Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Timeframe inside current timeframe - is that possible? 2 replies

Seeing higher timeframe indicators on lower timeframe 4 replies

Oanda not of the preferred brokers? 3 replies

Simple, VT preferred 0 replies

CMC/preferred broker 4 replies

Attachments: Preferred Timeframe for Trading
Exit Attachments

Preferred Timeframe for Trading

Quoting PipMeUp
I never said anybody had read his thread. Just that a lot of traders trying randomly all the possible timeframes will converge towards the optimal by trail and error. Call it Darwinism if you like. .
Maybe there is a smoothess which is subconsciously detected by the majority after detailed analysis and watching of currencies for a long time.

But then if H4 is the most optimal, and most of the traders use it, then why does the majority of traders lose ? There is a contradiction here.

Evolution is about adaptation to the enviroment, this implies that retail traders evolve into losing machines?

Quote
{quote} I find you cast a lot of your own believes onto other people. Personally I never used a NN. Especially not a "very very complex one" which would be the best way to overfit any data. I don't know for the others but I don't try to predict any bar and I don't optimize parameters. Egyptian certainly used rational reasoning to approximate π like Babylonians and Greeks after them did. They didn't pick random formulas

Yes i do, after all what is the point in having a forex forum, if not for the sharing of opinions with eachother.

I was talking about professional quants in Wallstreet, almost all HFT firms and HF use some sort of NN which is semi or completely automated.Most hedgefunds which operate on mathematics instead of economics, dont really have traders there, only mathematicians, quants and PC scientists who program the algos and set up the hardware.

Overfitting i dont think is an issue for them, they are certainly more qualified to handle NN technology than any of us in this forum.

Quote
Egyptian certainly used rational reasoning to approximate π like Babylonians and Greeks after them did. They didn't pick random formulas

Yes , but i didnt talked about the ancient methods where you measure the circle's circumference (approximately) and then divide it by the diameter, which is a very non precise way, but the methods in calculus that were used to calculate the 6000th digit of pi accurately.
There are many many formulas with less and less margin of error for it, and now with modern technology we can calculate the 1 billionth digit for it precisely, which you obviously cant if you use a ruler to measure the circle's properties
"There's a sucker born every minute" - P.T. Barnum
Please folks especially newbies learn and train yourself to look at higher time frames that is daily chart , weekly chart and monthly charts.then use lower time frames to get into trades.thiswill help you to spot trends assuming you have a system.
Quoting Proximus
{quote} Yes indeed, that is the problem, because there is 1 formula for the MEAN of the market which is measuring it perfectly, the problem is that it is unknown.Heck i believe there is a formula that can measure anything with some degree of certainty, there is none that can eliminate uncertianty completely, but there should be one which is the most optimal. So with the markets, there is a formula that can predict X+1 bars ahead always, the problem is that nobody knows it, if you knew it then you would be a billionaire and make sure that nobody...
Assuming that one believes that Markets are driven by trader feedback,

Mathematically, the market is recognized as a "non-linear, non-homogeneous differential equation with non-constant coefficients", impossible to solve in closed form (basically insoluble).

Looking at the complex market from a differential equation standpoint, a descriptive equation would likely be nonlinear (feedback does not behave linearly), non-homogenous (variables most likely could not be separated, even if they could be defined) and the coefficients are non constant (an increase of ten percent in the soybean crop would rarely translate into a change in price of ten percent). It would be most unlikely that a complicated differential equation could be translated into a distribution function.
Overarching mathematical formulation of a market distribution is not possible with current market understanding and mathematical capabilities In view of the complex nature of auction markets, a closed form solution may never be possible.

Markets are driven by Trader Feedback. Mathematical “manipulations” explain little or nothing about the market itself.

Complex systems concepts have only recently begun to be applied to financial market analysis. Complex systems are generally nonlinear with feedback acting to continually adjust the system.

