No indicator can really lead any market

Chart Forum » Indicators » No indicator can really lead any market

«« Previous Next »»

Thread Start New Thread
Last Poster Posts Pages Last Post

Indicators - time period alexm 1 1 22-Jan-08 04:24 pm

Author

Message

jross

Member

Post Number: 7
Registered: 03-2004

Rating:
Votes: 5


Friday, April 02, 2004 - 03:08 am:

Once a person understands price-bar action it is possible SAFELY say that indicators do virtually nothing! As a matter of fact I have on many occasions successfully sold short while more than one indicator was oversold.

The problem with indicators is that contrary to what some may say about them “leading” the market, the truth is that no indicator can really lead any market. An indicator can only register after the fact, never before.

I have consistently been against the wrong use of indicators. Indicators can be useful in revealing something that is difficult to see just viewing a chart. For example, Bollinger Bands reveal the location of two Standard Deviations, which is rather difficult to see just looking at a chart. An indicator can confirm what is seen on the chart.

However, most traders use indicators in a wrong way. For example, I have seen traders use RSI to confirm what they see with Stochastics. Doing that reveals a complete misunderstanding of how an indicator can be used. RSI, Momentum, %R, Stochastics, DEMA, MACD, and MACD Histogram all are measures of momentum. How can one confirm the other when they are all measuring the same thing? What you need for confirmation is something that measures volatility, as do Bollinger Bands, or something that measures volume for confirmation of momentum and price action.

Like most people, I had to fool with and play around with indicators. But I gave them up after a short time and went back to the way I originally traded – without them. Once you're in, the whole trade becomes a matter of management. All an entry method can possibly do for you is to help you be on the right side of a trade at entry. After that, it is all management. Who knows where the next tick will be???

Barry Lind (Lind-Waldock) once did a study showing that 80% of traders were on the right side of a trade at entry. He wondered, then, why 90% lost. I don't know if he ever figured it out, but I did. The problem was management – or the lack of it.

-----------------------

This article was contributed by Joe Ross

stevo

Member

Post Number: 140
Registered: 01-2003

Rating:
Votes: 2


Friday, April 02, 2004 - 01:59 pm:

Believing "leading" indicators is a recipe for disaster - it took me a while to figure that out!

I have found that I am better buying when the momentum indicator is "overbought" and selling when it is "oversold". Simple Momentum and Rate of change indicators can show strength in the market and, I believe, are more useful when used in this manner.

Management comes into play before the trade is even taken in determining how many dollars to risk on each trade. Position sizing, combined with exit strategies are all I can really control on each trade.

stevo

riddler

Member

Post Number: 84
Registered: 11-2002

Rating: N/A
Votes: 0


Saturday, April 03, 2004 - 10:22 am:

I think indicators are most useful in selling software. Which ever one(s) you use, if it is not more reliable than 50% of the time you are better off flipping a coin. For me, rsi divergence scores the best.

r

williamat

Member

Post Number: 216
Registered: 09-2002

Rating: N/A
Votes: 0


Monday, April 05, 2004 - 11:28 am:

I guess we are putting a finer point as it were.
Joe Ross, I am in no position to argue and therefore quote:- ELDER: Trading for a Living. P 120 wherein he refers to MACD as trend follower and Slow Stochastic as Oscillator, which can often turn ahead of prices.
While there is more to entry signals than indicators, I use searches for SS crossovers and MACD crossovers, as separate searches.
Respctfully and out of interest: could you describe for me "Price-bar Action. It's a new one for me.
Good Wishes and thanks for arousing me into thought. Bill.

The difference between intelligence and education is this- intelligence will make you a good living. Charles C Kettering.

jross

Member

Post Number: 8
Registered: 03-2004

Rating:
Votes: 1


Thursday, April 08, 2004 - 06:22 am:

Hi Bill,

Price bar action is chart reading. Looking at the individual price bars to see what each one reveals. The price bar contains the most complete and accurate statement of what is currently going on with price. Learning to read that action to see what is actually revealed is for example, picture a daily price bar that looks like the following:

High is at 55 and low is at 25. Open is at 35 and Close is at 50. Go ahead and draw it

What can you tell me about this price bar? How would you characterize what took place that day, assuming this is a daily bar.

You should be able to name at least half a dozen things that this bar reveals. Can you name them? Think a bit before you answer. My answers are below, but you shouldn’t peek until you’ve thought for awhile.

Here are my answers, there may be more:

1. Prices were in a trading range between the high and the
low.

2. There was volatility that day. It can be measured as
the arithmetic difference between the high and the low.

3. There was momentum.

4. There was both buying and selling.

5. Buying pressure was greater than selling pressure, thus
forcing prices up. The bulls won over the bears.

6. There was liquidity to the extent that people were
buying and selling.

7. There was volume.

Speaking of A. Elder, I see you may have fallen for his doublespeak. MACD and Stochastics are both measured of momentum. Since they both measure the same thing only in slightly different ways how can one be a “leading” indicator and the other one not be leading?

