If there is a narrrow range day with an open near the low and the close near the high, I would not call that a trending day. Trending means that price moves ( volatility++) and that it moves into one direction only (further north in this case). So you need to include
-> the volatility++ criterion (requirement that the range of that day is above average)
-> the directional criterion (marubozu candle opening at one end and closing at the other end)
For the volatility threshold you can select an absolute value (for example a minimum of 10 points for ES) or a relative value (for example a larger range than the average range of the preceding N days).
I understand needing a minimum volatility for the current day, that depends on how many points on average your method needs to reach average targets. But, in terms of prior day requirements, if the ES had a range of 10 points each day for the last 3 days, and today it opened at the low and closed at the high and also traveled 10 points, would that not be a trend day because the range was not greater then the last 3 days?
I think a trend only has to do with how a market moves during the day, namely starting at one extreme, ending at the other extreme, and only having relatively shallow retraces during the the entire movement (HH/HL standard). The total amount of points traveled is only dependent on how many points your specific method requires the market to move in order to be profitable.
Last edited by monpere; March 22nd, 2012 at 05:32 PM.
Today's open > yesterday's open
Today's high > yesterday's high
Today's close > yesterday's close
Today's low > yesterday's low
I compare everything to the previous day, previous week, previous month. That's how I trade.
It's my benchmark. I don't use it to predict. Only to set traded prices and their values based on the volume profile.
Maybe you could take the lows off. There's always fake outs that break the previous days low that end up being strong trending days.
Last edited by Massive l; March 22nd, 2012 at 05:46 PM.
@Massive l: This definition is not possible. Today's open does not need to be higher than yesterday's open. It is possible that yesterday was a down day, and that today is an uptrending day reversing that trend from yesterday. In that case today's open can be considerable lower than yesterday's open.
Now imagine the case that the market gapped down overnight and that a strong trending day followed closing just below yesterday's high. In that case you have a strong trending day which does not meet your conditions 2 and 4.
That means that conditions 1,2 and 4 cannot be defended at all. Condition 3 should be met but it is not ehough to define a trending day.
The following 2 users say Thank You to Fat Tails for this post:
This is a great discussion and one that I have a slightly different take on.
I've observed that during the best trending periods that "Volatility" is actually LOWER than during chop periods.. The " Directional Range" is Larger but the vertical volatility around the trendline is lower.. see attached photo- although not a great example, you can clearly see that during the trending period that the vertical range parallel to a LinReg line (volatility as I define it) is much smaller during trend periods than during the chop period...This is something that Mike COULD define in Excel as well..
Define a "minimum" directional price movement over time (slope) and low range about a Linear regression line (or MA) to define historic trend.
Just an idea...not sure if I expressed it well enough either..but it is easily observed.
The following user says Thank You to sptrader for this post: