What's your favourite combo of indicators?

[wayneL]

Hullo Wayne,
Just had to ask you about your use of CCI, I don't use this , I think its a tool for trading cyclical trends & trying to predict some sort of cycle of moving averages.
Its another osillator !
What is it in your trading application ?please.

Regards Bob.

First lets looks at how the CCI is computed and what it measures

1/ Stage one is a simple price oscillator using typical price i.e

Typical price { (H+L+C)/3 } subtracted from an x day simple MA of the typical price

2/ The result is divided by the x day Mean Deviation. This basically corrects the above raw oscillator by a crude measure of volatility.

3/ Finally, the result is divided by a constant (0.015) which corrects the value of CCI so that ~70% falls between 100 and -100.

So what the CCI is, is a volatility corrected measurement of the TP in relation to it x day moving average.

It can be used as a measure of trend (as if you can't see it already)

It can show proximity of the price to its 20 day MA, and if price finds support there, by the CCI "rejecting" or "kissing" the 0 line.

Extreme values of > 200 or <-200 can suggest a regression to the mean, i.e. a counter trend trade.

Also can show divergences like any other oscillator

This is all subject to observations of the actual price movement. Best suited to swing trading IMO. You could just as easily use the 20DMA.

 

[Bobby]

Thanks Wayne,
Wow that was bloody good mate !.
You sure know your stuff.

But they suck when you lay down the $buck. (too often)

Regards Bob.

 

[tech/a]

Hmm no takers on stochastic?

Wayne has shown a powerful aspect of Oscillator analysis which is worth a look when analysing any chart particularly one in an on going move positive or negative.


DIVERGENCE.

 

[bullmarket]

Hi WayneL

:iagree: 100% with your earlier post where you said people should try to understand the maths behind the indicators they are using.

Below is a post and chart I made a while ago describing my understanding and usage of the Stochastic indicator and how some people can misuse/misinterpret the stochastic if they do not fully understand the maths/mechanics driving the indicator.

Sometimes I also 'calibrate' an indicator to match the historical price turning points as closely as possible thus giving a higher probability that any future buy/sell signals will be correct. But I wouldn't advise tinkering with indicator settings too far away from the defaults unless one understands the maths behind the indicator pretty well.

Finally, as I said earlier I go on price action first and then look to the indicators to hopefully confirm the price action. The indicators I described in my earlier post in this thread have served me well, as an investor, to help time buying points over the years.

The stochastic is basically a 'change of momentum' type indicator. It shows where the current close is, in percent, relative to the highest high minus the lowest low range of the last X days...eg if in the last 10 trading days a stock traded between 2.00 and 2.20 and the last close was 2.10 then the stochastic would read (2.10-2.00)/(2.20-2.00) = 50%......This value is typically called the %K (or fast line) on the stochastic display. The %D (or slow line) is simply a moving average of the %K line.

Buy/Sell Signals

BUY signals are given when the %K crosses above the %D line and when especially both are coming up from below the 'oversold' line.

SELL signals are given when the %K crosses below the %D line and when especially both are coming down from above the 'overbought' line

But be aware that the stochastic, like all indicators, can give false signals at times and so should not be used on its own but in conjunction with other indicators and/or chart patterns

What I see fairly often is that traders will often call a stock 'overbought' as soon as the stochastic reads above 75 (overbought line) when in reality it only may or may not be actually overbought at that stage. By the nature of how the stochastic is calculated (as shown above) an uptrending share price will consistantly read above 75 shortly after the uptrend has begun.

The chart below shows where the buy/sell signal points from the stochastic occur on the price chart. You can see that at points A1 and A2, where the stoch starts to read above 75, the stochastic then continues to read above 75 for a period of time while the share price continues to rise (and hence the stock is not actually overbought at all during this period) until the sell signal is generated later on. Only then is the share price actually overbought

You can see that traders who called the share price overbought at A1 and A2 and sold out would have missed out on significant profits.

So the moral of the story imo is: The sell signals is given when the %K crosses below the %D and preferably when both are coming down from above the overbought line, and not simply when the %K and %D lines reach 75 from below. The reverse applies for buy signals.

Food for thought and I hope this helps someone.

In this chart I have set %K = 10 days, %D = 5 days

cheers

bullmarket

 

[Prospector]

Hi prospector



geeeeeez.....you're hard to please $4 to $12 in 4 years = 31.6% pa compound return.........I'll take that any day

But I get the jist of what you're saying

btw.....in addition to my previous post, I always go on what the price action is saying first and then look to the indicators for confirmation hopefully. If the indicators confirm the price action then I am more confident my interpretation of the chart is likely to be correct. All indicators regardless of the settings one uses will give false signals occassionally.

cheers

bullmarket

ps....CSI NY is about to start....have a relaxing evening whoever is reading this

OK, time to brag a bit - well in my first 12 months I bought MBL at $32 and within a few months sold for $58. I bought into PDN maybe 18 months ago at 11c ( you can buy a lot of 11c shares) and still have them now valued at $3.80. Biota at 47cents 12 months ago and sold them at $2.25 a few months months later. COH bought in at $20.

