Recognizing Trend Days

Here's how market wizard Linda Raschke describes trend days:
http://www.traderslog.com/capturing-trend-days/

- Price opens near LoD
- Price closes near HoD
- shallow retracements
- breakout strategies in favour of trend consistently works



My additional observations of trend days:

- price is moving away from VWAP, which itself is trending



- TICK is biased, ie. on a downtrend day, there are few +ve TICK extremes, and the ones that do occur are on lower highs.



By definition, a trend day must be trending for the majority of the day. Therefore, if there is no successful breakout of the 1st hr IB, or if the IB breakout fails, chances are that it is not a trend day.



I am by no means an expert on trend days (just an observer!). I don't have any mechanical criteria that you can enter into an automated trade system to tell you definitively whether the day will be a trend day or not. Everything is all about probabilities. I look at the 5IB initially, and build on the ongoing assessment of whether it is a trend day or not.



I invite everyone to take a look at these charts, and study for some common patterns to help better identify trend days...

I think there have been some good suggestions.

I was interested because I recognize a trend day when I eyeball the chart, but if I were to write code to identify a trend day to search for repeated patterns on the day following the trend day, I obviously would need to define it.

phileo, when you use the term IB, what exactly are you referring to, the initial 5 minute bar? or the initial 60 minute bar.

Copied and pasted this from another topic since it fits here. Some of the discussion had to do with Don Miller's MATD ... and how to trade it. But the BIG question was what constituted a TREND DAY. And I posted here for anyone who knows what Don Miller's definition/description of a "trend day" is.

Since the MATD (Morning After Trend Day) is a Don Miller thing, I've been searching the net for his definition of a "trend day" and have yet to come up with any useful description. MATD definition link.

Anyone here know or have any idea of Don's definition of a trend day for his subsequent MATD trading? Would be helpful since it's come up several times recently on other topics/threads?

Originally posted by day trading

Originally posted by PAUL9

phileo,
can you describe in some sort of quantitative way what a trend day is?

in other words, is it a day where RTH 30 minute bars only have 1 or 2 bars where previous 30min bar low is undercut...

is there any mathematically expressed condition that exists on a "trend" day?

I know that this question is directed at phileo but I can try and answer it as it's a question that I've been struggling with for years. Here are some rough rules from memory for how I have flagged a trend day after the day's trading has completed. (i.e. can't determine during the day.)

• Open is within 10% of high or low and close is within 10% of opposite high or low. (Percentage based on range. e.g. if range is 20 points then open must be within 2 points of high/low.)
• Range is no less than 75% of 5 day average range.

That does not guarantee that you'll pick out a range day (for back testing for example) but it does catch a lot of them. There are other techniques like calculating the linear regression of the average of the 5 minute OHLC bars and then examining the price points along that regression line to determine how far they fall from the line. I have never tried that though...

Copied from other thread as it's appropriate here:

It's a question that I've been struggling with for years. Here are some rough rules from memory for how I have flagged a trend day after the day's trading has completed. (i.e. can't determine during the day.)

• Open is within 10% of high or low and close is within 10% of opposite high or low. (Percentage based on range. e.g. if range is 20 points then open must be within 2 points of high/low.)
• Range is no less than 75% of 5 day average range.

That does not guarantee that you'll pick out a range day (for back testing for example) but it does catch a lot of them. There are other techniques like calculating the linear regression of the average of the 5 minute OHLC bars and then examining the price points along that regression line to determine how far they fall from the line. I have never tried that though...

Originally posted by PAUL9

I think there have been some good suggestions.

I was interested because I recognize a trend day when I eyeball the chart, but if I were to write code to identify a trend day to search for repeated patterns on the day following the trend day, I obviously would need to define it.

phileo, when you use the term IB, what exactly are you referring to, the initial 5 minute bar? or the initial 60 minute bar.

Hi Paul,

IB is the initial 60minute bar. The more i think about it, the more I think it is an appropriate "yardstick" for assessing whether the day is a trend day or not.
Also, after thinking about "Shallow" retracements, I think I can better quantify that too:
- retracement does not violate previous swing hi/lo
- retracement does not violate VWAP

Now you've piqued my curiousity, are you developing an ATS for trading MATD's ?

Originally posted by PAUL9

phileo, when you use the term IB, what exactly are you referring to, the initial 5 minute bar? or the initial 60 minute bar.

