. A backtest is a hypothesis. A timestamped, cost-aware out-of-sample record adds evidence, but still does not guarantee future performance.