Trying to rank the distribution of events over time

15 Views Asked by At

I am trying to figure a way to rank a set of financial trades done across multiple instruments.

This takes place across of a period of 1 year. For each instrument, there is a list of trades that are either positive or negative.

I am trying to find a way to assign a score to each instrument, where:

  • It is desirable that positive and negative trades happen close to each other and may alternate.
  • It is undesirable to have a cluster of positive trades and a cluster of negative trades.

For example, if I was ranking the result of each month:

  • +-+-+-++--++ would be desirable
  • +++++-----++ wouldn't be desirable because of a large sequence of negative results
  • +++++++--+++ would be ok, despite large monotonous chunks

I would like to determine a score, as a single number, that will be used to rank each instrument.

What would be a good approach for this?