Theory on the first creation of indicators

Lets assume that in the beggining, a few traders decided to play around with prediction of future price movements based solely on price itself.

After not long, they understood that methods forecasting changes of variables in the future based on the variables themselves already exist. They found them in meteorology, predicting earth quakes, rain etc...

Those forecasters uncovered that it is possible to anticipate the most likely changes in a variable based on the nearest past states of the variable.

Why did those methods work in the markets ? Its not about why or when someone was buying or selling. It was simply that the nature has quantifiable correlation in changes between variables.

Now assuming this did take place, what exactly did those traders do ? What was available to them ? Formulas and concepts from maths and statistics. After all the changes in price were nothing more than a data set.

So what did they first use ? We can crack this together.

The data set of price changes, were sampled in to time frames/ranges/etc. Its all about what goes on the vertical Y, and horizontal X scale. This gave the ability to start quantifying the "random" price changes.

Im assuming that the HLC bar was the first successful data set.
On the range of which quartiles were applied along with standard deviation. I really need everyones critique, opinion, knowledge !!!

Quartiles probably gave birth to the so called "overbought/oversold"

Indicators don't work.

It's an illusion, created by the creators of the markets itself.

It's made to make you think it works, to make you believe in something but you can never prove it...just like religions... the real purpose of the indicators is to keep you believing,trying... suck money from you... then you tell yourself "Why me"......well just the religion again.

Good luck figuring it out.

Hi,
If you read about speculative trading on markets 100 years ago and more, you find there were no charts or any of the various indicators we take for granted now. The price feed was a ticker tape and the core skill was thought to be good tape reading to pick up the flow of buying and selling from the printed trade record. The most widely used frame of reference and analysis was to the high and low of the previous day, and not much else.

It was really only from the 1920s on that hand maintained daily charts were used to any extent and more rigorous statistical analysis of price was pretty rare until the 1960s, no small part due to the lack of computing power available to the average person. I believe the first widely used derivative indicator was moving averages to try to highlight underlying trend. As trading became shorter and shorter in time horizon the statistical analysis of regression and mean reversion started becoming more widespread and things quickly expanded from there to what we have today. And I believe it was Japanese rice traders who created candlestick charts to routinely analyse and use chart patterns.

I'm only replying because I saw your message in the chess thread. I'm not a quant trader therefore I can't contribute much. Some traders better equipped to assist you with your research might be Copernicus, FXEZ, PipMeUp, algoTraderJo, olsen-yersen, Sis.yphus, ezcurrency, 9047, kprsa, stt, mikkom (if he's still around)..... and no doubt others who've slipped my mind. Good luck with your research.

Thanks allot for your replies ))

Price Action vs Indicators;

from a mathematical point of view; everything is derived from price; even OHLC bar is indeed an indicator; it is derived from the tick by tick data, and time.

indicators do work; and I am making a living from them myself.

take for example; moving average; the most simple indicator of them all. It simply works. you just have to design a system with few rules; and play with them till you find something that works for you.

see this thread for the strategy rules

Hi Max. The way I tend to treat this problem is akin to a problem with thermodynamics. The activities of participants in a complex system such as the markets gives rise to GROSS statistical features of the markets themselves such as the overall volatility and market structure. It is not the individual buy/sell and trade size decisions of each participant that gives rise to the Gross statistical features of the market but rather the collective activities of participants that do.

This collective action gives rise to gross features of price such as divergent and convergent momentum just like the waves in the ocean and these features tend to become more 'enduring' when participant behaviour becomes coordinated....but more often than not, the vast differences in participant behaviour give rise to apparent features that are simply no more than a random collective summation of these activities. Indicators inherently rely on a particular set of circumstances to unfold in relation to price to indicate a future predictive outcome....but they are often too prescriptive in nature. Emergent features of collective price action can take many forms.

It is more a game of constructive and destructive interference when you sum up all the activities of the participants just as the overall features of a cloud are governed by the summated activities of the mechanics of it's individual water molecules. Market structure is an 'emergent' feature of collective participation. More often than not the price action is a pure random outcome based on the summated constructive and destructive impacts of the participants.....but occasionally the emergent feature has 'real' substance that persists for an extended period and reflects overall participant behaviour. Price action itself holds no value to me as a trader.

If we base our trade forecasting decisions on price action alone we will find it extremely difficult to achieve a result that is anything other than a random outcome.....but if we base our forecasting decisions on the mechanics of all participants over a period of time reflected by Gross features, then it is more likely that we will be able to jump on board the enduring features of the market when they are present. That therefore provides a way to capitalise on the gross features by only taking part in a trade when these activities are more likely to be more coordinated than the everyday chaotic behaviour. For example if I was trading the weather I would avoid the normal days and only trade during the storms when the gross ensemble are more likely to behave in a coordinated manner. Some of these storms will turn out to be cyclones/hurricanes and this is where your bread and butter will be borne. All other activity will simply be a churn and random outcome.