OK, the PDN were my gamble shares but were bought after research and making predictions that happened to come true. My worst mistake - thinking that Reefton would do the same thing but only lost around $500 on that one.


So if I can do it, then it cant be that hard, can it? And yes, the price movement is maybe one of my first triggers too.

If you want action, I trust you are watching Prison Break tonight! Excellent stuff!

 

[tech/a]

Hmm sorry wrong glasses.

Sell signals should not be considered in any trend.
There is a case in this chart for ranging.However signals should ONLY be taken in the direction of the predominant trend.

Now as another exercise try using stochasic as a stand alone REALTIME indicator.

IE post a signal both with and against the predominant trend.
Out of 10 you'll find accuracy well below 50%

Using oscillators to confirm oscillators is crazy.
Try running a commentary similar to the above in REALTIME not hindsite.

Vastly different once price action is complete ofcourse the overlay of an indicator will appear flawless.

 

[tech/a]

Prospector.

Seems you have an ability few have.

Being able to hold winners and cull losers quickly.

100k parcel of PDN equates to $300K profit.

Years ago when doing Tech Analysis at Securities Institute one of the guys had bought 100,000 Davnet at 6c while he was on it to $7 (Think it went) he eventually sold at $4 not bad for a 26 yr old.

 

[bullmarket]

Hi Tech/a

Interesting.
Daily stochastic settings used for a weekly chart.

Hows that work?

if you look closely at my chart you will see there are 20 odd price bars (1 per day) for each month and so it is clearly a daily chart....so not sure what point you are trying to make.


bullmarket

 

[tech/a]

See above I'm clearly blind.

Point is now clearer.So are my glasses.

 

[bullmarket]

no problem tech/a

See above I'm clearly blind.

Point is now clearer.So are my glasses.

good luck in your endeavours

bullmarket

btw...I had to look below (not that it matters)....I have my thread settings set to display the last post at the top.

 

[Prospector]

Prospector.



100k parcel of PDN equates to $300K profit.

Years ago when doing Tech Analysis at Securities Institute one of the guys had bought 100,000 Davnet at 6c while he was on it to $7 (Think it went) he eventually sold at $4 not bad for a 26 yr old.

Ah, but I didnt have the confidence to buy that many shares! The profit was still good though! Now, I need to work on my confidence issues

 

[bullmarket]

Continuing on with the theme that chartists should understand the maths behind indicators, below is a post I made on another chat site (I'm not allowed to mention it ) describing my understanding of the maths/mechanics driving the MACD/MACD-H indicator.

If anyone is still having a problem trying to work out what the MACD is displaying and how it relates back to the price chart, maybe try this practical experiment to help visualise what is going on.

Firstly remember that the MACD line is simply just the difference between 2 moving averages (typically 12 and 26 days, but I use 5 and 15) eg....if on a certain date the 12 day MA is $3.50 and the 26 day MA is $3.40 then the value of the MACD line on that date will be +0.10. I'll discuss the MACD trigger line later.

To visualise 'the reality' of what this means, set up a chart with just the 12 and 26 day MA's on the price data and a MACD display. You will see that when the 12 day MA is above the 26 day MA the MACD line will be positive, when the 12 day MA is below the 26 day MA the MACD line will be negative and when the 12 day and the 26 day MA's are equal to each other (eg..cross over each other) the MACD line will be zero.

So assuming a buy/sell signal is given when 2 MA's cross over then the MACD line gives a buy signal when it crosses thru zero from below to reading a positive value. When this happens on the MACD line the 12 day MA on the price chart will be crossing over the 26 day MA from below it to above it. Vice versa applies for sell signals.

The 'trigger' line on a MACD display is simply a moving average of the MACD line (typically 9 days - I use 10 days).

The MACD-H is a histogram displaying the difference between the MACD line and its trigger line. When the MACD line is above the trigger line the MACD-H will be +ve and when it is below it will be -ve.

The basic purpose of the trigger line is to preempt the expectation that the MACD line will soon cross above zero from below and so give a buy signal.

So using the MACD line and trigger line combination, a buy signal is given when the MACD line crosses above the trigger line and the cross over occurs below zero. Vice Versa for sell signals.

Therefore when you see a MACD-H reading negative values but beginning to turn up, that indicates the MACD line is beginnng to head up back towards its trigger line. Some traders will see this as an 'early' buy signal in anticpation that the MACD line will continue to move up and cross the trigger line at which the MACD-H will read zero.

I hope all this makes sense, but if anyone would like to discuss further, just post away with any queries.

cheers

bullmarket

 

[tech/a]

Have a look at September sell signals for the CBA chart.

There are 5 false signals which according to your rule for sale would have been taken.So my point of being able to show perfect signals in hindsite I think becomes clearer.

The same applies to all oscillators.

Those signals (buy) taken with the predominant trend show a clearly better performance.