Another question: Since no indicator can possibly register its final result until a price bar is complete, how can any indicator possible lead the price action?

All the best,

JR

deanrosario

Member

Post Number: 254
Registered: 11-2002

Rating: N/A
Votes: 0


Friday, April 09, 2004 - 12:53 am:

Hi JR

I really enjoy reading your views on trading - I too try to keep it very simple.

Just a couple of things in relation to your example above ..

(a) how do you know there was volume? I know it is unlikely, but, in theory, there could have been 4 trades all day for 1 share at the open, high, low and closing price, or have I missed something?

(b) same goes for liquidity and volatility - without the intraday chart or course of trades, how can we be sure that prices were not basically trading around one price with the odd trade at the open, high, low, close?

Interested in your thoughts - especially your views on the importance, or unimportance, of volume when taking a position.

Regards - and keep up the postings they are very useful!
Dean

jross

Member

Post Number: 9
Registered: 03-2004

Rating: N/A
Votes: 0


Friday, April 09, 2004 - 05:13 am:

Hi Dean,

I know there was volume, I just don’t know what it was. You are right there could have been only 4 trades, in which case volume was 4 J

As for liquidity, there had to have been some. You can see that prices changed direction at least 3 times. Liquidity is made up of two things: volume + participation.

If there were only 4 trades, then we had both volume and participation. Therefore we had liquidity.

Volatility in the case of a single bar is the arithmetic difference between the high and low. We had a high and a low, therefore we had volatility.

In the cases of volume and liquidity, I do not know how much there was, unless I had some number for volume. But as far as liquidity is concerned, I never know how much there was. Liquidity means buyers are hitting the offer and sellers are hitting the bid. Does any one ever know how liquid a market is? Only if you know both – volume and how many participants were actually trading.

Is volume important when trading? I believe it is. If I’m trading the e-mini S&P I like to see contract volume of at least 3,000 contracts/minute. I also like to see tick volume of at least 160 ticks/minute. When I see that, I assume participation is good and liquidity is also good. For the e-mini Nasdaq, I like to see 1,000 contracts/minute and 140 ticks/minute. All of the above are subject to adjustment from time to time. What I’m really looking for is how much in the way of movement size am I getting at those volume numbers. In other words, if S&P mini is consistently getting good momentum with tick volume a only 100/minute and contract volume at only 1,000/minute, I would lower my volume target and say, “that’s good enough for me.”

Keep well,

Joe Ross

deanrosario

Member

Post Number: 255
Registered: 11-2002

Rating: N/A
Votes: 0


Friday, April 09, 2004 - 09:25 am:

Thanks for the explanation, Joe.

I now understand the point you were trying to make with the example - that the plethora of price indicators do no more than provide a picture of information that is contained in the PRICE BAR.

I too use a similar metric of "volume traded per minute", to assess "interest" in a stock at any particular time.

Best wishes for Easter.
Dean

david_louisson

Member

Post Number: 6
Registered: 02-2004

Rating: N/A
Votes: 0


Saturday, April 17, 2004 - 12:56 am:

Hi Joe

Great post, I totally agree. All indicators are ultimately derived from raw price bar data.

The only practical use I’ve found for indicators is where they can summarize aspects of the OHLCV bars that might not be immediately obvious at first glance. For example, a custom indicator that weights several comparisons between yesterday’s and today’s price bars (I am an aspiring short term trader), and then displays the consolidated result as a single value. Like other indicators, moving averages can be used to visually summarize the events of the previous ‘n’ days, but the longer the period, the greater the introduced obsolescence (‘lag’).

With all indicators, one has the additional problem of how to arbitrarily ‘calibrate’ them (i.e. assign parameters to best fit their curve around the changing cyclical patterns of the market). Many traders seem to simply use the default values supplied by the software retailer, regardless of the stock, the instrument, or the timeframe, that they are seeking to trade.

Indicators derived from the same OHLC values do not confirm each other, or increase the probability that a trade will succeed. Confirmation is subject to two entities being of independent origin. If the OHLC values rise, for example, then this must necessarily trigger rises in all OHLC-based indicators according to their construction, and calibration.

The most reliable signals I’ve found (for picking short term price reversals) are changes in the price bar ‘storyline’ and direction, and their position relative to simple channels based on (adaptively calibrated) standard deviations and prior support/resistance.

Cheers
David Louisson
Hamilton, New Zealand

hilarius

Member

Post Number: 32
Registered: 04-2004

Rating: N/A
Votes: 0


Saturday, April 17, 2004 - 09:58 am:

Good Morning David

(I say this realising that you have already beaten us to the afternoon showing once again what a progressive nation you inhabit)

Wormald proposes that many (if not all) stocks have dominant angles, and that future price slopes will mimic past ones.

Do you give this any credence?

With Best Wishes

Hilarius (a humble friar)

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

andrewk

Member

Post Number: 121
Registered: 09-2003

Rating: N/A
Votes: 0


Sunday, April 18, 2004 - 01:40 am:

Posted by david louisson:

The only practical use I’ve found for indicators is where they can summarize aspects of the OHLCV bars that might not be immediately obvious at first glance.

Exactly.

Love them or hate them, that is the point. Realistically, you need some sort of leading indicator to tell you when you should enter a trade. Whether this is derived from subconscious analysis of price-bar action, or simply a quick glance at a price or volume derived indicator. The key to using indicators is to understand what they are really telling you. To simply say *insert indicator name here* is telling me to buy/sell/hold whatever without understanding why will of course lead to trouble. In the same way it is possible to successfully trade by simply watching price movements, it is technically possible to successfully trade just by using a single indicator.

I can't say I personally use them, but that does not mean they cannot be useful.

"We are generally the better persuaded by the reasons we discover ourselves than by those given to us by others." - Blaise Pascal

hilarius

Member

Post Number: 34
Registered: 04-2004

Rating: N/A
Votes: 0


Sunday, April 18, 2004 - 07:27 am:

Good Morning

I think David Louisson's messages have been outstanding for their clarity of expression, and I totally agree with his conclusion that indicators clarify past market action by highlighting aspects of price behaviour visually, in a way which may not be obvious from a table of data.

I believe it would add value to the discussion if we were to clarify what is meant by a "leading" indicator.

As a humble friar searching every dark corner with his candle for new insights (not to mention the Holy Grail) I would be interested to know what "leading" means.

I do know what "lagging" means ... and I can best summarise it by saying that a 150 point indicator lags by more than a 30 point indicator ... because the 150 point based indicator reviews a longer period of past data. Like the QE2 it takes longer to turn than a small sailing dinghy, that is to say it lags in its response to changing conditions.

The direct opposite of a lagging indicator might then be a BF entry as defined by Rocky ... which uses as its basis only the most recent action within the single latest day and perhaps 3 recent days to draw a conclusion.

Day traders will be even more sensitised to recent data by using intra-day 1 minute charts.

This, however, still does not seem to get to the essence of the term "leading".

Economists use this term when observing past data that signals new emerging future levels of activity ... or in other words they seek to find in recent current data (eg phone connections, new car sales, retail sales etc, bank lending, confidence indices etc) signs of a reviving or slowing economy.

In short they are seeking to find in recent data indications of new trends.

It seems to me that share traders fall into 2 categories :-

(1) Those who believe that certain patterns or trends have a higher than 50/50 tendency to continue in a certain direction into the future ... and are thus tradable signals

(2) Those who believe that the market is so chaotic and random that it is never possible to deduce future moves from past data and thus their entire focus must be on money management.

In the second case a share list pinned to a dartboard and the throwing of a dart would be as useful a method of selection as any other. There would be no need to study charts for clues.

So I am interested to see who is in each group of selection method :-

(1) Those who believe past data does create patterns that offer better than average clues to the future

(2) Those who believe that nothing can be reliably deduced from past price action

Note that I am discussing here the selection process ... not the post entry management process.

While I wait for the wisdom and clarity of the Forum I shall return to my monastery duties. Hopefully I will also have a little time to read my texts on trading with my trusty candle while quietly humming to myself "Lead Kindly Light" [Source: Hymns Ancient and Modern]

With Best Wishes

Hilarius

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

vermante

Member

Post Number: 142
Registered: 11-2002

Rating: N/A
Votes: 0


Sunday, April 18, 2004 - 01:31 pm:

Hilarius

Try the RSI as a leading indicator.

Extract from Dr Elder's Trading for a living-

"The RSI is a leading or coincident indicator - it is never a laggard".

There is an art in using the RSI as a leading indicator -explore its possibilities . Try to establish why Elder made those comments.

Cheers

Vermante

hilarius

Member

Post Number: 35
Registered: 04-2004

Rating: N/A
Votes: 0


Sunday, April 18, 2004 - 01:46 pm:

Thanks Vermante

I will focus my attention on RSI and its leading qualities

Hilarius (re-lighting his candle ... preparing to burn the mid-night oil)

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

shaguar

Member

Post Number: 298
Registered: 10-2002

Rating: N/A
Votes: 0


Sunday, April 18, 2004 - 05:16 pm:

Hi hilarius,

I think your question do not have a clear cut answer.

I do believe certain patterns on chart do repeat itself occasionally, and the most reliable/profitable pattern ever is trend continuation pattern.

Yet, I cannot refuse that the market do act chaotic some other time and often gave us surprise. I believe this is going to continue forever.

That why both chart and money management is important to trading success. Use chart to identify opportunities, then money management is to protect and enhance profit.

Hi vermante,

I also believe no price based indicator can lead the market.

RSI do highlight the momentum changes thus make you feel it's leading the market. What happen is it will turn earlier than the market sometimes, but also gave much more false signals trying to act ahead.