The non-stationarity and feedback aspects of the Market argues for markets to be "non-linear"…………
So why would you want to view it or analyze it that way? (Time series chart)

Price is a linear, one dimensional variable. Whereas "Value" is a two dimensional variable (price over time). ("price over time "and "time at price” are different concepts)

Problem: Time series charts show large scale but you trade in the here-and-now. This gives you a time warp. TF's modify and distort the data.

"Linear thinking” makes an incorrect market assumption of periodicity. TF's imply the existence of an underlying periodicity. (moving averages, same thing)
Charvo made a great point……it depends on your Methodology. Some methodologies, the concept of TF’s would not be an allowed “practice”.

May I ask in which camp do you belong?

Do you subscribe to the idea that the Market is driven by Trader Feedback?
Or
Do you believe the Market is driven by “Mathematical manipulations”?
Markets are not efficient, rather they are effective - Jones
Quoting mzvega
{quote} Assuming that one believes that Markets are driven by trader feedback, Mathematically, the market is recognized as a "non-linear, non-homogeneous differential equation with non-constant coefficients", impossible to solve in closed form (basically insoluble). Looking at the complex market from a differential equation standpoint, a descriptive equation would likely be nonlinear (feedback does not behave linearly), non-homogenous (variables most likely could not be separated, even if they could be defined) and the coefficients are non constant...
Price itself is driven by 2 main things, that is the liquidity provider book balancing when the net orders are in a big unbalance due to excessive acumulation of one orders, so the price balances itself to a more stable place, in layman terms called mean reversion,crash,retracement,etc.
Secondly it is moved by events, such as news, which change the fundamental enviroment of the market => and the human psychological reaction of the traders to this.

So basically its 3 basic concepts, however from a more aggregate position its simply the imbalance of SUPPLY & DEMAND, if the SUPPLY>DEMAND then price goes down, if SUPPLY<DEMAND then price goes up.

Price as time series, contain all the action that has happened in them, so you dont need to find the value/price ratio and expect the future of the price which is a very complex analysis, but just look back at the already "scribed in price" type of action and guess the future from it.

Its basically the same thing if you look in it, however maybe the price as a time series lost many information about itself (LV2 QUOTES, LV3 QUOTES).

But still if there is any underlying pattern which repeats itself, then time series analysis works.The market is too inefficient and there is always an exploitable phenomena which can be traded.Wall street quants succeded at this, if HFT can trade profitably at a so tiny unpredictable timescale, then it is possible to trade at higher timescales aswell, and even more reliable (given that retailers dont have the resources as a HFT firm).
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
{quote} Price itself is driven by 2 main things, that is the liquidity provider book balancing when the net orders are in a big unbalance due to excessive acumulation of one orders, so the price balances itself to a more stable place, in layman terms called mean reversion,crash,retracement,etc. Secondly it is moved by events, such as news, which change the fundamental enviroment of the market => and the human psychological reaction of the traders to this. So basically its 3 basic concepts, however from a more aggregate position its simply the imbalance...
May I ask in which camp do you belong?

Do you subscribe to the idea that the Market is driven by Trader Feedback?
Or
Do you believe the Market is driven by “Mathematical manipulations”?

Does your concept of Mean reversion dictate where price will go next?
Or
Does trader consensus dictate where price will go next?
Markets are not efficient, rather they are effective - Jones
Quoting mzvega
{quote} May I ask in which camp do you belong? Do you subscribe to the idea that the Market is driven by Trader Feedback? Or Do you believe the Market is driven by “Mathematical manipulations”?
If you think about that question for a while you find out that its meaningless.

It cannot be "driven" (in the meaning of the word) by those 2 things.The mathematical manipulation is a result not a cause, you can only analyze it mathematically after it happened, so all quant models of the markets are retroactive.Whereas trader feedback is not neccesarly imply action, if a trader is thinking about bullish positions that doesnt move the market until he puts the trade.

But if you meant by the word driven the sense that which one determines or influences the future price then i would say the trader feedback, as the mathematical manipulation is only available after you have the data to begin with.

Quote
Does your concept of Mean reversion dictate where price will go next?
Or
Does trader consensus dictate where price will go next?

This is actually funny because nobody knows the real mean of the market, so because many guys use different moving averages it will probably not work in this way.
But not even the trader consensus dictates it, after its overtraded.So the answer is in some cases the trader consensus.

Mean reversions are caused my market makers.

EDIT: Its basically like an elastic band, until the band is not stretched too thin the MM let the traders play in their market by collecting spreads & comissions, but after its stretched too thin, they intervene to reset the band to its original form.
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
EDIT: Its basically like an elastic band, until the band is not stretched too thin the MM let the traders play in their market by collecting spreads & comissions, but after its stretched too thin, they intervene to reset the band to its original form.
If the market would not be manipulated at all, so that all traders would trade between themselves not trough a 3rd party or with a liquidity provider AND assuming that residual effects like inflation and other stuff would not corrode the currencies, then the FX market would look pretty much like a sine wave, all traders would be long if the price would rise, while those who would sell in this rally would take profit at the next swing.

The only problem would be when the price is exactly at the mean, because at this point it would lie flat because nobody with brains would trade at this point.But i`m sure there would be some suckers who would buy/sell into this.

Also the news events themselves could also be distorting on the price, its not the MM who are ultimately the responsible for this kind of manipulation but the news agencies, and the news events themselves,because the MM's only protect their money, if they would not do this they would go broke instantly, it is that "mania" which the outside world is putting on the trader that is responsible for it
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
{quote} Yes indeed, that is the problem, because there is 1 formula for the MEAN of the market which is measuring it perfectly, the problem is that it is unknown.Heck i believe there is a formula that can measure anything with some degree of certainty, there is none that can eliminate uncertianty completely, but there should be one which is the most optimal. So with the markets, there is a formula that can predict X+1 bars ahead always, the problem is that nobody knows it, if you knew it then you would be a billionaire and make sure that nobody...
Well proximus, He knew it and found how to use it.

Shared theory here..

Cool isnt it ?

Just my 2 cents. Take he's words as gold.. You will thank me later.

PS. Oh about billions,

Wake up, will ya, pal? If you're not inside, you're outside, okay? And I'm not talking a $400,000 a year working Wall Street stiff flying first class and being comfortable, I'm talking about liquid. Rich enough to have your own jet. Rich enough not to waste time. Fifty, a hundred million dollars, buddy. A player, or nothing
Neural Network Algorithm and Econophysics
Quoting Proximus
The mean is very hard to measure in a heteroscedastic market like FX.
I have a feeling that you didn't really understood heteroschedasticity. It has nothing to do with the mean, but with the variance. And the market is heteroschedastic only in wall-clock time. In trade time it is not, so you can get rid of heteroschedasticity if you want, but for that you have to give-up candles.

Quoting Proximus
If one could precisely measure the SD correctly then it would be the holy grail, because you just need to put the SL outside lets say 3 deviations and profit, but unfortunately the market is a very tricky thing.
You are approaching this the wrong way. The SD is already measured by the whole market. And in fact you can even bet on it/buy it/sell it. If you can measure it in a better way, you can literally profit from this (without caring about the direction). The SD that I'm talking about it's called implied volatility and it's contained in option prices.

So if you want the best possbile SD computation that you can get, you can look at the current implied volatility for your term. If you don't want to pay for this, you can see it here updated twice a day: https://fxowebtools.saxobank.com/otc.html

Using these, you can build a volatility term structure surface. For example in this GBP/USD IV term structure surface you can see the recent SD spike due to the Scotland referendum, and it's subsequent collapse. A simple MA would never caught this (since it implies politics and real world news). Note that the spike is BEFORE THE FACT, not measured after as many would assume:

https://dl.dropboxusercontent.com/u/...m_scotland.png
Quoting Vigovski
{quote} Well proximus, He knew it and found how to use it. Shared theory here.. Cool isnt it ? Just my 2 cents. Take he's words as gold.. You will thank me later. PS. Oh about billions, Wake up, will ya, pal? If you're not inside, you're outside, okay? And I'm not talking a $400,000 a year working Wall Street stiff flying first class and being comfortable, I'm talking about liquid. Rich enough to have your own jet. Rich enough not to waste time. Fifty, a hundred million dollars, buddy. A player, or nothing
Who is "He" ?

Quoting Adal
{quote} I have a feeling that you didn't really understood heteroschedasticity. It has nothing to do with the mean, but with the variance. And the market is heteroschedastic only in wall-clock time. In trade time it is not, so you can get rid of heteroschedasticity if you want, but for that you have to give-up candles. {quote}
Yes i know the definition, but in a heteroskedastic market the mean is ever changing, because the distribution is so.That is the relation between the variance and the mean.The mean is the mid point of the distribution, how can i define the mean if the distribution is ever changing around it? This is my current research.

My current project is to find a formula which defines the mean as best as possible and to try to normalize the market, or find a mean which is the mean of all the distribution which the price is/was/ever be.

I`m experimenting with AFIRMA models at the moment.

Quote
You are approaching this the wrong way. The SD is already measured by the whole market. And in fact you can even bet on it/buy it/sell it. If you can measure it in a better way, you can literally profit from this (without caring about the direction). The SD that I'm talking about it's called implied volatility and it's contained in option prices.

Correction, the past SD is the market itself, the fact that the EUR/USD hit 1.6009 on the upside would imply that thats the maximum that it will go, perhaps if the fundamentals dont change, but what if the FED starts to print money again like crazy and dilute the USD further, then who is to tell me how high can the EUR appreciate vs the greenbuck.

I`m talking about forecasting SD on a week by week basis with a very accurate method.

You just gave me a good idea about implied volatility, thanks
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
I`m talking about forecasting SD on a week by week basis with a very accurate method. You just gave me a good idea about implied volatility, thanks
I'm not sure you understood, so I'll say it again: implied volatility IS THE SD FORCAST on a week by week basis to the best of the whole market knowledge (ie: everything that might affect the SD is priced in). So the SD needs not to be forecasted, it already is. Of course, if your own forecast does not agree with the market forecast (the IV price), then you can sell or buy volatility directly (with delta-hedging)

That's why I said that IV IS ABOUT THE FUTURE, not about the past. A lot of people have this misconception, that IV somehow is historical when it is not, it is a prediction about future volatility, not a measure of past volatility.
Quoting Adal
{quote} I'm not sure you understood, so I'll say it again: implied volatility IS THE SD FORCAST on a week by week basis to the best of the whole market knowledge (ie: everything that might affect the SD is priced in). So the SD needs not to be forecasted, it already is. Of course, if your own forecast does not agree with the market forecast (the IV price), then you can sell or buy volatility directly (with delta-hedging) That's why I said that IV IS ABOUT THE FUTURE, not about the past. A lot of people have this misconception, that IV somehow is...
Yes thanks for the explanation i understand now.But i dont really want to trade option markets for now.

Besides my primary goal is to find a more suitable measurement of the SUPPLY & DEMAND relationship, volatility only tells me the size of my stoploss.
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
{quote} Yes thanks for the explanation i understand now.But i dont really want to trade option markets for now. Besides my primary goal is to find a more suitable measurement of the SUPPLY & DEMAND relationship, volatility only tells me the size of my stoploss.
You don't need to trade options. The idea that if you need a volatility/SD estimation, the best you can find are the IV prices. Much better than ATR/GARCH/or custom MA.

