The Contents of a Bar

[ducati916]

It was this [partial] post that got me thinking, as there are a number [lots?] of systems traders out there:

I don't think I curve fit and use price action patterns as a basis for all my code , the majority of my indicators are custom written by myself ( why I will not divulge or discuss ) and represent tangible price action events . Median reversion is the basis of my systems and I filter what I see as important aspects of price action . The counts used in charts here are only a part of the process but they do measure trend and reversals and are integral part of what I do on every time frame .

So my question is: is every bar created equal?

The answer [to my mind] is clearly no.

Some bars contain nothing but trading noise, whereas other bars contain important information.

Will the market [the sum of all participants of a given bar] treat the same information in the same way as they did previously?

Again, my answer would be clearly no. The participants could, and likely would be, completely different, with the commensurate difference in subjective views. That is just one example, there are hundreds of reasons why the reaction could be different.

Therefore, building a trading system/methodology, based on historical data is prone to randomness, exactly what you are trying to eliminate.

Is this a futile undertaking?

jog on
duc

 

[Triathlete]

It was this [partial] post that got me thinking, as there are a number [lots?] of systems traders out there:

I don't think I curve fit and use price action patterns as a basis for all my code , the majority of my indicators are custom written by myself ( why I will not divulge or discuss ) and represent tangible price action events . Median reversion is the basis of my systems and I filter what I see as important aspects of price action . The counts used in charts here are only a part of the process but they do measure trend and reversals and are integral part of what I do on every time frame .

So my question is: is every bar created equal?

The answer [to my mind] is clearly no.

Some bars contain nothing but trading noise, whereas other bars contain important information.

Will the market [the sum of all participants of a given bar] treat the same information in the same way as they did previously?

Again, my answer would be clearly no. The participants could, and likely would be, completely different, with the commensurate difference in subjective views. That is just one example, there are hundreds of reasons why the reaction could be different.

Therefore, building a trading system/methodology, based on historical data is prone to randomness, exactly what you are trying to eliminate.

Is this a futile undertaking?

jog on
duc

You bring up some very good points and I also think along those lines..

I am a totally discretionary trader and have no experience with systems.

My thinking around systems is that they are made for short term trading as I cannot see how the systems would be of any benefit if your trading was over the medium term, other than generating more trades and taking the human emotion out.
The other point with systems is getting caught in a large drawdown.

A discretionary trader would have an advantage over the longer timeframes as they are more likely to position trade and the system would need to see each bar before it gives out a signal.

I would also think that you would need different systems to trade in different styles of markets.

Can the system traders make some comments on the timeframes and holding period that they trade using systems...Thanks

 

[tech/a]

My current view is to have the program identify price action which is tradable then rate and implement it

This is an on going eduction process for the program on a chart by chart basis

As not all bars are equal
All charts are not equal
All patterns and volumes related to price bars and groups of bars are also not equal

Computers can and do find things HIDDEN to us
Some of the results I see are lightbulb moments

 

[Quant]

My current view is to have the program identify price action which is tradable then rate and implement it

This is an on going education process for the program on a chart by chart basis

As not all bars are equal
All charts are not equal
All patterns and volumes related to price bars and groups of bars are also not equal

Computers can and do find things HIDDEN to us
Some of the results I see are lightbulb moments

100% + , these conversations/debates are ones I will not participate in for " Is this a futile undertaking? " is not a relevant question to me , I know its not futile and I have no need to convince anyone either way . I am on the right path and no-one can tell me or convince me otherwise

And exactly what Tech says all bars are not equal with many irrelevant , low volatility ( small range ) inside bars are a prime example , the key to this whole price action algo type thing is measuring/filtering the things that matter to help disregard price that doesn't matter .

What is meaningless chop ?
Where are the most reliable signals produced ?
What are the rules to help define edge ?

Systems traders find solutions

Statistical analysis can and will produce an edge .. I think that's beyond debate

I know what I wanted to define when I started this journey . I wrote a bullet list and I worked tirelessly to breakdown each point and work on solutions to measure/define and to assign a set of rules to get an empirically proven edge .