Cheers,

Shaguar.

alsoran

Member

Post Number: 202
Registered: 02-2003

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 08:43 am:

The best leading indicator has to be the humble trendline.

The trendline is easy to apply, projects trend, provides clear & consistent reversal, continuation and exit signals, ... and works!

cheers,
A.

[Voice Over]
The Trendline -- Free with any purchase of Incredible Charts. (Batteries sold separately.)

david_louisson

Member

Post Number: 7
Registered: 02-2004

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 11:12 am:

Hilarius:

Not sure about NZ being a ‘progressive nation’ (esp following our recent showing in the rugby world cup), but I am fast learning that the IC forum is an extremely stimulating way of riding out insomnia. There have been some fantastic posts here, making it a great resource for a newbie like myself.

‘Wormald proposes that many (if not all) stocks have dominant angles, and that future price slopes will mimic past ones. Do you give this any credence?’

I’m not 100% sure what is meant by ‘dominant angles’. I should perhaps point out that I am aspiring to be a short term (3-15 day turnaround) CFD trader of UK sector indices, so I perhaps pay less attention to longer term patterns, where I suspect such ‘angles’ are more likely to have greater relevance. Moreover, the ‘angle’ idea conjures up ideas of Gann, or Elliott wave, type projections, in my mind.

Anyway, (if I may be permitted to return to basics), here is my roundabout response. I also seek to further your discussion on the selection process, in your second post.

MARKET PREDICTABILITY
Fear and greed are two of the most irresistible human emotions, emanating from which average crowd behaviour is the overriding factor behind any recurrent, cyclical, semi-predictable character that exists in the markets. In theory, the more actively traded the stock, then the larger the participant ‘sample size’, hence the more predictable the behaviour. There is also the concept of reinforcement (or ‘self-fulfilling prophecy’), where, if enough heavyweight participants believe that a price will top or bottom out at a certain point, and act accordingly, then it must necessarily do so.

MARKET RANDOMNESS
Offsetting this are unforeseeable economic and political events that seemingly cause impulsive and frequently erratic trader behaviour, and consequent convulsions in the price patterns, bringing us to the unpleasant conclusion that no amount of back-testing, however exhaustive, can ever be totally predictive. Probabilities in casino-type games can be precisely calculated; in the markets, they can not. Hence there is at best a flimsy basis for assuming that because an event has occurred x% of the time previously, that it will converge, even over a large number of trials, toward that same probability, in the future.

Future price slopes may on occasion mimic their immediate predecessors’, but my view is that the overriding patterns of self-correction are more ‘evolutionary’ than ‘recurrent’ in their nature, giving rise to the market’s changing ‘character’ and rhythms (hence a system can give excellent signals for an encouragingly long period, and then suddenly – and otherwise inexplicably – give poor ones). I am also conscious of my eyes’ tendency to be drawn excitedly to obvious chart patterns, while overlooking related ‘non-patterns’ that are equally relevant, but whose lack of pattern makes them indiscernible.

Perhaps the best we can hope for is to (paraphrasing Martin Pring) ‘uncover a trend, and then ride it until such time as the weight of probability suggests an imminent reversal’. It is one thing to ascertain the direction of the prevailing trend, but quite another to accurately project its gradient (‘dominant angle’?), or duration. Market analysis is an occupation that involves a vast number of vaguely interrelated parameters, complex imponderables, and projections that are at best mathematically approximate.

However, the fact does remain that past price/volume action is all that is available to us in our arsenal of technical data. Given that there are trends whose duration is long enough to overcome one’s costs, and that these occur frequently enough, one must ultimately profit. But, at least for me, the above highlights the need to operate conservatively – to seek to exit losers as quickly as possible, to mitigate risk by every means available, and size positions modestly enough to make it possible to fight another day.

Regarding stock selection, unless you have a reliable boardroom news source(!), the only way to guarantee picking the stock that will fly, is to pick all of them, when they give entry signals. Assuming that you are willing to manage several open positions, diversification is a great way of alleviating risk. Mathematically, there is no additional risk in trading (for example) 10 independent positions, each of x%, simultaneously, than consecutively, from the same capital pool. Everything else being equal, the outcome (death or glory) will simply occur up to 10 times more quickly. Although this is mathematically bulletproof, some very important caveats: (1) allowance needs to be made for the fact that ‘market risk’ significantly undermines the independence element, (2) leveraging time in this fashion gives one less time to react, and make adjustments to one’s system.

Shaguar : ‘RSI do highlight the momentum changes thus make you feel it's leading the market. What happen is it will turn earlier than the market sometimes, but also gave much more false signals trying to act ahead.’

I agree. Everything else being equal, indicators giving earlier signals will provide better price entries, but are prone to lower reliability. Momentum-based indicators, because they are based on rates of change, can tell us if a trend is accelerating (increasing in gradient) or slowing. Given that every trend must necessarily decelerate before it reverses, then this (provided it occurs slowly enough to allow perceptibility and reaction) could be viewed as an advanced warning of the reversal.

