Evaluating weekday close direction with a counted baseline

Direction labels for each weekday close

Ordered daily closes are first categorized by direction and then treated as weekday strings so a later close can be compared with the counted baseline at that sequence-fork.

Monday’s close is always coded as a continuation-label, whether that close is higher or lower than Friday’s close. Price-change-categorization maps each later daily close to a same-direction or opposite-direction label relative to the Friday-to-Monday move.

Each later weekday close receives a continuation-label if it matches the Friday-to-Monday direction. It receives a reversal-label if it moves opposite the Friday-to-Monday direction.

A branching grid of weekday strings

Weekday close strings are arranged on a branching grid so each new close is read as a sequence-fork from the preceding sequence. Pattern-recognition enumerates those weekday close strings and compares a later close with the count accumulated at that sequence-fork.

One illustrated grid summarized weekly cattle close sequences. At each sequence-fork, remaining observations are split by the next close.

A counted baseline after each sequence-fork

A runs-test partitions ordered closes into continuation stretches and reversal breaks so a later session can be scored against the sequence already observed.

Replication is a tally of closes under each sequence label, expressed as a share of the observations that remain after the previous sequence-fork.

The same encoding in another market

A related study asked whether the session after any five-day close sequence finished lower and allowed sequences to begin with a down close.

The same close-direction encoding can be rebuilt for another market by counting sequence members rather than by fitting a priced rule set.