Walk-forward system evaluation with a commodity channel index and chandelier exits

Evaluation as one procedure

Walk-forward evaluation fits parameter values on an in-sample window and then applies those same values to a later out-of-sample window. The in-sample window is the historical window used to fit or select parameter values. The out-of-sample window is the later window used only to apply already chosen parameter values.

Editorial interpretation: treat that sequence as one procedure. The protocol should decide whether a mean-excursion signal and a volatility-scaled stop still belong together after the parameters leave the fitting window.

How walk-forward evaluation works

Walk-forward analysis is a staged evaluation that fits parameter values on an earlier sample and then holds those values fixed on a later sample that was not used for fitting.

Backtesting is a two-stage check that develops a strategy on historical observations and then applies the same rules to later observations to see whether the behavior stays consistent.

Optimization shapes system rules so they fit the historical sample under study as closely as possible. Because that practice matches the sample as closely as possible, walk-forward evaluation keeps the later window free of a second fitting step.

The commodity channel index as a fixed reading

The commodity channel index is a price-momentum reading that measures how far the current price has traveled from its mean over a stated lookback. It reports how far price has moved away from its mean.

A discretionary approach interprets indicator signals with personal judgment rather than a fixed procedure. Discretionary judgment means interpreting an indicator signal by personal decision rather than by a fixed, testable rule set.

Editorial interpretation: writing the reading as a fixed rule lets it enter the walk-forward procedure with the rest of the system, instead of remaining a matter of discretion.

The chandelier exit as a fixed stop

A chandelier exit is a volatility-scaled stop placed from the peak high or the last close, using a chosen multiple of average true range to bound remaining exposure. The distance is a multiple of average true range, measured from either the highest high or the close.

Average true range is a volatility measure used to set chandelier stop distance from a high or a close. A stop-loss is a risk action that closes a trade to prevent additional decline in value.

Editorial interpretation: the chosen multiple and the choice of high or close stay with the already chosen parameter values when the later window begins.

Judging the pair after fitting ends

After the in-sample window, the later sample is used only to apply already chosen parameter values. Those values are held fixed.

Editorial interpretation: the later window is where the designer learns whether the commodity channel index rule and the chandelier exit still belong together as one procedure, not whether each piece can be reshaped on its own.