# Multiple Conundra, from Kim Zussman

March 6, 2010 |

Here is the trouble with multiple hypothesis problem:

Say one wants to test "what happens tomorrow if today is down". You take the mean of days after yesterday was up, and compare this to all days. But you could have asked "what happens tomorrow if today is up". So that is a hypothesis you could have thought of too.

what happens tomorrow if today is down, and bonds are up/down?
what happens tomorrow if today is down, and \$/eur is up/down?
what happens tomorrow if today is down, and last year was up/down big?
what happens tomorrow if today is down, and there is a Republican in office (or not) what happens tomorrow if today is down, and the US is at war?

i.e., any hypothesis I can come up with might have been a different one (and would have been, in the case of a capable trader), and any hypothesis I come up with can also be conditioned by other concomitant variables.

## Ken Drees replies:

This is exactly why I have problems with a lot of these ideas or theories. I am not saying to not explore or to find relationships, but the future is what it is and that random unknown can match, not match, match slightly, or match inversely whatever likely input one is testing. Everyday is different, yet each day is lived and used and imprinted with the similar human behavior from the previous day or a patterned day and the final resulting price changes are the end net result.

We are all fishing. Some have the latest rods and tackle, some use an old pole and gut instinct. Flexibility of techniques is important — sometimes the fish are hitting on blue, some days its green and sometimes the wind is coming from a different direction than normal and they don't bite much on anything.

I had to unlearn a lot from my h**d** broker who had me listen to Peter Eliades every day trying to pick a top in the Dow in 1995 when the market just went up and up and up. I mean every number imaginable was sliced and diced and arranged to show that the market had to go down, every Elliott Wave turned into a triangle and then another triangle and so on.

Even if you get a "go" sign on an idea and you hit, did it really work because your idea said so or was it some other factor and you got lucky? Got a match?