Automatic Elliott wave counting: prepare the input before the run
A practical sequence for checking chart data, selecting the range and preserving the result of an automatic calculation.
Published: (UTC+3)
Русский · English · Deutsch · Español · Français · Italiano · 繁體中文 · 日本語 · 한국어
Automatic counting still begins with a human choice of data and boundaries. An algorithm cannot make two different exports into the same input. Complete the data checks first so that a surprising result can be investigated rather than merely rerun.
Separate data preparation from interpretation
OHLC means the opening, highest, lowest and closing price of a price bar. Check that your file represents the intended instrument and timeframe. A transaction history is a different dataset even if it is also saved as CSV.
Preserve the downloaded file. If you make a cleaned working copy, describe every change. Removing a row, rounding a price or changing the timestamp format changes the input and needs a reason.
Check chronology and price relationships
Look for duplicate timestamps, an unexpected order and missing periods that need explanation. A market closure and a broken export are different causes of an aphigher-degree wave gap; do not insert invented price bars to make the spacing look regular.
For each ordinary price bar, the high must not be below its open or close, and the low must not be above them. Finding an inconsistent row is a reason to inspect the source, not evidence that a particular importer will automatically repair it.
Choose and record the working range
Keep the file coverage and the selected counting range in separate fields. Note the intended origin and the last price bar included. A changed left boundary can give the calculation a different structural problem.
Run the calculation only after these choices are explicit. Save the result alongside the input record so that an alternative can later be compared on the same conditions.
Review the output as a candidate
Check the proposed structure, its internal subdivisions and its rejection conditions. Where the available price bars do not resolve a point, record that uncertainty rather than filling it with a preferred label.
The practical benefit of automation is a repeatable research step. It does not make the resulting wave interpretation a guaranteed forecast or an instruction to open a trade.
A wave interpretation changed after you shortened the file. Did the same test fail?
The input changed, so it is a different test. Keep both files and compare their time coverage and selected ranges before judging the difference.