Sampling and Aliasing Distortion in Timeframe-Based Price Analysis
Summary
The article examines how market quotes become discrete tick observations and timeframe bars, arguing that sparse sampling can distort the underlying price signal. It applies sampling theory and aliasing as a conceptual framework: when a series is sampled too slowly relative to its frequency content, spectral components overlap, so the sampled data no longer faithfully represent the original. The author notes that tick arrival intervals vary and that the market signal’s true frequency range is unknown, making precise distortion estimates difficult.
As an illustration, the article compares low-pass filtering of GBPUSD M5 quotes with a thinned series treated as M75 data, adjusting the filter period to preserve its cutoff. The resulting curves differ, which the author presents as evidence that resampling may alter analysis. This is a demonstration, not a quantitative validation of the assumed signal model; the discussion relies on simplifying assumptions and acknowledges gaps, skipped bars, and platform time-series construction as additional complications. It recommends using the smallest available timeframe for mathematical analysis and cautions that OHLC values may add uncertain information beyond closes.
Key ideas
- Variable tick arrival and timeframe aggregation make the sampling process difficult to characterize precisely.
- Sampling below the rate needed to represent a signal can cause aliasing and distort analytical results.
- A comparison of filtered GBPUSD M5 data and a thinned M75 series illustrates differences after resampling.
- The demonstration depends on assumptions about the signal and does not quantify distortion generally.
- The article favors smaller timeframes for analysis and warns that bar OHLC values may carry timing uncertainty.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.