The post on the site about introducing wolves into Yellowstone and seeing changes in river flow a while back got me thinking. It's like adding one or more random but significant variables to a relatively stable process and not expecting any change. It's almost the negative definition of insanity, doing same thing and expecting different results… I see it in manufacturing all the time. It seems to affect hubristic management the most. And seems like politicians do both all the time too.

Example: Stable system + some random change(s) = *total surprise it blows up*


Stable system + system left alone = *total surprise it doesn't blow up*

The safest method I've found is introducing small variables for a short time, recording results, then removing variables, and recording the results. Repeat two more times. If ALL results are similar, then the variable + system can be characterized. Just cranking a knob or dial and hoping is roulette with nearly loaded gun.

A nice method is Evolutionary Operation or EVOP. I bought this software, but couldn't get it to work any better than manual method. I just think it might have a viable use in trading schema.





Speak your mind


Resources & Links