 

[wayneL]

Continuing on with the theme that chartists should understand the maths behind indicators, below is a post I made on another chat site (I'm not allowed to mention it ) describing my understanding of the maths/mechanics driving the MACD/MACD-H indicator.

bullmarket

Good stuff Bull

I seen a tendency among noobs to sometimes to assighn more importance to the MACD than the price itself, as if MACD leads price.

Understanding the maths dispells that, 'onya!

Cheers

 

[bullmarket]

No problem wayneL

Tech/a: if you interpret 5 false signals then that's fine....each to their own.

Personally, I see a clear buy signal late August and only 1 sell signal about 1/3 into September where both the %K and %D lines cross below the 'overbought' line. But as I mentioned earlier I go on price action first and then indicators and since the price action at that time in Sep showed that CBA was still in an uptrend I would interpret that sell signal from the stoch as being false and as an investor and in accordance with my plan I would not have sold at that time - but a trader may or may not have sold, each to their own. I then see a clear sell signal on the stoch in the 1st week of October.

Also, bear in mind that all indicators will give false signals on occasions (as I and others have mentioned before) and so no one indicator should be used in isolation imo especially given the fact that some indicators as a result of the maths driving them will work better in some cases than others.

cheers

bullmarket

 

[bullmarket]

For the sake of completeness I've added the 5 and 15 day EMA's (Exponential Moving Averages) and the MACD indicator to the CBA chart I posted earlier with the Stochastic indicator.

So hopefully my earlier post (without diagrams) describing the MACD indicator might become clearer.

So you can see on the chart below that the black MACD line simply plots the relative positions of the 5 and 15 day EMA's to each other on the price chart as per earlier post.

I've also displayed the buy/sell signals generated by the MACD for comparing to the relative positions of the buy/sell signals from the stochastic. Overall they are reasonably close together - but it is very unlikely any two indicators will give exactly the same buy/sell signals.

Hope this helps someone

bullmarket

 

[tech/a]

It is very simple to overlay a chart with indicators and show how well they have performed.

Thats because it HAS PERFORMED.

My point and one which only active traders who have tried to trade by indicator/s will know is that when trading off of signals in REALTIME the completeness and accuracy are far from that seen on a chart in hindsite.

As the whole idea of a site like this is to help those who are struggling with some issues that active traders know and understand,I will take your 3 indicators (or you can) and present 5 charts for analysis of a simple entry.

Incidently as no buy signals coincide when then in your veiw is it appropriate to enter/exit? After all 3 have signalled?

Happy to do the same with exits if you think the use of indicators for exit is appropriate.

For the sake of completeness.

 

[bullmarket]

Hi tech/a

Imo it's not in hindsight at all because at each of the buy/sell signals when they occured you obviously didn't have the price data for the future. So it then comes down to whether and/or how the chartist uses the signals according to his/her investment/trading plan and strategies.

Regarding trading/investing in real-time as you put it, I use only end of day data and look at charts either in the evening after the day's data is available for download or in the mornings. So for me personally as an investor, that is essentially looking at charts in real-time. I don't have any need for intraday data/charts.

All I have done in earlier posts is describe my understanding of the maths driving those 2 indicators and how I use them according to my strategies and objectives and they have served me well over the years as an investor.

If what I have desecribed is not applicable to your style of trading then so be it - I don't have a problem with that at all What I described earlier works for me and if it helps someone else all well and good and if it doesn't because they have different views or strategies then that is totally fine with me......trading/investing is very much a case of each to their own objectives, methods and risk tolerances

cheers

bullmarket

 

[tech/a]

An entry signal is an entry signal regardless of timeframe.
Exit will determine trade length.

I both invest as trade.

Back to my question.
When do you determine an entry to be confirmed---at what point with your oscillator selection and or price action would you take a signal.

Software will select the appropriate conditions and if you tell me what that is I will code it and as I said run some entries using EOD charts.

As for exit that can be determined at any time you wish.
Do you trade stops?
If not then I'll leave them out of the exercise.

 

[bullmarket]

tech/a

An entry signal is an entry signal regardless of timeframe.
Exit will determine trade length.

I both invest as trade.

Back to my question.
When do you determine an entry to be confirmed---at what point with your oscillator selection and or price action would you take a signal.

Software will select the appropriate conditions and if you tell me what that is I will code it and as I said run some entries using EOD charts.

As for exit that can be determined at any time you wish.
Do you trade stops?
If not then I'll leave them out of the exercise.

Now you're just asking me to repeat what I have already posted. If you look back at my posts in this and other threads you will see that I look at the price action first (ie....support/resistance/trend breaks preferably with volume support) to help determine entry points and then to the indicators that hopefully confirm what the price action is suggesting to me. More often than not the indicators confirm my interpretation of the price action but not always obviously

How someone determines signal confirmation is up to the individual and should be written into their plan....ie.....which indicators they will use and in what priority will they need to give buy/sell signals for confirmation) Imo most of the parameters for these criteria will depend on one's risk tolerances and so there is no single right answer for determining confirmation for buy/sell signals.

cheers

bullmarket