Time Series Prediction and Neural Networks

Discussion in 'Strategy Building' started by Optionpro007, Nov 3, 2006.

  1. Yesterday a guy posted market turns in the ES before the market open as per pic enclosed. I thought it was pretty incredible.

    Hope we can discuss this in further detail.

    <img src=http://www.elitetrader.com/vb/attachment.php?s=&postid=1253082>
  2. and the accompanying chart.

    <img src=http://www.elitetrader.com/vb/attachment.php?s=&postid=1253088>
  3. wow.. thanks for sharing

    didnt even know about that post.

    I would love to hear more about its theorys...

  4. thanks. is prem a chart style indicator? never heard of it, or jint.

    the fact that there was a forecast to start with was pretty incredible.

    weren't the results good ? I believe the poster had racked in 7 or 8 points profit per contract around 2pm.

    unfortunately all of his posts were deleted yesterday by what seems to have been a mistake of identity with another poster.

  5. I'm sure you have heard of PREM!

    It's the premium that futures are trading to cash. So no, it's not a chart style indicator!

    Depending on your charting package/data vendor you can get access to it as a symbol e.g. $PREM.


    The above two sites are oriented around arbitrage program trading when premium gets out of line with fair value.

  6. spinn


    That is the way I m trying to trade. I have 7 charts and only one of them shows the price.

    And the only reason that one chart shows the price is so I see the previous highs and lows and sometimes the bands.

    I could trade without seeing the price at all.....and would probably do better. The moving price is the money in your account, it takes the focus away from what is truly important.

    The only thing that really matters are the cycles of depth of market participation.
  7. Thanks mo. You da man. :)

    Enclosed is an esignal chart for $prem. Does it look accurate to you guys ?

  8. I am not a daytrader, but this subject has really awoken my interest.

    This is what I have been able to conclude so far.

    1.- Time prediction with precision as it was presented by 2cts only works for the next 3 to 4 market hours. Meaning one can predict at the close what will happen the next day till about 2pm. then after midday can one predict the rest of the day.

    2.- Information needed to make the prediction is taken from what has happened during the last 3 to 4 hours. all other information is meaningless.

    is this correct?

    #10     Nov 3, 2006