Conversely, however, the gradient of any trend is never completely smooth, hence there will be accelerations and decelerations along the way. A momentum-based indicator whose calibration makes it susceptible to these decelerations will, in these cases, give a false signal.

Just my sixpence worth...

Thanks for your posts, and good trading!

David

shaguar

Member

Post Number: 299
Registered: 10-2002

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 04:39 pm:

Hi David,

Another great post! I think both of us sharing the same view to financial market, and you just layout the facts and arguments so clearly.

Thanks and keep it up! ;-)

Cheers,

Shaguar.

perler59

Member

Post Number: 98
Registered: 09-2003

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 10:14 pm:

I find Best-Charts (B-C) very interesting as a mechanical trading system that tells you which indicator (and what settings) to use on which stock, and its free.

http://www.stock-anal.com/

Here is an explanation of how one person uses it to great advantage (Note that "BT" is Back testing and "AIO" is All Indicator Optimization):

What follows is a step-by-step guide to how to use B-C effectively. If this method seems too complex then B-C may not be the tool for you. However, if this methodology seems like something that you can handle, then B-C is probably the ideal trading tool for you, as it has been for me ("trading" is the operative word, positions may be held for as little as a week or up to a couple months, occasionally longer)

Step 1: Open B-C, click "Stocks", then select "Switch between 100 and 200 data"

Step 2: Click on the "BT" button.

Step 3: Select "88" for "Number of quotes for BT and AIO" (There are an average of 22 trading days in a month, thus you will be Back-Testing and running the AIO over the last four months of trading activity. Selecting a shorter time frame tends to produce less predictive indicators because they lack a significant-enough "look-back" period; much longer than "88" days and the indicators lose potency.)

Step 4: Select "closing price on signal day" (With this option, B-C will always be including the current price in its calculations of BT and AIO)

Step 5: Select "long and short" (Regardless of whether you intend to short only, long only, or long and short, because when optimizing indicators (AIO) B-C will tend to produce more predictive indicators with "long and short" selected).

Step 6: "Indicators" may be on any setting you prefer.

Step 7: Enter a stock symbol in the box next to the word "symbol" in the upper right of the window.

Step 8: When the information has loaded for the equity you selected, click the "AIO" button and select all the indicator options (i.e., MACD, MA, CCI, etc.), then click "OK".

Step 9: After the screen refreshes; click "Stocks" at the top of the window, then select "All Indicator Optimization", then select "Save Optimal Parameters", then click "opd.txt for daily quotes" to save your optimized indicators.

Step 10: Click the "BT" button. Click "OK" and a screen will open in your browser showing the results for each of the indicators in B-C. You should be primarily concerned with the Gain %/Trade column which indicates the average gain for each buy and sell signal generated (a sell is always taken into account with the most recent price even if a "short" signal has not been triggered). If the number in one of the boxes in this column reads "5.3%" then you should double that number in your head to read "10.6%" average gain per trade.

Step 11: Selecting the "best indicator" to use for entries and exits. There are many subtleties at this point. Sometimes an indicator will produce the highest average return but will manifest an unreasonable tolerance for lack of momentum (i.e.; the PVT indicator, which sometimes maintains a "long" or a "short" position for months even though the price of the stock is gently moving counter to the signal). In such circumstances, an indicator with the second highest average gain per trade may be more useful because of its greater responsiveness to changes in price direction. Typically, the best indicator will be one that has generated four to six trades (two to three changes of position) over the eighty-eight day BT/AIO period and generated the highest average rate of return. With some chart-reading experience, one can compare the indicator's behavior with the price changes of the equity. If an indicator would have kept you in a position for one to two months, during which time you could of had your money invested elsewhere, then that indicator, even if it had the highest average rate of return, might not be the best one to use for an upcoming buy or sell decision.

Bottom-line
The best time to enter a long position in an equity is when the best indicator, after optimization with the settings I mentioned above, indicates a "buy" from the bottom of its chart while the BBI (the top indicator in the window) is also at or near the bottom of its chart. Note: B-C is not a black box system (like Indigo or BlackBox.com) — it is better, if you know a little bit about technical indicators, reading stock charts (not to be confused with tea leaves) and market behavior.

hilarius

Member

Post Number: 39
Registered: 04-2004

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 11:02 pm:

Hello Perler

Once again my candle is burning as I sit all alone at the refectory table after all the other monks have gone to bed.

Now you add to my misery by offering yet another software package to be researched and analysed (only kidding)!

I can't resist new software ... but ....

I am already using :-

Incredible Charts
Metastock
Excel

Since the software you've recommended is free you obviously are not grinding a commercial axe ... and you are genuinely offering a tool for use ... could you just mention what it is (from your perspective) that Best Charts does best that Incredible Charts and Metastock do not already do?

