CHART INDICATORS

[TerryA]

Could one, or all, of the Charting Gurus tell me which Indicators they find of most use for following the NZX and the parameters that give the best indications.

I have been trialling ATR, RSI, MACD, Stochastic, Williams%R, Directional Movement (all 4) and the 30/120 Simple Moving Averages.

Thanks,

TerryA

[scamper]

If you mean shares on the nz exchange, different indicators should be used for different horses on different courses etc. [:p]

If a stock is trending: trendlines, ADX ribbon of DMIs, parabolic SAR, and moving averages are useful.

If a stock is ranging, momentum oscillators such as RSI and stochastics are better.

I also give different emphasis to different indicators when wanting to buy from when considering a sell.

The way individual stocks leave indicators depends greatly on their liquidity and so on, so I always check whether a particular indicator has proved valuable for a particular stock in the past. It is always good to pull up a three-year chart and see which clues have proved historically valuable.

Sorry if that's not very helpful -- am definitely not a charting guru, but always consider TA for timing. Of course, inadequate discipline negates all utility...!

[Dazza]

which free thing on the web has trailing adr as one of their indicators? i cant find it on national bank, or big charts services, yet i always see it on P's charts, is it only avaliable in metastock?!

and i agree, if some of u TA gurus, could use this post as a teaching thread for us, it would be great :D

[limegreen]

Phaedrus had a great thread on sharechat with a whole load of references. Personally, I mostly watch the OBV (general sentiment), plus the RSI for buying, the rest is trend and resistantance/support lines.

Still learning though. Failure to follow TA cost me a good whack on NOGOC, but think the learning from that will be positive in the long run.

[TerryA]

Limegreen,

I am hoping that Phaedrus will pay this thread a visit !

Many of the various texts I have seen give the parameter but not why that value has been chosen. For instance, In some postings there are a number of recommendadtions for Moving Averages but nothing that I have seen, to date, that argues the merits of any particular duration or type.

The results of using the extremes in all the indicators are obvious but I was hoping to get the benefits of others experience in finding the "best fit".

Thanks for your contribution.

TerryA

[limegreen]

Terry,

Summarizing what I think I've heard/read (but possibly misrepresenting), is that it depends on both your investment outlook, and the particular stock. For example, I take a fairly long term view, so my default moving average is 200. However, for some stock's I've seen Phaedrus go out to 365 (e.g., FBU), and I think that's because it fits past data better. I've also seen elsewhere that Phaedrus sometimes uses a higher value to make RSI less twitchy. However, the RSI parameter is not one you can usually change on free sites.

From the little reading I've done, I believe that back-testing is quite important. That is, to look at the indicators you've selected for a stock, and see how well they would have served you in the past. Obviously, the past is not a perfect guide to the future, but will hopefully help you pick indicators which match the volatility etc. of the stock in question.

NB: Phaedrus has also said that he likes the OBV for precisely the reason that it is parameterless (well not strictly, but it takes a day as it's unit, and unless you were to calculate it weekly, the parameter is effectively fixed).

[TerryA]

Limegreen,

Thanks for those comments.

Having missed the signs of a couple of "drops", which cost me relatively dearly, I recently invested in Metastock and InfoScan's EOD service.

I can therefore change as many parameters as are available and I do as much back testing as I can and am gradually devoping my own template(s) My question was put in the hope that I could speed up this process.

Will let you know if I see the light at the end of the tunnel.

Regards,

[scamper]

Agree, Limegreen!
Terry, it truly depends on what sort of an investor you want to be: long, med, short-term; high, med, low risk etc which usually reflects your age, asset status, available time for nitty-gritty research etc. There really is NO correct answer for parameter length.

Some people trade at +/- 10%, others let profits run until a major reversal, or until they need some cash, or until they see a better opportunity. From this it follows that some stocks get flipped within days, or for (mere)hundreds of dollars, others are left to make 300%, some are held for a decade.

Good advice (which I seldom follow) seems to be to WRITE DOWN why you bought a particular stock and what your expectations are of that stock. This record should be revisited regularly, and the stock sold when it no longer meets your expectations.

It's all fairly fraught if you don't stick to the plan, or if the game plan proves faulty -- like picking a dodgy performer to begin with. It is interesting to visit the 'one-liners' thread on this forum -- you will see that every conceivable strategy has its moment!
Best wishes, Scamper.

[Phaedrus]

Terry, Backtesting tells us that for any given stock there is always a "best" indicator and period. Now, this does not guarantee that it will continue to be so, but I feel that past performance gives some indication of signal reliability. If a certain indicator gives a Buy signal, and 7 of the last 8 were "wrong", would you act on it? Nah.
I am a staunch advocate of optimisation - the process by which you find and use the indicator period that has proved to be the best historically. Do appreciate though that you can optimise for many different characteristics. Maximum profit, minimum number of trades, best hit rate, best win/loss ratio etc etc. This is what Scamper was getting at.

Limegreen, Once a trend is established, I like to find the shortest period moving average that just skims below the reaction lows of the trend. I then use this in exactly the same way as a trendline drawn using the same points. The aim is to get early warning of a weakening in the trend. Re RSI period - I often shorten the 14 day default value to 8 to make the RSI more responsive, and also often use 40 or higher where I want a less active indicator.

