Detecting Sideways Market?

[ephi144 10]

sorry i am noob so forgive my asking:

is their a such a thing as indicator that detects sideways market?

i would be interested in such an indicator for Tradestation.

 

thanks

[Soultrader 12]

sorry i am noob so forgive my asking:

is their a such a thing as indicator that detects sideways market?

i would be interested in such an indicator for Tradestation.

 

thanks

 

Hi ephi,

 

Nothing to do with indicators here but you can compare volume to previous average volumes. For example, today the Nikkei is stuck in a dull range. Notice the volume in comparison to Friday. Low volume today showing not much participation in the markets... these tend to create rangebound days.You can simply wait for the opening 15-30mins as well to gauge activity.

 

[torero 10]

Seeing the price action should give you enough clues that a sideway markets are in progress. If you see the prices don't break the highs or lows of the time chart you're looking, then it's a sideways market. Pure and simple. Of course determining which time chart to use is another problem.

[ephi144 10]

hi guys, yeah i agree about looking at the price pattern to see if it breaks above/below that level but sometimes it takes a while for that to come in to play. i am just trying to devise a system based on some emas etc to see if i can come up with something.

 

thank you for your input.

[zdo 20]

ephi,

 

What platform are you using?

 

zdo

[ephi144 10]

tradestation.

 

thanks

[stanlyd 10]

many traders use a volume weighted MA to look for "sideways " action. If MA period is usually short, and when it goes flat it is considered sideways. I have seen this in practice, but I personally don't use it as I still think it lags compared to being able to read price action itself.

[zdo 20]

...when it goes flat it is considered sideways...it lags compared to being able to read price action itself.

 

ephi,

 

lag and responsiveness are issues with regular averages but this principle works well with some of the more adaptive MA's. For example, Jurik, etc. set to 4 periods on a low timeframe ndx chart gets flat quickly enough to be useful in algorhythmic trading.

Have you looked at Joe Ross's definitions of congestion found in Law of the Charts book(s)? These are also easily coded.

 

hth,

 

zdo

[ephi144 10]

guys these are great ideas for noob like me thank you so MUCH.

i have been working with (experimenting with) screaming 2 fast MAs that are found on this site along with 55 period EMA. I have incorporated BB Squeeze indicator (though i have yet to figure out how i will use it along with the MAs). Next i am wondering if i can someone figure out the choppy or sideways market.

 

I will look in to Joe Ross' book for more information. Please do send more ideas my way.

[BlowFish 10]

No need to look for a book. You can download the Law of charts from his web site. Joe uses a 'measuring bar' to define congestion...all the time subsequent bars have an open or close within the measuring bar you have congestion. His definitions are pretty robust actually.

[rangerdoc 13]

What you're asking is really the million dollar question: If there was a way to reliably know when a market was transitioning from sideways to trending or vice-versa, we would all be very rich indeed. One can only really know after the transition has occurred.

 

A problem here is defining a trend. After that, a "sideways market" is a lack of trend as you define a trend.

There are about as many techniques to answer this dilemma as there are indicators. Multiple MA's, Comparing two different ATR periods, Rate of change in the ADX line or just looking at the chart. By it's very nature, you can not tell a trend until it is underway.

 

Welles Wilder's ADX (sometimes called DMI) is specifically designed to differentiate between trending vs. sideways markets. Like any indicator, there are limitless ways to interpret it. Personally, I largely ignore the level of ADX (contrary to Wilder's methodology). What I use to determine trend is the direction and rate of change in the ADX line. Interpretation is largely subjective and very visual. Since it, like any indicator, is simply a manipulation of price, many would argue that you are better off visually detecting trend vs. range directly from price and not from an indicator, which will always lag behind price.

 

Take a look into ADX. It's not a very intuitive indicator and therefore not very popular. It takes some getting used to. Those who are comfortable with it swear by it, and it often becomes their favorite indicator. Personally, ADX is the first indicator I want to attach to any chart and sometimes is the only indicator I use.

Edited November 12, 2008 by rangerdoc
spelling

[BlowFish 10]

You might want to look at John Ehlers work if you have a mathematical bias. A lot of his stuff is based on separating the cyclical component of a 'signal' (he was an engineer and applies signal processing techniques) from the non cyclical (trend). He nearly always uses the Hilbert transform for this. Fast Fourier transforms are another popular way. Quite a few of his algorithms have been coded and are in the public domain.

 

Personally I'd just stick to HH's and HL's. If you have those you have a trend. In the long term being able to identify a trend by glancing at a chart is likely to serve you best. (Unless you have some other objectives like automation?)

 

Cheers.

[nburgess 10]

The trouble with ADX is that it's so laggy that by the time it tells you the market is going sideways a trend is starting to form again. It's best to augment ADX with more responsive AMAs.

[rangerdoc 13]

You're right, ADX is about the slowest thing out there. Personally, I don't use hardly any indicators anymore. I'll use one medium oscillator for divergances and one volume based indicator... that's it. And in their own window, never on the chart window. I keep the price naked. All these indicators are just derivatives of price. Now you can use the latest computer (quad core) that processes about 59,000 MIPS (Million instructions per second) to determine trends and price action... or you can use your brain which has a processing speed of 100,000,000 MIPS. Which do you think can do a better job of evaluating a price chart? I found that I was trading the indicators, not the price. Once I stripped all the indicators away, my profitability increased about 40%. This takes time though. Personally, it took me five years. Others maybe much more or much less.

 

If that just won't work for you, then print the chart, hang it on the wall and call a five year old kid over and ask him if it's going up or down. If he can't tell, it's sideways. (In other words: use his 100million MIPS computer if your own isn't working) Kids have the huge advantage of not being biased in any way. Hmmm, Maybe I need to put a daycare on retainer???

[stockmarcus 10]

Hi, im also a noob, though here are some ideas im developing:

 

Using bollinger bands at one deviation, if they're moving sideways, and with the price oscillating between the inner and outer bands at 1 deviation, thats a good clue its going sideways. Also the if the width of the channel formed by the deviation lines is narrowing thats another good clue prices have stopped. (if prices are trending and this is happening thats a whole new story..)

 

Also more loosely, if you can get a historgram of macd, if the volume of the histogram is near 0, thats also a seems like a good indication its sideways.

 

Disclaimer, im also a noob

[masterg 10]

sorry i am noob so forgive my asking:

is their a such a thing as indicator that detects sideways market?

i would be interested in such an indicator for Tradestation.

 

thanks

 

Try using an ADX chart.

 

That will let you know if it is trending or not.

[ephi144 10]

i added BB with 1 diviation and see how that works.

i am going to ADX soon as soon as i get some feeling for the BB.

thanks guys.

[Fredd 10]

Hi guys. Nice discussion about trending and trading markets. I agree that an ADX is too slow to use in its original format, and I wonder if someone found a way to make it respond a bit faster, because the concept is good.

[stockmarcus 10]

A random from last night watching europe do it, another obvious clue is just look at the prices, around your favorite time period. The channel literally goes sideways.

[sevensa 10]

Hi guys. Nice discussion about trending and trading markets. I agree that an ADX is too slow to use in its original format, and I wonder if someone found a way to make it respond a bit faster, because the concept is good.

 

By definition any indicator dedecting trending/sidweways markets will be slow and lagging and it can only tell you that the past x period was sideways, but it doesn't tell you anything about the future. The only way to accurately detect a trend/sideways market is after the fact. Unfortunately you will never know early enough if a market is trending/sideways.

 

This is why trend following systems have such low winning percentage, because they assume all breakouts will be the start of a trend and they keep cutting losses until they hit a trend.

[sep34 10]

I dont have the skill to predict chop with accuracy that will occure in 15min or 3 days etc.., however I can stay out of choppy price action as price is unfolding...anyone that tells you its not possible...well they are telling you their experience, not a fact of point about technical analysis. Its up to you which you decide to belive.

 

Learn about ease of movement principle, or volatility, or study characteristics of price when its choppy vs. when its moving free/smooth, or define swing points and then compare swing point price composition to next swing point price composition, or even better do all of these. Take notes and see what you come up with.

 

Eventually Im sure you will be able to see chop as its unfolding and before you get trapped in it.

 

Price can leave a lot of clues, but you wont see it if your not paying attention or even worse, if you have already decided that its not possible.

 

 

Regards..

Edited December 31, 2008 by sep34

[darthtrader3.0beta 10]

I dont have the skill to predict chop with accuracy that will occure in 15min or 3 days etc..

 

Neither does anyone else...

to me this whole question is flawed...