That's what I'm using when I need to estimate where to put a volatility based stop. GARCH will never increase in advance of a referendum, neither will a custom MA, however good the mean is, because the mean also doesn't know about the future referendum (or NFP, or single name event)
This is not a simple question to address. You must first discover/define if you are an immediate term trader, intraday trader, medium term, or longer term. This usually has to do with the amount of time you have available to trade currencies or your preference for trading.

I would not trade lower than M5 and even then for only for entries with extremely tight stops. Price action strength probability tends to degrade as time frames decrease. But for very short time frames, a tick chart is probably more effective. Tick chart diagram screenshot below.
Attached Image (click to enlarge)
Click to Enlarge Name: 2014-08-07_EOVM2.png Size: 341 KB


In my opinion, multiple time frames are more effective because they provide more evidence to support entries

First time frame/chart(lowest time frame) should be the chart you use to take your entries, second would be your primary price action chart, third time frame is your Map of longer term Supt/resistance levels supported candle stick patterns.

Here are chart progressions that I would recommend
Entry Chart/ Price Action Chart/ Map of Supt Res & candle sticks
Immediate term: M5/M30/H4
Intraday: M15/H1-H4/Daily
Medium Term: H4/Daily/Weekly
Long Term: Daily/Weekly/ Monthly

Below are a few chart set-ups that I have used in the past. I trade intraday using median renko charts for my entry and midrange price action charts.
I use M15/H1-H4/Daily these are my timeframes from left to right

Attached Image (click to enlarge)
Click to Enlarge Name: 2014-08-21_CT.png Size: 63 KB


EU/GU tandem setup

Attached Image (click to enlarge)
Click to Enlarge Name: SErver_charts.png Size: 152 KB



Below is my current Mt4 setup detailing a trade I took

Attached Image (click to enlarge)
Click to Enlarge Name: 2014-08-29_CT.png Size: 80 KB


Hope this helps. Thanks for starting the thread
Quoting Adal
{quote} You don't need to trade options. The idea that if you need a volatility/SD estimation, the best you can find are the IV prices. Much better than ATR/GARCH/or custom MA. That's what I'm using when I need to estimate where to put a volatility based stop. GARCH will never increase in advance of a referendum, neither will a custom MA, however good the mean is, because the mean also doesn't know about the future referendum (or NFP, or single name event)
Maybe they dont but the market has also some boundaries between which it can go or not, that is why historical analysis is good, to find patterns, obviously.

The market is tightly limited between the current market capitalization and the willingness of external participants to come and trade in that market, if the traders dont want to trade or they are not exposed then obviously the market wont move much.

Now you cant know ahead anything, and certainly a referendum like this is rare so you cant know the outcome for sure, but no matter what the outcome is, the market still has boundaries.

What if Scotland would have seceded? How much would the pound lost ? 50% or so?Some stupid analysts said that it could lose up to 70% of its value ?
Obviously this is ridiculous, the pound would have not lost more of its value than the GDP[PPP] of Scotland, so if Scotland makes up X% of the UK GDP [PPP], probably thats as much as the pound would have lost, in the worst case scenario of course, if Scotland it were to adapt other currencies of course.Because the news and other stuff would have skewed those facts, many guys would probably bought it before it it reached the bottom.If Scotland would have still used the pound then obviously it would been just another whipsaw abusing of retail suckers,as many would have sold it because of non-established fear without any reason and many pro's would have stayed on the other side waiting to cash in.

