Mar

28

Data, from anonymous

March 28, 2014 |

 I often look at the amount of past price action used to attempt to predict future price action.

Some things that are useful to ponder, in my opinion, are:

1. Is more past data really going to help to make the future prediction more accurate?
2. Should there be a balance between look-back period and forecast horizon?
3. How important is data accuracy (tick level to daily range)?
4. Should reference points & times be changed every second, minute and hour of a day?
5. Should the definition of 'big move' and 'small move' be a fixed thing or relative to the market's current level?

For me, it's NO, NO, VERY, YES & RELATIVE.

Go Well.

Leo Jia writes: 

In Schwager's book "Hedge Fund Market Wizards", Jaffray Woodriff addressed this in the following way. Any comments? It does have to do with what one is trying to get, doesn't it?

"Do you give the same weight to data from the 1980s as data from the 2000s?

Sometimes we give a little more weight to more recent data, but it is amazing how valuable older data still is. The stationarity of the patterns we have uncovered is amazing to me, as I would have expected predictive patterns in markets to change more over the longer term."

Larry Williams writes: 

As I see it we certainly cannot compare data from the old pit sessions to today's electronic markets.

And how do we handle Saturday trading in the real old days or that markets were close on election day…or in 1967 the markets were close on Wednesday… or there used to be a massively important bond report the goosed bonds on Thursday??

We need to understand what the data represents.


Comments

Name

Email

Website

Speak your mind

2 Comments so far

  1. Bill Wolfe on March 30, 2014 9:15 pm

    1. Is more past data really going to help to make the future prediction more accurate? Yes. It is absolutely imperative to find the beginning of the current wave.
    2. Should there be a balance between look-back period and forecast horizon? Yes. This is how one predicts the new target line.
    3. How important is data accuracy (tick level to daily range)? Of course, yes.
    4. Should reference points & times be changed every second, minute and hour of a day? Yes. On an as needed basis depending on the price/time relationship of the wave.
    5. Should the definition of ‘big move’ and ’small move’ be a fixed thing or relative to the market’s current level? No. Each wave should be traded as a separate instrument.

    Visit: www.wolfewave.com for examples.

  2. Bill Wolfe on March 30, 2014 9:18 pm

    1. Is more past data really going to help to make the future prediction more accurate? Yes. It is absolutely imperative to find the beginning of the current wave.
    2. Should there be a balance between look-back period and forecast horizon? Yes. This is how one predicts the new target line.
    3. How important is data accuracy (tick level to daily range)? Of course, yes.
    4. Should reference points & times be changed every second, minute and hour of a day? Yes. On an as needed basis depending on the price/time relationship of the wave.
    5. Should the definition of ‘big move’ and ’small move’ be a fixed thing or relative to the market’s current level? No. Each wave should be traded as a separate instrument.

    Visit: www.wolfewave.com for examples.

Archives

Resources & Links

Search