Discretionary traders aren't as discretionary as the word implies , all the successful discretionary traders have rules even though they may not be used in a binary sense . Binary removes doubt , it improves performance , its measurable and therefore easily analyzed . If you can measure it , its easier to improve it .

I believe as an ex price action discretionary trader I have an extreme edge over the average number crunching curve fitting algo writer . Mathematize your thoughts , where it leads is mindboggling , nothing will get you closer to the illusive " holy grail " . When x , y & z happen I have an 80% chance of having a positive outcome with outstanding expectancy is as close as the holy grail gets . There are no absolutes

In any of these threads don't be put of by other peoples expressions of their limitations . Its not impossible and it is not easy

Id rather spend 5 hours on systems building than spend 5 hours discussing whether systems trading is effective . I already know .... good luck to all ... rock on

 

[Triathlete]

Mathematize your thoughts , where it leads is mindboggling , nothing will get you closer to the illusive " holy grail " . When x , y & z happen I have an 80% chance of having a positive outcome with outstanding expectancy is as close as the holy grail gets .

Then I must be at the holy grail for the time being.....my x , y & z = price , pattern & time...Looking back at my results over 3 years:

Win % = 84.9%
Reward to Risk =1.96
Expectancy ratio = 1.52

over 12 months
Win % 85.2%
Reward to risk = 2.26
Expectancy ratio = 1.77

Happy trading....

 

[rb250660]

What's your average holding period Tri?

 

[Gringotts Bank]

Then I must be at the holy grail for the time being.....my x , y & z = price , pattern & time...Looking back at my results over 3 years:

Win % = 84.9%
Reward to Risk =1.96
Expectancy ratio = 1.52

over 12 months
Win % 85.2%
Reward to risk = 2.26
Expectancy ratio = 1.77

Happy trading....

wow, nice work.

 

[ducati916]

My current view is to have the program identify price action which is tradable then rate and implement it

Your words imply a single program.

I am however assuming a number of different screens within that program identifying different types of price action. Is this a fair assumption?

This is an on going eduction process for the program on a chart by chart basis

This implies some form of AI, where the machine learns. As far as I know, this is not yet possible.

As not all bars are equal
All charts are not equal
All patterns and volumes related to price bars and groups of bars are also not equal

We can agree on this.

Computers can and do find things HIDDEN to us
Some of the results I see are lightbulb moments

Here is the crux [sort of]:

(a) Historical data is the past;
(b) it is 1 realised path;
(c) there are billions of possible paths that are unrealised

These statements are also true.

Therefore:

(a) the content of a bar contains the market reaction to 'X';
(b) which occurred [the bar] at that time; and
(c) no-one knows what 'X' was [which can include of course market reactions in of themselves]

Some 'Xs' are important, many are not.

Successful systems traders have at a point in time, identified correctly, the prevalent pathway in the market. An easy example that I have seen recently in various threads is 'reversion to the mean'.

Having quickly glanced at the ASX chart, the ASX has exhibited exactly this quality for a number of years, thus, programmes that identify this quality, should be successful.

Markets change [they still go up/down/sideways] over time periods, reflecting [a different] 'X'.

Are the following statements true?

(a) How quickly the trader adapts is a function of the trader.........not the programme.
(b ) How the trader manages risk is a function of the trader..........not the programme

I saw later in a reply '85% expectancy. This is high. This will not [one assumes] be discarded lightly. Some thoughts:

(a) the use of a '%' indicates a sample size [in this case total trades taken] which is 'n'
(b) the expectancy is calculated inductively, therefore realisation of the new 'n' advances only at the square root of 'n', viz more slowly.

Even with excellent risk management [a skill in itself] a very successful programme can turn to custard overnight, but not be detected for a long time. Would you agree?

jog on
duc

 

[ducati916]

Id rather spend 5 hours on systems building than spend 5 hours discussing whether systems trading is effective . I already know .... good luck to all ... rock on

Therefore I shall respect your wishes and leave you in peace to pursue your interest. I didn't wish to be rude and simply ignore your [initial] response.

jog on
duc

 

[Triathlete]

What's your average holding period Tri?

The 3 year results based on 90 day average and the last 12 months has been 19 day average

 

[barney]

The 3 year results based on 90 day average and the last 12 months has been 19 day average

Nice work on those stats "Tri"!!

