Number of distinct behaviours in a basket
An approximation of how many separate movement patterns actually repeated in a set of coins.
The number of positions alone does not say how many different things happened in them. Six coins that moved identically historically describe one behaviour, not six. The measure is computed from the average mutual association in the basket. The higher it is, the more the results duplicate each other and the lower the value. The scale is natural: 1 means everything moved as a single position, while a value equal to the coin count means each went entirely its own way. The result is clamped between 1 and the number of coins, and with a single position it is not computed at all. It describes what already happened in a given window, not an assessment of a portfolio and not a hint on how to change it.
In plain words
It says how many genuinely different behaviours were hidden in a set of coins. Six positions that travelled together historically behaved like one. A value close to the coin count means each went its own way.
Example
A portfolio of 6 active positions with an average mutual association of 0.55 gets a value of 2.4. Historically, about two to three separate movement patterns repeated in it. Had those same 6 positions had an average association of 0.05, the value would come out close to 5.
Also known as: effective bets, efektywna liczba pozycji