LC Simulator: using “Number of Steps” and “Run Count” values for enhanced Model robustness

Description
LC Simulator can suggest a minimum number of experiments needed for modeling (for example, for gradient optimization). However, this minimum guarantees only that the model can be built mathematically; more runs are usually required for robust and reliable optimization.

Solution

1. In LC Simulator’s optimization panel, note the “Required number of runs” indicated by the software.
2. Treat this number as a lower bound:
   • It is the smallest number that allows the model to be calculated without errors.
   • It is not necessarily sufficient for a high‑quality, robust model.
3. Plan additional experiments beyond the minimum:
   • Increase the number of runs such that:
     • Key regions of the gradient space are sampled.
     • Critical peaks are properly modeled.
4. Use the Number of Steps or plan settings to distribute experiments logically across the gradient domain (for example, varying starting and ending %B, or gradient time).
5. After collecting data:
   • Rebuild the model and check residuals, predicted vs experimental retention times, and resolution of critical pairs.
6. Adjust the experimental design if:
   • Model residuals are large.
   • Important peaks are poorly described.