To "Duc" ..... Nothing really concrete to add to the thread, however to state the obvious, the importance of the "bar" being traded needs to be studied in context with the idiosyncrasies of the instrument being traded .... particularly the "time of day" the bars in question are being printed.

I am a great believer that everything in relation to "Time" is the key to the market's mystery

 

[CanOz]

This guy uses quantitative analysis to generate signals and discretion for context to execute trades ... he's pretty much explained away my edge as well as his. Chat with traders @HF_Trader

 

[ducati916]

I am a great believer that everything in relation to "Time" is the key to the market's mystery

Market data is historical.
Trades are taken in the present.
The past, does not predict the future.

Therein lies the dichotomy.

This is why mechanical traders talk in terms of probability and expectation.

Asymmetry in the market deals with probabilities and expectations. Probability is nothing more than a math calculation that tries to deal with uncertainty or the unknown, which is of course the future.

It is the asymmetry that seems to be inaccurately calculated, but on this I may be wrong, which is why I was interested in generating this discussion.

jog on
duc

 

[Quant]

Duc you seem to have tried to become systematic , what did you measure and how did you measure it ? Did you find any measurement that had potential ? Surely you found something that had a modicum of edge . What would you consider a significant edge ? Many people find an edge but are put of by the amount of transactions at a small size to extract the profit out of a marginal edge .

 

[tech/a]

Fuzzy logic

Computers will find common price action whether it be a single bar or a combination of bars,range and volume
It needent be perfect but it will reflect
Behaviour in the instrument.

Just as you can calculate crowd behaviour
In a given space so you can in a given instrument.

Like Quant I'll leave you here
I know as he does

 

[barney]

..... which is why I was interested in generating this discussion.

jog on
duc

And a great discussion it could well create Duc!!

In my haste I probably missed the true essence of what you were questioning initially(Nothing new lol) ....
Before I babble any further ...... the "partial post" you originally quoted ... was that yours or someone else's? Cheers.

 

[ducati916]

Computers will find common price action whether it be a single bar or a combination of bars,range and volume
It needent be perfect but it will reflect
Behaviour in the instrument.

It will.

It will all be historical. Which is my point.

jog on
duc

 

[ducati916]

Just as you can calculate crowd behaviour
In a given space so you can in a given instrument.

You are not calculating crowd behaviour, which implies a 'prediction' of the outcome, viz, 2 + 2 will if added, equal 4.

You are merely observing the numerical volume that actually transacted at a given price, at a given point in time. Past tense. Again.

jog on
duc

 

[ducati916]

Like Quant I'll leave you here
I know as he does

And that is of course fine.

jog on
duc

 

[ducati916]

Duc you seem to have tried to become systematic , what did you measure and how did you measure it ? Did you find any measurement that had potential ? Surely you found something that had a modicum of edge . What would you consider a significant edge ?

I do consider myself systematic, just not possibly with quite the same definition that you may have.

What do I measure?

Now this is an interesting question. My immediate response would have been nothing, but that is actually inaccurate. I measure only when I enter a new trade and I measure:

(a) the historical P/E ratio; and
(b) that the ETF is well diversified; and
(c) the capitalisation; and
(d) the average daily volume

So a mix of some fundamental factors and 1 technical. I have to accept going forward, that these will change, perhaps in a material way. This is a risk that I manage, so this if you like is an ongoing measure.

After the trade is entered, I only observe price in real time, ie, price that I can transact at.

The reason for this is that I no longer rely on the security trending higher/lower/sideways. Any combination of the 3 possibilities will suffice.

In this way I embrace the [incomplete] randomness of the market.

I cannot manage a complete blow-up of the security/bourse/country/currency/etc. That is a black swan beyond my control.

Therefore I diversify my [total] wealth away from purely financial markets and have land/job/ to manage that risk....as best you can.

What is a significant edge?

A significant edge to me is an edge [as you may have gleaned already] is an edge that is independent of market conditions at any given point in time. It must be robust and embrace either high or low volatility, trends or chop, news or no news, etc.

The edge must be scaleable and therefore the edge must be able to compound if required.

jog on
duc