Now everyone knows when a storm is upon them and these storms take many forms and shapes.....but it is almost an impossible task to predict them. That's why I prefer to 'follow price' rather than 'predict it'....but I only enter a position when a storm is right on top of me. I do not take a future position when it is normal weather. I rely on the principle that the storm has durability and persistence for a period of time of unknown extent. The path you take when riding the storm can be varied so you need to scope out the limits of that price path to ensure you remain with the storm. Those limits define the broad features of the storm itself to be able to define what is 'in the storm front' and what is 'out of the storm front'.

Meteorologists base their forecasts on gross statistical measures such as pressure, humidity etc rather than the individual mechanics of the air molecules......and in the same way for the markets we should be looking at gross statistical measures such as volatility, momentum, market structure etc as opposed to the infinite ways price action can unfold.

To be able to ride the storms you need to be present for those unpredictable rare opportunities. To be in this position you need to defend your capital base at all other times and also be diversified. This means that you need to be very selective when you trade as the random chaos will progressively chip away at your capital.

Randomness means that simply with luck alone some may be profitable and some may not but with the Law of Large numbers and the frictional costs of trading, time will get you in the end. To be a trader you must accept that the markets are not Gaussian and that they possess fat tails. If you do not believe this philosophy then you may as well pack your bags and join a new profession. The fat tails that exist in price distributions occur in any timescale and it is your task to identify when there is the greatest likelihood that fat tails will arise outside of the normal everyday churn.

The way I tackle this problem is as follows:
1. Identify periods when it is more likely that participant behaviour is coordinated. This will be when market conditions are NOT normal. Use gross statistical tools such as volatility, market structure and momentum to define these zones rather than price action. Look for zones outside normal volatility limits or periods when major news events will trigger a major coordinated move. By avoiding the 'normal' churn you will tend to avoid 'random features' that have no durability of forecasting potential;
2. When you have defined the zone within which participant behaviour is more likely to be coordinated than not, then treat every signal in that zone as valid and take them all. Avoid the propensity to predict as emergent features in these abnormal zones can take many forms. Take every trading signal using a systematic approach to ensure that you do not miss the potential whales;
3. Abnormal market conditions are by definition unpredictable and rare events so diversification is your key to increasing trade frequency to lift your trade activity. Rather than chasing local storms, become a storm chaser of the world to bring home the bacon; and
4. Strictly manage any downside risk but leave your profit targets open to allow for infinite yield.

I hope this helps a bit mate and also hope I haven't taken your thread too off track :-)

They probably said...

"I wonder if it's above or below its average price..."

and then, "is it above or below the body of the previous candle".

interesting question maxd !
I see it this way: any indicator including "candlesticks" is a way to "smooth the noise". tick data is just too large hence difficult to interpret !

Yes mere statistics are used to predict the weather in the long term.

except that the market changes so frequently.. that we are more like pioneering rather than making actual predictions.

I like it.

Putting theory to practice. Have you made your first million trading fat tails?

My Econometrics professors use to always tell students they predicted this and that crash through their modelling but made no money.

I dropped out of their classes

Thanks allot for your reply )
The analogy is golden.
I simply don`t have anything to add since you covered it from a to z

Why the average exactly though ? And not the median for example or mode. Was it only because it is the easiest to calculate, or because it was the average which was used in the calculation of Standard Dev ?

Also is there an exact term or method used in statistics which deals with above/below average ? I`m a total newbie at it unfortunately at the moment.
The problem I`m also facing with the average is that at least to my understanding its usually close to the median. And is affected by extreme values, and adapts to them which kind of hides their importance. I assume that there is something similar to correlation coefficient but anchored to the average.





Why just the body though ? For what reason are the wicks disregarded ?
And I guess the body of the previous candle could be easily a reference range. Movement away from which could be quantified using percentage as an indicator.

How I see it;

A chart itself, is already an indicator by itself in the purest sense. Tick, line, time bar, range bar, renko, Jchart...etc

The core business of indicators is to organically organize the price/time data in such a way, that it reflects the approximate reality of the market, into summations (hopefully with workable edges) that the limited and perhaps lazy (but smart in a way) human brain could understand quickly, and make logical decisions from it.

those who claim they predicted a crash; have survivership bias. If you dwell enough; you find that they predicted 100 things; and only 1 event happened. Thus they are 99% wrong about their models.

These kind of educators are everywhere; educate with no experience

Who cared a dime about HLC bar??
How about OHLCC bar??

The first traders, probably in Rome, Mesopotamia, Atlantis... used demand charts!!!
I'll explain!!!
The Nile river is down a lot, so wheat harvest is weak....so price up!!
Simple.

Just like today with price of oil - more wells - drop the price.
More green stocks -more oil demand. See the chart Friday afternoon 15:30 GMT.

Nothing new on this little planet

But still on demo ?