Hilarius (needs new software as much as he once needed new socks before abandoning his vow of poverty)

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

hilarius

Member

Post Number: 40
Registered: 04-2004

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 11:06 pm:

Perler

I forgot to mention I also use CMC for real time charts

I presume Best Charts in offering "real time" analysis would have to rely on delayed ASX quotes if no licence fees are involved for the user?

How exactly does the "real time" bit work?

Hilarius

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

hilarius

Member

Post Number: 41
Registered: 04-2004

Rating: N/A
Votes: 0


Monday, April 19, 2004 - 11:37 pm:

Hello Perler

One thing which brings me back to Incredible Charts time and time again is the quality of the adjusted data which Colin Twiggs and the IC team provide. [I just noticed though that the adjustments for the ALN and CRT rights issues are not yet showing in my current IC charts ... even though I understood that raw data is no longer used?]

In Best Charts, as good as its features are, the lack of adjusted data for the dilutive effect of rights issues at discounted prices is a major concern.

At this stage then my vote is still for Incredible Charts as I imagine the adjusted dats for the 2 stocks I mentioned will flow through in due course.

Perler your views on this issue would be appreciated.

With Best Wishes

Hilarius

(Message edited by Hilarius on April 19, 2004)

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

hilarius

Member

Post Number: 42
Registered: 04-2004

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 12:03 am:

Ooops

I could be wrong .. the raw data in Metastock for ALN shows a high of 7.77 ... but the high is lower in IC so perhaps there is an adjustment ... it was just I was expecting a larger dilutive effect for the rights issue

Hilarius

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

perler59

Member

Post Number: 100
Registered: 09-2003

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 10:26 am:

Greetings Hilarius :-)

Best Charts is not a replacement for IC or Metastock. It uses free data from various net sources. For Australia I think it uses yahoo but I haven't bothered to snoop on its traffic to verify that. The main strength of B-C appears to be the quality of the end of day trading signals that it generates. I say "appears" because I have been very impressed with the gains it makes in its back testing after optimizing in hindsight, but I have not tested its ability when its optimized indicators are used to make live trading decisions. That is my next project. The continual re-optimization is an interesting twist and I'm keen to see if it produces favourable results. So, in summary B-C may be a good tool to aid medium term trading decisions, but is untested (by me at least).

david_louisson

Member

Post Number: 8
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 12:34 pm:

Dear Hilarius

I’ve just spent around 6 hours examining ‘Best Charts’. The software was so intuitive and navigable that I felt I learned a lot in that relatively short time. Here is a summary of what I discovered.

Installation (of v 4.32) was fast and simple. The download is only just over 1 MB.

The software appears robust and accurate, but what impressed me most was its versatility, and simplicity and speed of operation.

You need to be connected to the web while using it, as it downloads all of its OHLCV data (‘quotes’) directly from Yahoo Finance. That is a plus, because it means that you have instant, up-to-date data of stocks and indices from virtually any world exchange (including Aussie). These are instantly loaded when you type in the ticker of the stock you want analysed.

It uses only the following indicators (MetaStock and IC offer many more) –
* Moving averages
* Intelligent Technical Analysis
Bollinger bands / MA envelopes / Candlestick patterns with Volume
* MACD
* Stochastic
* RSI
* Commodity Channel Index
* DMI
* Best Charts Index (this is their own invention)
* Price ROC
* Elder Force Index
* Williams %R
* Percentage Price Oscillator
* Money Flow Index
* OBV
* Price & Volume Trend
Bullish/Bearish Indicator (also their own invention, this summarises those marked * above)

Once you have chosen the stock/index you want analysed, it displays all of the above as separate charts, vertically over each other, with buy and sell signals on each chart, and narrative on the RHS. Narrative includes whether the stock is currently trending up/down (bullish/bearish), current price/indicator values, and back-testing stats. This is very well presented – easy to read and assimilate. The charts are basic – you can’t zoom, add your own trendlines, indicators, text, etc, like you can in MetaStock.

Now, here is the interesting part. You can run OIA (optimise all indicators), and it will run an optimisation of all of the above, recalibrating them according to the optimum back-test profits. Back tests are run across whatever time period you specify. Then the charts are redisplayed with the optimised indicators, buy/sell signals, and the narrative updated. Instead of the optimised calibration, you can use other presets, or even define your own, and then re-run the back tests. A full back test of all 15 indicators across 4 months worth of OHLCV runs is less than 30 seconds on my AMD-1700 PC. You can save the optimised parameters, for later use in portfolios (see below).

You can run the BT (back-test) option to change the back test parameters, re-run, and then view the results in two tables. The first of these is a list of all trades made, showing entries, exits, gains/losses, and cumulative gains. The second gives a summary of indicator performance, including which one is delivering the best/worst return per trade.

Note that, in the back-testing, all gains/losses are compounded, and there is no provision to set manual stops (the exit of each long position becomes the entry for the next short position), have additional ‘set up’ criteria for entry/exit, or allowances for costs such as brokerage or spread.

