Skip to content
Strata
How it works

From an idea to a sized allocation — without a thousand browser tabs.

Strata’s workflow is one loop: idea → build → backtest → analyze → allocate. Scroll through each stage.

Step 01
Idea

Start from the question, not the code.

Most strategies die because the idea was wrong, not because the code was wrong. Strata starts you with the question a good portfolio manager would ask: what edge are you trying to capture — what session, what timeframe, what's your hypothesis for why this works? An AI assistant that pressure-tests the idea with you is coming soon.

Prompt
// what edge?
When ES is trending on the daily and pulls back to the 15m EMA during the NY session, fade the pullback into the trend.
Assistant: What defines "trending on the daily" — ADX, price-vs-MA, or higher-highs?
Step 02
Build

Indicators or Python — your call.

Drag in standard signals (EMA cross, ATR breakout, vol regime filter). Or open the Python editor and write it yourself with full libraries. Describing it in English and getting an AI-drafted first version is on the roadmap. The output is the same: a strategy object you can backtest.

Strategy
entryema(close, 21).cross_below(close) & adx(14) > 22
exittrail_atr(2.0) | session_end('NY')
filterregime.daily.trend == 'up'
sizingrisk_per_trade(0.5%)
Step 03
Backtest

Intraday realism, no curve-fitting safety net.

Run against 1-minute CME data from Databento with realistic slippage, commission, and session-aware fills. Walk-forward windows are on by default — you can turn them off, but Strata won't pretend an in-sample 4.0 Sharpe means anything. Monte Carlo on the trade-list when you want to stress the equity curve.

Backtest · ES · 2020–2025
CAGR
9.4%
Sharpe
1.42
Max DD
11.2%
Win rate
54%
Step 04
Analyze

How does this play with your other strategies?

This is the step every other tool skips, and it's coming soon to Strata: a correlation matrix that maps the new strategy against everything else in your book — correlation, return concentration, drawdown overlap. Sometimes the strategy that looks best in isolation is the one you shouldn't add.

Correlation matrix
ES Pull
NQ Vol
CL Rev
ZN Mom
ES Pull
1.00
0.62
-0.18
0.04
NQ Vol
0.62
1.00
-0.22
0.11
CL Rev
-0.18
-0.22
1.00
-0.05
ZN Mom
0.04
0.11
-0.05
1.00
Step 05
Allocate

Size it for the account that's actually trading it.

Risk-based sizing across your book. If you're trading a funded account, Strata simulates the daily-loss and trailing-drawdown rules so you know whether the strategy survives them. Broker bridges for routing the sized allocation live are on the roadmap — today you run the exported Pine on your platform.

Proposed sizing · $150K PA
ES Pullback35%
NQ Vol Breakout25%
CL Mean Reversion20%
ZN Momentum20%
Sim daily-loss limit: $2,200 (within $3,000 PA cap) · trailing DD: passes

Want to run the loop?

The platform is live. Spin up an account in under a minute — your first three backtests run on us.