Write the idea.
Use the Pine Script editor, built-in examples, and reference docs to define your strategy and its inputs.
Test trading ideas
across a range.
Run thousands of historical backtests across the inputs of a supported Pine Script strategy. Compare the results and inspect the trades behind them.
No credit card. Bring a supported Pine Script strategy, or start with an example.
Highest single result
–
- Result
- –
- Neighbors, median
- –
- Neighbors above 0%
- –
Best-supported region
–
- Region median
- –
- Neighbors, median
- –
- Neighbors above 0%
- –
Hover to check a setting's neighbors. Click one to replay its trades below.
Open a result. Inspect every trade.
Price, strategy plots, entry and exit markers, and the full trade list in one view.
- Trades
- –
- Combined, no fees
- –
- Winning trades
- –
- Buy and hold
- –
Real Binance BTC/USDT candles, July 20 – August 10, 2025. Crossovers are replayed in your browser with next-candle fills and no fees; TrainBard runs the full backtest.
Test the range.
Enter exact values or ranges such as 10_50:5. See the combination count before running the sweep.
Follow the evidence.
Check the batch, compare nearby settings, and open individual results to see what produced the ranking.
Analyze the whole sweep.
Analyze a completed sweep or upload results to investigate. Start with batch health and coverage, then look at parameter neighborhoods, effects, and competing objectives.
Strong results need neighbors.
Regions groups nearby settings and shows their support, pass rate, and outcomes against a quality bar. See whether a promising result sits in a broader region worth testing again.
- All settings
- –
- Best-supported region
- –
- Around the highest result
- –
Settings at or above the bar are blue.
Keep the tradeoffs visible.
Trade-offs compares non-dominated candidates across selected objectives, such as return, drawdown, and win rate. Inspect the Pareto frontier without reducing every choice to one score.
- Non-dominated
- –
- Highest result, drawdown
- –
- Region center, drawdown
- –
Each dot is one setting from the same sweep.
Look closer with Drivers. Inspect main effects and interaction heatmaps to see which inputs deserve a narrower test.
–
–
Change your perspective in Explore. Choose axes, color, and opacity to examine the result set in two or three dimensions.
These views describe the history you tested. Local consistency and a Pareto frontier do not establish future performance.
Build a Study
across sweeps.
Study brings related sweeps together across markets, candle sizes, and periods. Compare candidate neighborhoods, check their coverage, and see where support thins out as the context changes.
Move between the evidence across sweeps and the details of one batch. Export a study when you want to carry the work with you.
Study is available on Pro 5x.
Inside the Study workspace–
–
- Region median, →
- –
- Highest result, →
- –
Outlines mark the same settings on both sweeps.
The same SMA parameter grid, compared on 2-hour and 4-hour candles. Historical evidence across two sweeps.
Run the next sweep
through MCP.
Connect a compatible AI assistant through the Model Context Protocol. Give it your supported Pine Script strategy and test conditions, then work with the returned evidence.
- Submit a parameter sweep
- Specify the market, candle size, period, fees, and ranges.
- Inspect a particular result
- Retrieve performance metrics and the full trade list for one parameter set.
- Keep the outputs you need
- Have your assistant save results to your files. Your plan limits and weekly research allowance apply.
Choose what
to keep.
Running a backtest does not automatically save your strategy or results. Save selected work to your private Library, or export it to your own files.
- Strategies and versions
- Save a strategy explicitly and keep versions as the idea develops.
- Tests and studies
- Keep selected tests and related research together for later inspection.
- Portable exports
- Export CSV, JSON, and research packs. MCP results remain available for up to 30 minutes after completion; your client may retain its own copies.
Beyond the backtest.
Live (Beta) · Paper trading
Follow a supported strategy on live market data with simulated funds. Inspect its runs, holdings, and recorded activity as new candles arrive.
Beta access depends on your account. No real orders are sent.
Start with a question.
Scale the test when you need to.
The free plan gives you room to try the workflow. Larger plans add more combinations, longer historical ranges, and higher throughput.
Free
$0
- 500 combinations per batch
- 1 year date range
- Standard queue
Standard
$10/mo
- 10,000 combinations per batch
- 2 year date range
- Priority queue
Pro 2x
$25/mo
- 20,000 combinations per batch
- 4 year date range
- 2x throughput
Pro 5x
$100/mo
- 100,000 combinations per batch
- Unlimited date range
- 5x throughput · Study
A few useful details.
Compatibility, storage, and the limits of what a backtest can tell you.
Can I use my TradingView Pine Script?
TrainBard supports a defined subset of Pine Script. Paste a supported strategy or start with a built-in example, and check the compatibility reference for accepted functions and behavior. Full TradingView compatibility is not claimed.
Why test a range of settings?
One parameter set can look good by accident. A sweep lets you compare nearby settings, inspect sensitivity, and see how return changes alongside drawdown and trading activity. It is a way to investigate historical behavior, not proof that a strategy will work in the future.
Are my strategies and results saved?
Only when you explicitly save them. Ordinary backtests stay live in the workspace; selected strategies, tests, and studies can be kept in your private Library. You can also export your work. MCP results are temporarily retrievable for up to 30 minutes after completion, and your MCP client has its own storage policies. See the security notes.
What do I need for MCP access?
Use a remote MCP client that supports custom authorization headers and create a revocable token in TrainBard under Settings → MCP. OAuth-only clients cannot connect yet. Read the setup guide for supported tasks and account limits.
Does Live (Beta) trade real funds?
Live (Beta) currently supports paper trading with live market data and simulated funds. No real orders are sent. Access depends on your account and the beta rollout.
Bring an idea.
See what the tests reveal.
Start free with a supported Pine Script strategy or an example from the editor.