All of the investigation I did was end-of-day OHLCV, but there is provision there for intraday analysis also. You can also switch from daily to weekly charts. The software links directly to Yahoo for market snapshots, company news, forums, etc.

There is also a facility to import OHLCV data from suitably formatted ASCII files.

You can set up portfolios of up to 20 stocks/indices, and run the same sets of tests across the whole portfolio. The output is excellent, it gives a tabular comparison of the 20 stocks, sorted by advance/decline probability, and showing buy/sell signals, the number of indicators that are bullish/bearish, and any candle patterns. Then there is a second table, showing the cumulative % gain/loss for each indicator (columns) back-tested across each stock (rows).

This is a great tool if you’re a fan of these indicators, running back-tests to optimise their calibration, and then viewing the profitability, and adjusted buy/sell signals. MetaStock will do much the same (I haven’t used IC), but not as quickly and seamlessly. You are limited to these 15 indicators, however.

As Perler points out, this can be used as a mechanical tool, but given that you can adjust the parameters, and see how this adjustment affects the results, it is not a black box. Using daily charts, the buy/sell signals generated would suit a swing rather than long term trader.

I feel that this is an excellent tool for a newbie who wants buy/sell signals generated automatically and in user-friendly fashion, without having to learn the intricacies of system testing in a more full-fledged product like MetaStock. A more advanced trader can vary the parameters, to some extent, to suit him/herself.

There must have been a significant amount of development work go into this product. I can’t believe that it’s freeware. I’m sufficiently impressed with its speed and simplicity that I intend to use it as an adjunct to my existing MetaStock-based trading system. Many thanks, Perler, for bringing it to my attention.

Hope this is useful.

Cheers
David

new2waix

Member

Post Number: 720
Registered: 10-2003

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 01:09 pm:

Thanks Perler for the link.

I am just playing around with Best Charts and so far am impressed.

Twenty minute delayed data, indicator analysis, backtesting all for free!!! Looks like a very useful tool in testing indicator setups.

http://www.stock-anal.com/

Maybe there are some ideas IC can borrow from Best Charts?

:-)

david_louisson

Member

Post Number: 9
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 02:43 pm:

Agreed, ‘Best Charts’ is an interesting and powerful tool, for the reasons stated in everybody’s prior posts.

A few points, though –

1) The back testing produces some impressive looking gains, but remember that no dealing costs are being considered. Where the gains made by different indicators are similar, the one achieving it with the lower number of trades is likely to incur dealing costs less frequently, and therefore deliver a superior real-life profit.

2) The default option is to use today’s Closing price to generate an entry signal, and the way in which the trading gains are calculated assumes that entry is made at this signal price. (The same logic applies to exits). Unless one is trading intraday, today’s published data can not normally be acted upon until the markets open tomorrow morning. In this situation it would perhaps be more true-to-life to change the ‘Buy/Sell Price’ option to ‘opening price the following day of the signal’. I suspect the ensuing overnight gaps would, on average, and everything else being equal, have the effect of deflating overall profit.

3) Apart from Best Charts' own proprietary indicators (BBI and BCI), most of the other indicators can likewise be back-tested for parameter optimization in other charting products (e.g. MetaStock). Agreed that Best Charts puts it all together very effortlessly and seamlessly, but where the non-proprietary indicators are concerned, there is no real ‘magic’ attached to their buy and sell signals, that can’t be achieved (without a little hard work) by back-testing elsewhere.

Hope this makes sense.

David

hilarius

Member

Post Number: 43
Registered: 04-2004

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 03:00 pm:

Hi All

Agree with everyone's conclusions about Best Charts ... it has seamless chart production, backtesting and ease of use which are great benefits ... and impressive development though IC is also exceptional in the development area ...

I remain concerned about the data issue ... and this is where IC scores so highly with its premium data ...

Looking forward to hearing whether the adjusted data for the ALN and CRT rights issues will be in the IC charts soon ...

Any word on this Robin or Colin?

With Best Wishes

Hilarius

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

blondpanic

Member

Post Number: 42
Registered: 10-2003

Rating: N/A
Votes: 0


Tuesday, April 20, 2004 - 10:14 pm:

Hi everyone, I have been fiddling about on best charts on and off after someone mentioned it here some time ago. The thing that amazes me is how quick it is. i have a slow dial up connection and play endless games of solitaire while waiting for my charts to load on IC but on B-C the multiple charts appear almost instantaneously. Must be very smart programming and it's fun to use.
Regards, BC

cjb

Member

Post Number: 31
Registered: 02-2003

Rating: N/A
Votes: 0


Wednesday, April 21, 2004 - 12:18 am:

Hi All,

I had a look at B-C for an hour or so and though it does look interesting, for me if I was to rate as a tool to help me trade.... I would put it somewhere between dangerous and useless. What is the point of running a bunch of stock through such a system and having a different strategy for each one? It appears to me that they are trying to do is "curve fit" indicators to stocks. As we all know, stock price action does not remain the same forever. How would we know that the indicator chosen will perform as expected into the future? We don't and my guess is it wont. It does have some value as a training tool ie looking at how different indicators perform on different types of stocks and the software is kinda easy to use though the user interface could do with some work. It also crashed on me twice.

