Stop making changes, go back to your last known-good baseline, and change exactly one thing at a time from there. Most setups get worse not because the driver made a bad decision, but because too many things changed at once, so it became impossible to tell which change caused which result.
Why this happens
In pursuit of a perfect setup, it’s tempting to make several changes between runs: springs, anti-roll bars, and alignment all at once. Each adjustment changes how the car responds, which means you’re not just testing a setup, you’re relearning the car’s limits every single run. Time that should go toward refining your driving instead goes toward adapting to new variables you introduced yourself.
How to recover
- Return to your last baseline. If you have a setup you know was reasonable, go back to it completely rather than trying to patch the current mess.
- Change one variable at a time. Make a single adjustment, run enough laps to feel its effect clearly, then decide whether to keep it before touching anything else.
- Diagnose by phase before changing anything. Identify which phase of the corner has a problem, and whether it’s understeer or oversteer, before picking a component to adjust. Guessing at components without this step is exactly how setups spiral.
- Write down what you changed and why. A simple log prevents you from re-testing the same failed change twice or losing track of what’s actually been tried.
The bigger picture
A stable baseline with small, deliberate adjustments will always beat a setup that’s been changed a dozen times in an afternoon. Predictability and familiarity are worth more than a theoretical maximum you can’t consistently drive.
This is the same phase-and-balance framework we built into onRails: identify the corner phase, identify the grip imbalance, apply the smallest effective change, one at a time. If you’re losing track of what you’ve already tried, onRails keeps that process disciplined for you.
Related questions
How many changes is too many at once?
More than one. Changing a single component per test run is the only way to know for certain what caused what you’re feeling.
What if I don’t remember my last good baseline?
Start over from a simple, symmetrical baseline setup rather than trying to reverse-engineer your way back through a chain of changes. It’s faster and more reliable.
Is it ever okay to change several things at once?
Only once you deeply understand the car and are making small, related refinements in the same direction. Early in the process, and any time you’re troubleshooting a problem, stick to one change at a time.