Reading Optimization Results

After optimization completes, the results show more than just the best parameters. They reveal the landscape of your strategy's parameter space.

Result Components

Parameter Surface

A visual representation of how different parameter values affect the target metric. Look for:

• Plateaus — Broad flat regions where many nearby parameter values produce similar results. These are the most robust zones.
• Sharp peaks — Narrow spikes where only one exact combination works. These are overfitted and unreliable.
• Valleys — Regions where the strategy fails. Understanding where it breaks is as important as where it works.

Stability Score

Each parameter cluster receives a stability score measuring:

• How consistent the metric values are within the region
• How large the region is (broader is more stable)
• How far the metrics drop at the edges

Higher stability scores indicate more robust parameter regions.

Top Results Table

The results table shows the best parameter combinations ranked by your target metric, along with:

• All performance metrics (return, Sharpe, drawdown, etc.)
• Stability score for the parameter region
• Trade count and win rate

Choosing Parameters

Do not simply pick the row with the highest return. Instead:

1. Look for clusters with high stability scores
2. Within those clusters, prefer the center of the plateau (not the edges)
3. Check that the trade count is sufficient (ideally 30+)
4. Verify that the drawdown is acceptable

A parameter set with slightly lower returns but higher stability will almost always perform better on unseen data.

Next Steps

• Walk-Forward Analysis — Test your chosen parameters out-of-sample
• Optimization Overview — The full validation pipeline