So it can be predicted with other tools aswell, but my point is that the market swings as such in the historical data also contain much much information about the market capitalization and the likelihood of reversals at a certain point.Not certain, but i believe it can be used to forecast the price with an edge.
"There's a sucker born every minute" - P.T. Barnum
Quoting Proximus
{quote} If a trader knows what he is doing then he will probably, sooner or later, find the best period suitable for him. If the strategy he is using, and he already has settled the time horizon of his trades, are matched with the TF of his choice then its good. But perhaps i wanted to know why X trader chooses Y timeframe, particularly.So as if why scalping > swing trading or vice versa, if the majority wants to scalp then why M15 instead on M5 for example. Because there is a clear tendency towards, larger timeframes, but not that large, so looks...
I agree with your views about one choosing a time frame that suits his/her trading style/method.
It was good to learn from the poll that most people here use H4, maybe because it is half of a session's length and they can see what each session does in its first and second half.
The Thief of Wall Street
Quoting Proximus
{quote} Obviously this is ridiculous, the pound would have not lost more of its value than the GDP[PPP] of Scotland, so if Scotland makes up X% of the UK GDP [PPP], probably thats as much as the pound would have lost, in the worst case scenario of course, if Scotland it were to adapt other currencies of course.
That line of reasoning is wrong. The sum is greater then the parts.

See the great Larry Summers make fun of your line of reasoning taken to extreme. 30 sec from 9:00

Inserted Video
Quoting Adal
{quote} That line of reasoning is wrong. The sum is greater then the parts. See the great Larry Summers make fun of your line of reasoning taken to extreme. 30 sec from 9:00 http://www.youtube.com/watch?v=KYpVzBbQIX0
The only thing he said which is relevant to my post is that if the "electricity would go off then it would not affect the economy by only 4% even if electricity only makes up 4% of the GDP"

How is that analogy indentical to what i`ve said ? Scotland is yes an important financial/economic center of the UK, but definitely not of a ultimate importance.

Which one would have a higher negative impact on the pound: Scotland seceding or no electricity at all for 2 years in the UK?

You see my analogy there

The fact is that its not core importance, take a look at Ucraine indexes: http://www.bloomberg.com/quote/UX:IND
Not really affected by the loss of Crimea.You could also look at Serbian indexes after the loss of Kosovo, or North Sudan indexes after the loss of South Sudan.
Its not the end of the world just because a huge part of the economy went missing.
"There's a sucker born every minute" - P.T. Barnum
Well guys according to the current state of the poll there is a very strong bias towards H4 timeframe (which was expected):

Attached Image (click to enlarge)
Click to Enlarge Name: SASDASDASD.png Size: 22 KB


But i find it strange that nobody used renko charts still.Well, anyway H1 and D1 as neighboring TF's are also popular, interesting.
"There's a sucker born every minute" - P.T. Barnum
TF's should be chosen that match your personality, general mood, appetite for stress, attitude toward volatility and belief of price action 'consistency' in the chosen TF.

TF Breakdown:

1m is gambling, random chance, fun loving, cocaine fueled..."That's the one. Get it. GET IT! Now! NOW!!!!1111oneone. Oh damn..damn..oh wait...wooohoo! No, noNooooo!"

5m is just silly. Really. Silly. It seems like it's working and then wanders off and goes and stars munching on a cheese sandwich...and then turns around and hammers a nail in your big toe. Whoa. Nope, can't cope.

15m is "fun", but still a bit silly at times.

30m seems more reliable and allows me to 'get in early', to beat the 1h crowd, but still can give false positives that 1h and 4h would just shake their heads at.

1h seems quick enough so I don't get too bored, holds a fair bit of 'price action', as it contains a chunk of 'serious' trade direction along with the noise from 1-30m.

4h is a great TF, but far too slow, thereby exists mainly as a complimentry (wise uncle) TF to nod sagely at my trade activity (based on 1h price action) or occasionally 'tut' at my naive and pre-emptive attempts to 'get in too soon'.

1d is a classic and must NOT be ignore for solid trend direction/reversal notification.

1w is interesting, as it gives a "looking at the stars for the first time - oh, wow", BIG view of whats really been going on, amongst all that quivering activity from 1h to 1d.

1m is a sloth, glacier, mountain (choose and old, slow, wise analogy) that only provides the true answers for those that wait...but ne're can ye trade from this land, but here is where the 'sages' will talk in detail of what should have been done, if only the ignorant savages could see it...
0 traders viewing now