Jason Cawley "NKS Time Series"

I stepped into this talk half-way through, so I think I missed the key points, but it was interesting.

The NKS book included a single model for market behaviour (rule 90 if I remember), and in this talk Jason Cawley presented a large number of experiments that show further evidence of the relevance of simple programs to explaining time series behaviour of market and economic data.

One point he made was that the models he presented capture some qualitative features of the real world data like "crashing" behaviour at different scales. Another theme was the general goal of NKS model fitting.

He has looked at Markov chain's from an NKS point of view (something not covered much in the book). I think the NKS type behaviour in such systems he shows is quite surprising. He has a novel and interesting analysis of it too.