I am still interested in finding out what others think about the original topic of this thread. Can an indicator lead the market?? From my understanding, if I have a signal X, the maximum information I can find out about X is X. If I decompose X into different parts it can only tell me what has happened to X. Decomposition cannot add information therefore I cannot know anymore about X than what X is telling me. Therefore I would think any derivative of X cannot lead X. Is this to say there are no leading indicators?? Of course not. I know that if the Chinese are buying Iron Ore the price of BHP will go up. So, in my opinion if you want leading indicators you must look else where....

This poses the question. Is price action in itself a lagging indicator?

CB

david_louisson

Member

Post Number: 10
Registered: 02-2004

Rating:
Votes: 1


Wednesday, April 21, 2004 - 08:06 am:

CJB

BEST CHARTS
Re curve fitting indicators to stocks – yes, I believe that is exactly what BC sets out to do. If one (1) places faith in indicators (the original discussion point in this thread), (2) believes that different stocks’ price action displays different character, and (3) believes that this price action changes its character over time, then tailoring different indicator-based signalling systems to suit different stocks, and then constantly re-optimising indicators to fit their price action’s most recent cyclical rhythm (calculated objectively by profitability back-testing), makes some kind of sense. BC allows one to re-optimise frequently and easily, perhaps supporting your point that ‘stock price action does not remain the same forever’.

Some traders apparently disagree, in that they stick with the default calibrations (e.g. 14-day RSI) supplied by the software retailer. I guess one could run profitability back-tests of adaptively optimised indicators versus those calibrated with a constant value. Over a vast enough sample size, the results would likely show us the extent to which the ‘most recent character’, relative to ‘general character’, repeats itself in the immediate future.

I am certainly not touting BC as any kind of Holy Grail. For, even assuming that constant tailoring and re-optimisation does have a practical (i.e. profitable) value, I’m sure there are other applications out there that do a similar job. I was simply impressed by its user-friendliness, smart implementation, and the fact that it was freeware. I’ve been looking for an adjunct to my existing system that uses a fresh perspective; BC will fill this gap in the meantime.

LEADING INDICATORS
I think Hilarius hits the nail on the head when he asks what is meant by ‘leading’. CJB, I agree totally with your logic resulting in ‘any derivative of X can not lead X’. All this aligns itself with the point that Joe Ross was making in his original post.

If, however, we take ‘leading’ as meaning ‘in some way predictive of future price action’, which is Hilarius’ point (assuming I understand correctly), then the question becomes something like ‘in what manner, probability and extent, does any price action, but especially the most recent price action, repeat itself?’. Believing that prior patterns (‘trends’) will in some way extrapolate themselves into the future is a keystone of technical analysis. Without it, TA has no basis.

Hence my conclusion is this. Given that all indicators (and trendlines, etc) are derived from raw OHLCV data, and that they can highlight or summarise aspects of this data that might otherwise be overlooked, then they are not only useful in this context, but are potentially as predictive as the price action itself is. However, indicators can never be more predictive than the raw data, and I would suggest that in certain cases their predictive potential is ‘diluted’ by their construction and calibration, i.e. the derivation process.

Good trading!
David

pksam

Member

Post Number: 1
Registered: 01-2006

Rating: N/A
Votes: 0


Sunday, August 27, 2006 - 06:25 am:

I think that indicators are neither leading nor lagging , but rather its the way of reading them that makes them leading or lagging.
Eg.: 1.A divergence in momentum can be taken after a confirmation or on on a early but risky premise.
2.MACD could be used only on a crossover of the lines or on the first downturn of the MACD histogram or coupled with divergence(much less lag here)... the difference lies in the enhanced risk.
3.Instead of a 5,3,3 stochastic a 5,1 stochastic can be used to anticipate moves. The time advantage is offset by the enhanced risk.
Exception: The only indicator I've seen that offers early signals is MESA. But only if you have means of filtering out the current cycle can you reduce the risk on following MESA signals. Else MESA signals hold good just 20% of the time.
So, I believe that most indicators can be used with or without lag, based on one's appetite for risk. The difference lies in one's preparedness to anticipate or react. Therein lies the choice and the difference.

Incredible Charts
Free Charting Software

Auto-Fit Trendlines

Trend Channels

Linear Regression
Channels

Raff Regression
Channels

Standard Deviation
Channels

david_louisson

Member

Post Number: 252
Registered: 02-2004

Rating: N/A
Votes: 0


Tuesday, August 29, 2006 - 03:26 pm:

Hi Pksam

Welcome to the forum.

For whatever it's worth, my take on the math that causes indicators to 'lead' or 'lag' here:

http://forum.incrediblecharts.com/messages/9/661670.html#POST86226

Best wishes
David