Maybe we should implement some more robist measure of association. For example Kendall's thau. Basicaly t is measure of. Oncordance/ disconcordance between two rank ordering variables. So, if we make assumption that all our indicator can rank order (favoirable, less favourable, not favourable) we can calculate Kendal's thau instead. Its uch more robust measure and it doesn't react on extremes quickly.
Thoughts? On Aug 5, 3:27 am, Eugene Kononov <[email protected]> wrote: > > Another bit of advice: for large amounts of accumulation, integer values > > may be more predictable, as they don't accumulate weird bits of noise from > > floating point errors. I have not verified this actually happens in the > > real world. > > Yes, it does. This entire formula is "notorious for its numerical > instability", as this article > mentions:http://en.wikipedia.org/wiki/Correlation_coefficient --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "JBookTrader" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jbooktrader?hl=en -~----------~----~----~----~------~----~------~--~---