Re using the cross of 2 moving averages - I have found that almost invariably this system is inferior to using an optimised single moving average. This is partly the reason why I do not use the MACD at all - though it is quite popular with many others.

Don't waste time looking for the perfect system or indicator. They don't exist. Watch out for "curve-fitting" if you do choose to optimise parameters. Don't get too carried away looking at indicators - of any sort. The more complex they are, the more removed they are from reality and the further they are from the ultimate oscillator - Price. Never forget that other than support, resistance and trends, all else is an artificial mathematical construct. Trendlines, moving averages etc are not "reality" and prices do not really bounce off them in the way that support holds, for example.

[TerryA]

Phaedrus, and others,

Many thanks for your valued contribution and I will continue with trying fo find a series of templates that best suit what, I think, are my needs.

Previously I have been a long term investor, and rarely sold anything, but have now changed put my shares in three "portfolios" 5years+, 2-5 years and 0-2 years (which may include some relatively short term trades.)

Fundementals will continue to be the major factor in any decision but I am finding the TA aspect an interesting intellectual challenge.

Best wishes,

[Tinker]

Hi TerryA

Good luck with your eploration of TA. I'm not sure how far along the track you are but one thing I found was that extensive searches can very easily ending up "data mining" ie that if you explore the data enough you will find a past relationship which is co-incidental and is neither predictive nor correlated to past movements.

An example I sometimes give is that if I search enough I could find a security that falls whenever my cat catches a bird.

Past matching of a formula to a security does not mean it is predictive.

Trying to help.

Cheers
Tinker

[kittydashwood]

I'm using the williams to sell on strength parts of my portfolio I feel are too risky or fall outside my parameters of inclusion.

[TerryA]

>>I'm using the williams to sell <<

Thanks Kitty,

I have also started to use that indicator but still have a long way to go before I will be confident of the results. However that and OBV in conjunction seem to giving a good corelation for both buy and sell.

[Phaedrus]

Terry,
You have to be clear about exactly what you want to do before you can start looking for appropriate indicators. If, for example, you decide that you want to be a trend-follower, then you will place heavy emphasis on trend indicators, as mentioned by Scamper, above. W%R and OBV are useful indicators but they are not monitoring the overall trend at all. At best, they would be no more than an adjunct to a trend-following system.
Re backtesting. As Tinker says, there is a risk of "data mining" or "curve fitting" when you do this. There is a relatively simple way by which this risk can be eliminated, though. Separate your data into 2 groups and do all your back-testing and optimisation on, say, the first 5 years (if you have 10 years data) Then, using the optimised parameters, test the efficacy of your tentative system and indicators on the second 5 years. You can then objectively assess the worth of your trial system. If you get fantastic results with the back-tested data and lousy results when you "forward test" using the second set of data, then your system is no good - you have been curve-fitting. My experience is that the more complex you make your system, the better the back-tested results and the poorer the "forward tested" performance. Keep it simple.
Do not get too technical too soon. I would advise you to start out trend trading - get a good feel for trends, trend indicators, charts and candlesticks first.

[TerryA]

Phaedrus,

>>Do not get too technical too soon. I would advise you to start out trend trading<<

Again my sincere thanks for your sage advice.

I am far from starting trading,in the true sense of the word, but am using some of the indicators as checks on (paper) decisions based on trend lines. Also to go back over some SP changes, almost all falls, where I did not have the information to spot the breaking of trend lines.

Hopefully I will not have the same problems again but I need the market to continue up for a period so as to establish some decent history.

Kind regards,

[Tinker]

Hi TerryA,

I do not mean to be adverserial nor follow Phaedrus but may I make a few suggestions. If you are serious about using charts to endeavour to increase your wealth then use the technology that is now available.

Buy a charting package ( Metastock is common), buy a decades worth of data ( standard open, close, high, low, volume - in the market of your choice) then rip into developing a system. You can test every idea you have and run wins/losses on spreadsheets to your hearts content. As a % of ones aims for profit it is a small investment, and as insurance it is an even smaller percentage of ones potential losses.

The ASX200 Acc index (which I follow) went up 23% 03/04 and 25% 04/05. If you flipped a coin 1/2 would have done better, half worse in either year. If you keep flipping the coin the right way every year you become very wealthy but not as rare as you might think given the large no. of people in the market. So the need to distinguish chance from skill is important.

Best of luck.

Cheers
Tinker

[TerryA]

Good Morning Tinker,

Thanks for your comments which are much appreciated.

I have already invested in metaStock and am coming to grips with it. Data is sourced fron InfoScan on an EOD basis for the NZX. This covers the bulk of my holdings and the relatively few shaores i have from the LSE and ASX are fairly easily covered by using the available public information, Yahoo for the UK and the feed from Direct broking for Aust.

I will not be entering any of the Stock picking competitions for a year or so !

Best wishes,

[trendy]

TerryA if you want to save yourself a bundle on buying EOD data you can get a shareware program that allows you to download the data for the NZX and almost any other exchange for free. It uses the yahoo finance site to access EOD data, you just need to adjust it for splits manually. Also provides the correct Metastock format.

Check it out. www.hquotes.com

[TerryA]

Trendy,

Wish I'd known about that before I took out the InfoScan subscription, although I did get a good bundled deal from them.

Will try it out when I get near the end of the subscription period.

Many thanks,