1988issue C081-2
Empirical price-change counts versus borrowed statistics
Counting how often prices actually move can bound a stay-or-exit decision. That empirical habit is not the same as dressing a method in scientific vocabulary the counts never earned.
- One technique estimates whether a price trend is likely to continue by counting how often moves of different sizes have occurred.
- Those counts are treated as empirical distributions that need not match any named mathematical family, and the same tables can ask how plausible further favorable movement remains after a given move has already appeared.
- Frequency tables were presented as a numerical check on whether a live trade still matches the historical pattern, while building and refreshing the arrays was limited by time, discipline, data, and programming.
- Marketing that borrows scientific-sounding labels without explaining the idea is treated as a reason for skepticism, and a shared glossary was proposed so technical language can be checked.
Counting how prices actually move
One described technique estimates whether a price trend is likely to continue by counting how often moves of different sizes have occurred. Those counts are treated as empirical distributions whose shape need not match any standard, named mathematical family.
A price-change histogram is a tally of how often price moves of different sizes occur over a chosen sampling interval and lookback. The empirical distribution is a frequency shape taken from observed price changes rather than from a named textbook family.
What further movement remains plausible
The same style of distribution is illustrated for asking, once a position already shows a given number of points, how plausible additional favorable movement remains. That reading is quantile analysis: an empirical move-size distribution used to ask what further movement remains plausible once a given amount has already appeared.
Frequency tables of this kind are presented as a way to judge, in numbers, whether a live trade is behaving as the historical pattern would lead one to expect. Expected value means using those frequencies to keep a stay-or-exit judgment bounded by how often similar situations historically succeeded or failed.
Time, discipline, data, and programming
Building and refreshing the needed arrays is described as limited by time, discipline, data, and programming, with then-current personal-computer software poorly suited to maintaining many of them. In that setting, descriptive distributions were kept on only four instruments.
When scientific language is borrowed
Marketing that borrows scientific-sounding labels without explaining the idea, citing other work, or revealing a method is treated as a reason for skepticism. The ordinary scientific word for a random process is called out as a poor name for a popular oscillator, because that meaning does not describe the indicator.
A misapplied statistical term is a scientific-sounding label attached to a trading tool when the ordinary technical meaning does not describe the tool. A shared glossary, flagged in articles with a dagger, is proposed so technical language can be checked against agreed definitions, and readers are invited to challenge entries that look wrong.
All readings on this track · 7 readings
- 1986Constructing price-change density histograms and symmetry tests
- 1986Calendar-locked averaging of price-change histogram categories
- 1986Chi-square test for clustered price-change histograms
- 1988Empirical price-change counts versus borrowed statistics
- 1991Constructing tick engines to match price-change histograms
- 1996Constructing a price occupancy histogram and a smoothed mobility reading
- 2003Treat an opening gap as a completed session event