Parameter optimization is the process of testing different strategy settings to find values that may improve a strategy’s backtest results.
In QuantPilot, you can ask the agent to review a strategy, identify important parameters, test different values, and suggest updated settings. These suggestions can help you compare versions, but they should always be reviewed before use.
What parameter optimization means
A strategy parameter is a setting that affects how a strategy behaves.
For example, parameters may include:
moving average length
RSI length
RSI entry or exit levels
stop loss percentage
take profit percentage
trailing settings
volatility filter settings
entry or exit thresholds
timeframe settings, when supported
Parameter optimization means testing different values for these settings to see how the strategy behaves in a backtest.
The goal is not to find a perfect-looking result. The goal is to find settings that make sense, improve the strategy, and remain reasonable when compared with the original version.
How optimization works with the agent
You can ask the QuantPilot agent to optimize parameters inside a strategy chat.
The agent can help you:
identify which parameters can be optimized
run or compare backtests
test different parameter values
suggest updated settings
explain why certain parameters were changed
compare the original version with the optimized version
point out possible risks in the optimized result
The agent may suggest new values, but you should decide whether the changes make sense for your strategy.
How to ask QuantPilot to optimize parameters
When asking for optimization, be specific.
A good request should include:
the market or pair you want to test
the timeframe
the strategy idea or existing strategy
which parameters you want to optimize
what you want to prioritize
whether you want conservative or aggressive changes
whether you want the original and optimized versions compared
For example:
instead of asking: Optimize this strategy.
Ask: Optimize this BTC strategy on the 1h timeframe. Test the EMA lengths, RSI threshold, stop loss, and take profit. Compare the optimized version with the original version and explain what changed.
Example prompts
You can use prompts like these:
Optimize the main parameters of this strategy. Test reasonable values for the indicator lengths, stop loss, and take profit. Compare the original and optimized versions.
Review this strategy and identify which parameters are worth optimizing. Do not change the core strategy logic unless needed.
Optimize this strategy, but avoid overfitting. I want settings that perform reasonably across different market periods, not only one perfect backtest.
Run a backtest with the current settings first. Then optimize the parameters and show a comparison between the original and optimized results.
Suggest a more conservative version of this strategy. Focus on reducing drawdown, even if total return becomes lower.
Optimize the parameters and explain which changes had the biggest effect on the result.
What users should review before accepting changes
Before accepting optimized parameters, review the changes carefully.
Check:
which parameters were changed
whether the new values are realistic
whether the strategy logic still matches your original idea
whether risk settings changed
whether the backtest period is long enough
whether the result depends on one unusually good period
whether fees, slippage, funding, or other assumptions matter
whether the optimized version has a reasonable number of trades
whether drawdown increased
whether the improvement is meaningful or only minor
Do not accept optimized parameters only because one metric improved.
Optimization vs overfitting
Optimization and overfitting are not the same.
Optimization is useful when it tests reasonable parameter values and helps improve a strategy while keeping the logic understandable.
Overfitting happens when a strategy is adjusted too closely to one specific backtest period. An overfitted strategy may look strong in historical results but fail when market conditions change.
A strategy may be overfitted if:
it only works on one exact date range
small parameter changes make results much worse
the strategy has too many optimized inputs
one or two trades create most of the profit
the settings look unrealistic or too precise
the strategy was repeatedly adjusted until one backtest looked good
A better backtest does not always mean a better strategy.
Why you should compare versions and backtests
Always compare the original strategy with the optimized version.
A comparison helps you see whether the optimization actually improved the strategy or only changed the result in one narrow way.
When comparing versions, review:
total return
drawdown
number of trades
win rate
average trade result
risk settings
behavior during losing periods
consistency across different time periods
The optimized version should make sense as a strategy, not only as a backtest result.
What results to treat cautiously
Treat optimization results cautiously when:
the optimized version only improves one metric
the strategy becomes much riskier
drawdown increases too much
the number of trades becomes very low
results depend on one short market period
parameters become too specific
the agent changes the strategy logic more than expected
the improvement disappears in another backtest period
If something looks too good, ask QuantPilot to explain why and test it again under different conditions.
Optimization is not a guarantee
Parameter optimization does not guarantee future performance.
Optimized settings are based on testing, historical data, and assumptions. Real market results can differ because of volatility, liquidity, fees, slippage, funding, execution delays, and changing market conditions.
Use optimization as a research and comparison tool, not as proof that a strategy will work in the future.
What to do after optimization
After QuantPilot suggests optimized parameters, you can:
review what changed
ask the agent to explain the changes
compare the original and optimized versions
run additional backtests
ask for robustness validation
adjust the parameters manually
save the version you want to keep
Only use optimized parameters if you understand the changes and accept the risks.