IB = Initial Balance

Originally posted by day trading

Copied from other thread as it's appropriate here:

It's a question that I've been struggling with for years. Here are some rough rules from memory for how I have flagged a trend day after the day's trading has completed. (i.e. can't determine during the day.)

• Open is within 10% of high or low and close is within 10% of opposite high or low. (Percentage based on range. e.g. if range is 20 points then open must be within 2 points of high/low.)
• Range is no less than 75% of 5 day average range.

That does not guarantee that you'll pick out a range day (for back testing for example) but it does catch a lot of them. There are other techniques like calculating the linear regression of the average of the 5 minute OHLC bars and then examining the price points along that regression line to determine how far they fall from the line. I have never tried that though...

If my recollection is correct, Market Profile (or is it Dalton ?) defines a trend day as any day which range is > 2*IB.

I would add to that requirement that the close is outside of IB.

I think mandating the close being within 10% of the HoD/LoD (in reference to day's range) is a bit restrictive ... many trend days end-up with a last range extension that is reversed into the close (close into last "bell" developed in the day)

There could be a lot of refinement added, in particular looking for multiple "bells" separated by single (or double) prints in the profile

phileo
1,568
249 Posts

Posted - 06/03/2010 : 12:34:47 Click image for original size

Here's how market wizard Linda Raschke describes trend days:
http://www.traderslog.com/capturing-trend-days/

- Price opens near LoD
- Price closes near HoD
- shallow retracements
- breakout strategies in favour of trend consistently works

I would add to this, look for the separating line pattern which
is a continuation pattern for further confirmation.

More info. on this pattern is in the "Candle Pattern" topic.
Just my 2cents

Well, one thing's for sure, Friday, June 4 sure looked like a trend day to me. So that makes Monday morning MATD.

I have not started to study the MATDs, but here is one thing I noticed with Thursday Jun3 (day after Trend day) and the example DT included in the dictionary, MATD = April 9, 2010, both of these days had the oscillating PA in the morning, and both generated Low prints in the early afternoon that were followed by some upside.

I have just moved backward in time eyeballing 5min RTH only charts and here are trend up days

Date of "eyeball" trend day UP

3/22, next day Low was 11:05, then ascending triangle with last our breakout to the upside.

3/5, maybe some would not consider this a trend day, but if you do, next day was super small range all day long, LOD 11:30am, retested at 12:05pm and 12:50pm

What about morning after trend down day?

Trend DOWN day 5/18, on the next day, the market ran out of downside 11:50-11:55 and then gained nicely from about noon until just after 2:00pm (gain from Low was 19.75 points, not shabby, could this be something possible on Monday?)

Does Miller mention any differences between expectations for Morning after Trend day UP and Morning after trend day DOWN?

I think I read him once that if price gaps down after trend day down, he looks to buy test of previous support or print of -700+ on $tick, no time to check files (for Miller comment) now.

OK

I found my file on Miller and if Friday wasn't a monster trend down day then I don't know what is/was...

Here is my original not to myself after being a reader of miller's journal for about 5 or 6 weeks in 2009:

***************************************
One of Don's favorite set-up is described in 03/09/09 TD blog:
1) monster decline previous day
2) next day, gap down to prior day support, BUY tick under -720

***************************************

Well, support from Friday looks like 1065 to 1059.

Whenever I can get to identifying trend days and viewing the PA in the morning and afternoon following a trend day, one thing I will be specifically looking at is

does afternoon tend to put in a trend that is in the same direction as previous day's trend.

The reason I raise this possibility is simple.

A trend day is generated because the market is lopsided, either buyers are aggressive or sellers are aggressive.

that buying or selling attitude probably cannot be extinguished in just a day.

the morning following a trend day sees oscillation due to short-term traders booking profits or exiting losing positions, but the late afternoon might see continuation of the original trend day.

Obviously, this would have to be viewed in the context of the market because in a strongly trending bull market, there can be "one day wonders" that counter trend down all day, but then reverse the following day.

current market is in a down trend so I wasn't surprised to see Monday oscillate in the morning and then finally head lower at 1:35pm. That 12:50- 1:30 window of time was also apparent as the beginning of continuation of orignal trend day evident in other "day after trend" days PA.

Nice to see some structure put into the type of day. Questions to ask as the day starts, Is it trending or range bound? As a day trader understanding the trend early in the session is crucial to effective day trading. We can ride the trend days or take profits according to Support & Resistance on a range day.