Statefulness in Algos
Mind that many trading algorithms are stateful, meaning that they maintain an internal state that is updated as new data becomes available. This internal state can include things like positions(qty), entry time, exit/entry price, indicators...

or other variables that affect trading decisions.
Because these algorithms are path dependent, the performance of the algorithm can be affected by the starting point of the backtest. If the algorithm's performance varies significantly depending on the starting point of the...

backtest, it may indicate that the algorithm is overfitting to the specific path of the historical data. This can be a sign of curve fitting, where the algorithm is too tightly optimized to the historical data and may not perform as well on new, **unseen data**.

To avoid curve fitting and ensure that the algorithm is robust to different starting points and market conditions, it is important to perform sensitivity analysis and backtest the algorithm on a variety of different starting points and market conditions.

This can/may help identify any weaknesses or overfitting in the algorithm and improve its overall performance and robustness.