The Marketplace
Strategies in TradeBricks are portable JSON specs — pure data, no code. That makes them safe to publish, browse, and fork. The Market is where the community shares them.
Strategies are portable specs
When you publish a strategy, what travels is the same spec the Lab compiles your graph into: a list of nodes (each a brick id plus its params), a list of edges (which port wires to which), and a little metadata. That’s it.
{
"meta": { "name": "RSI reversion", "version": 1 },
"nodes": [
{ "id": "es", "brick": "data_source", "params": { "source": "ES 5m" } },
{ "id": "rsi", "brick": "rsi", "params": { "length": 14 } },
...
],
"edges": [
{ "source": "es", "sourcePort": "close", "target": "rsi", "targetPort": "x" },
...
]
}A spec contains no code — only brick ids and numbers. Loading someone’s strategy can never execute anything on your machine; the engine just looks up known bricks and wires them.
This is true of every brick, not just one. The whole catalog is a closed whitelist of pure evaluators with no eval or exec anywhere, so the most a shared strategy can ever do is compute indicators from that catalog. Even the formula brick — the one place you type free-form math — stays inside it: its expr is parsed to an AST and run by a restricted evaluator (numbers, named inputs, a fixed set of operators and functions only), never a real interpreter. See every brick is a whitelisted evaluator. That whole-catalog guarantee is what makes a public marketplace of arbitrary strategies safe to use.
Publishing
Built something you’re proud of? Publish it so others can learn from and build on it.
- Open the strategy in the Lab and give it a clear name and description.
- Publish. Your graph is serialized to the spec above and listed on the Market under your name.
- It appears as a card showing the bricks it uses, plus its fork and like counts.
A great-looking equity curve on Dev isn’t an edge. If you share a strategy, share what its holdout and overfitting guards actually said. The community is better served by an honest “interesting but unproven” than by an overfit curve dressed up as a winner. See Backtesting.
Browsing
The Market shows published strategies as a grid of cards. Each card surfaces:
- The strategy name and its author.
- A short description of the idea.
- The bricks it uses, as tags — a quick read on its category and complexity.
- Counts for bricks (▦), forks (⑂), and likes (♥).
Forking
Forking is the heart of the Market. Found a strategy you want to understand or improve? Fork it into your own Lab and it becomes a fully editable copy — same graph, now yours to tweak, re-backtest, and re-publish.
Why forking works so well here
- It’s a real copy, not a black box. Because the spec is open data, a fork lands as actual bricks on your canvas — you can see and change everything.
- It’s safe. No code comes with it, so forking can’t run anything unexpected.
- It teaches. Reverse-engineering a good strategy brick by brick is one of the fastest ways to learn the Lab.
- It compounds. Forks track their origin, so good ideas get iterated on in the open.
When you fork a strategy, the honest move is to re-run it yourself — especially on the holdout. A strategy that only ever looked good on its author’s Dev slice will tell you the truth the moment you backtest your fork.
A forked strategy may use live or alt-data bricks. Because strategies declare a capability (like options) rather than a specific broker, a fork runs unchanged for anyone with a matching connected provider — and the built-in Congress/SEC bricks work for everyone.
Related reading
- Build something worth publishing → The Lab
- Be honest about what to publish → Backtesting & overfitting
- Run a fork against your account → The Trader