Like right now on the index, are we trending off a bottom, trading a range or both? If you get abstract enough, we are always basically doing both...to try to quantify that though to me misses the point that in any market under any macro condition, there are better places to put on bets than other places. Your ability to find patterns in the clouds means nothing if you can't wrap your head around the good places to bet.

If your talking sideways action your talking patterns that are on the chart vertically, don't know how you beat Dalton there to learn. What you don't want to do though is become a failed trader who ends up teaching bogus market profile "theory"...so don't get too into it.

[sep34 10]

If your talking sideways action your talking patterns that are on the chart vertically, don't know how you beat Dalton there to learn. What you don't want to do though is become a failed trader who ends up teaching bogus market profile "theory"...so don't get too into it.

 

 

I cant understand at all what you are saying. If you need a response then clarify your statement.

 

 

Regards..

[zdo 20]

By definition any indicator dedecting trending/sidweways markets will be slow and lagging and it can only tell you that the past x period was sideways, but it doesn't tell you anything about the future. The only way to accurately detect a trend/sideways market is after the fact. Unfortunately you will never know early enough if a market is trending/sideways...

 

At beginnings of / Going into congestion - It is possible to algorithmically and statistically

1 assign probabilities to whether a thrust or correction end will turn or congest

2 start putting probabilities on congestion on its very first bar

3 be getting good probabilities as to the ‘quality’ / type of a congestion by as early as its 3rd bar

 

Within congestion – It is possible algorithmically and statistically

1 project probable end of congestion / break out times

2 then, in real time, drill down and test the quality of those projections

3 and by separate measures, identify by end of one bar when a ‘release’ from congestion is likely to have been triggered (well before price has actually broken out)

 

… not picking on you personally Sevensa. Your post is just representative of all the “it can’t be done” assumptions about working congestions before and during instead of only after...

 

congesting is entangled up in the very core of market dynamics... 'Philosophically' / attitudinally, for me it is about a primary choice to be aggressive and creative instead of helpless and reactive in my interaction with this extremely crucial aspect of market behavior... and I'm definitely not saying that I still don't get the living sh!+ kicked out of me by congestions on a routine basis - but with the engaging instead of fleeing way, at least it is with far less frequency and degree. More than the money available around congestions is the benefit of being capable of sustaining 'in synch' both in and out of congesting

 

btw my congestion work is all on very short intraday time frames. For me it would be a complete burnout to attempt to gather this kind of tradeable information manually / observationally / with wetware… but that doesn’t mean it isn’t possible by applying analogues / adaptations of the same techniques to longer time frames and being willing to give up progressively more precision with each larger time frame quanta.

 

Happy new year all

[sevensa 10]

At beginnings of / Going into congestion - It is possible to algorithmically and statistically

1 assign probabilities to whether a thrust or correction end will turn or congest

2 start putting probabilities on congestion on its very first bar

3 be getting good probabilities as to the ‘quality’ / type of a congestion by as early as its 3rd bar

 

Within congestion – It is possible algorithmically and statistically

1 project probable end of congestion / break out times

2 then, in real time, drill down and test the quality of those projections

3 and by separate measures, identify by end of one bar when a ‘release’ from congestion is likely to have been triggered (well before price has actually broken out)

 

… not picking on you personally Sevensa. Your post is just representative of all the “it can’t be done” assumptions about working congestions before and during instead of only after...

 

congesting is entangled up in the very core of market dynamics... 'Philosophically' / attitudinally, for me it is about a primary choice to be aggressive and creative instead of helpless and reactive in my interaction with this extremely crucial aspect of market behavior... and I'm definitely not saying that I still don't get the living sh!+ kicked out of me by congestions on a routine basis - but with the engaging instead of fleeing way, at least it is with far less frequency and degree. More than the money available around congestions is the benefit of being capable of sustaining 'in synch' both in and out of congesting

 

btw my congestion work is all on very short intraday time frames. For me it would be a complete burnout to attempt to gather this kind of tradeable information manually / observationally / with wetware… but that doesn’t mean it isn’t possible by applying analogues / adaptations of the same techniques to longer time frames and being willing to give up progressively more precision with each larger time frame quanta.

 

Happy new year all

 

 

You can assign probabilities and projected start and ending of anything, using any kind of algorithm. How accurate that is, is a complete different matter.

 

Care to post a few charts with your probabilities assigned of future congestion and of charts in congestion with your forecast on when it will end?