Predicting Trending Days using Pivot Range Width

Discussion in 'Strategy Development' started by FaderTrader, Mar 28, 2006.

  1. I would be interested to hear anyone’s thoughts on how the width of the Daily Pivot Range can act as a predictor of trending days.

    Seems to me that volatility mean reversion is alive and well when it comes to the Daily Pivot Range. As you can see from the chart attached, days with unusually narrow pivot ranges were followed by days with unusually wide pivot ranges, and vice versa. I struggle mostly with exits, not entries – so it seems meaningful to me.
  2. Hi,
    doesn`t this simply mean that such a day did close in the middle of it`s range? To much math for so easy facts...But before i checked the simplicity, i did some test with an Openrange Breakout system, profiting most from trenddays, couldn`t find any useful. This condition tells nothing about volatility, cause the bar can be quite large or small, it gives only a definition that the close was ~ in the middle of the range. If you look for volatilityreversals, try Hist. volatility(3/10) or ATR(4-8) or NR4, NR7 stuff.

    greetings Michael
  3. Sorry for not being clear. I suppose the clearest application of this is to say that when the Pivot Range width tightens up relative to its 30-day average, then it's likely that the high of the day will be far away from the low - since today's pivot range measures where most of yesterday's trades went off, which is, IMHO, a decent proxy for measuring yesterday's volatility.
  4. Here some example why this stuff is useless...

    Low = 60
    High = 63

    range = 3 Point

    Daily Pivot Differential = 0

    Low = 60
    High = 70
    Close= 65

    range = 10 Point !

    Daily Pivot Differential = 0

    and so your Pivot Range width is also 0....

    You got 2 days, the last is 3 times larger, we expect this a very large widerange Bar in this market if we say the range of the first Bar with 3 points is more ~ the average daily range.

    So you got the same result, and only the information that the close was ~ (here exactly) in the middle of the bar.

    After a large widerange bar (trendday or here not) you get a higher Probability of a smaller day following with also a higher prob. to be a Z-day.

    You have a higher Prob. of a Trendday after 1-2 smaller Z-days. the size of the range is key, not so much where the close is...

  5. If you believe this working you can simply look for Doji`s, the candletraders believe a higher prob. of reversals after a Doji Bar....
  6. Not sure why you think that the Pivot Differential measures the width of the Pivot Range for tomorrow. It doesn't. It's just a number that is used to calculate the high and low of tomorrow's pivot range.

    Measuring the spread of that range relative to the 30-day average has nothing to do with the pivot differential.

    I think you need to rethink the math.

  7. I'm attaching a doc for your edification. Direct your attention to what happened in the following days:

    Day 4 had the largest trading range of the month. Betting that Day 5 would be a consolidation day, probably made you money since the pivot range going into that day was also the widest of the month.

    Day 13 had a fairly tight trading range and a pivot range going into Day 14 that was tight as well. You probably made money buying breakouts on Day 14.

    Day 23's data yeilded a tight pivot range going into Day 24. You probably made money buying breakouts on Day 24 as it yeilded the widest pivot range going into Day 25, which incidentally was a consolidation day with a pivot range going into day 26 that was well below average.
  8. Execuse me:

    You can simply look at a chart and get a good idea of where volatility compression has occurred. Look at the size of the bars. Look particularly at inside bars as a setup for increased volatility
    to come.

    Google or search (top right hand corner of home page) for the following subjects;

    NR4, NR7, Toby Crabel, Volatility Expansion, etc

    Good luck Newby

  9. Daily Pivot Price = (High+Low+Close)/3

    Second Number = (High+Low)/2

    Daily Pivot Differential = IF(Daily Pivot Price > Second Number, Daily Pivot Price - Second Number, Second Number - Daily Pivot Price)

    Pivot Range (Low) = Daily Pivot Price - Daily Pivot Differential

    Pivot Range (High) = Daily Pivot Price + Daily Pivot Differential

    That`s your math, if you expect Daily Pivot Differential ~ 0 you get a Pivot range width of ~ 0, the very only variable is your Daily Pivot Differential !

    let`s take your own numbers:

    Daily Pivot Diff ---PivotRange Width
    0.2033 --------0.41
    0.1783 --------0.36
    0.3300 --------0.66
    0.2150 --------0.43
    and so on...

    it`s simply a ratio of 1:2 of course, so why calculating the second at all to look for an effect on tomorrows day?

    You can use it directly for testing cause you just alter the numbers * 2 with additional math of a Pivot width range, you can test some conditions against it`s average(30) or whatever, and that`s what i did in Tradestation, with different AV lengths...

    And there is no difference if you multiply this number * 2 before :)

    I think you need to rethink the math, to ...

    condition1 = DPivotDiff < average(DPivotDiff, Length) * testfactor

    with Length from 5 to 40 and testfactor from 0.1 to 1

    found nothing useful

  10. @steve,
    Hi, exactly, that`s the stuff useful! With inside bars i got some surprise with testing as condition(day before) for a breakoutsystem(YM). The results were worse if the day before trading breakout was an inside Bar, but produced very good but rare trades if there was 2 days ago an inside bar, but not yesterday. So it seems inside bars like to come in pairs, at least nowadays. Whatever the reason is, it takes 2 days after an inside, to see fresh volatility (in YM, last 18 month)

    greetings, Michael
    #10     Mar 29, 2006