Mechanical systems need more than pretty charts

A chart is not a procedure

A mechanical trading system is a fully specified set of entry, exit, and abstention rules that can be tested as one procedure under stated costs and constraints.

Accepted technical ideas often fail a precise computer test once transaction costs and slippage are included. That outcome is the basis for skepticism toward sold systems, system optimization, and popular indicators.

System optimization is fitting rule parameters to past market data so the procedure looks stronger in sample than it may be in live use.

Bias still needs timing

A fundamental model can indicate a market bias but typically supplies no timing for whether to act today versus later, so execution still needs a technical timing layer.

Fundamental models remain exposed to surprise events and unrecognized structural shifts, after which the same model can intensify the losing stance as price moves further against it.

Trend-oriented mechanical rules usually self-correct by exiting or reversing when price turns against a position, whereas a long-horizon fundamental stance has no built-in reversal.

What oscillator charts hide

A stochastic oscillator is a bounded oscillator built from ordered prices over a lookback, often used to mark overbought, oversold, crossover, or divergence conditions.

Specific stochastic schemes, including crossover triggers and variants that wait for overbought or oversold conditions, did not produce results competitive with other tested procedures.

Divergence lacks a single unambiguous mathematical definition, so it is easy to mark after the fact and hard to apply without many apparent setups that fail or never complete.

Restricting stochastic sells to overbought readings can multiply countertrend exits during a strong uptrend. A reverse filter that blocked sells while overbought helped only marginally in tests.

Chart crossovers near turning points conceal the false signals that accumulate under any executable oscillator rule, including more elaborate filters, persistence requirements, and nonstandard lookbacks.

Optimized examples are not proof

Optimized historical results are not a realistic indication of later performance, and brochure or article examples should not be treated as proof that a system or oscillator has value.

The intended lesson is not that every sold system, optimization routine, or oscillator is worthless, but that many straightforward uses fail and any candidate must be tested before it is traded.