Nonparametric Association Measures in Systematic Trading | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Systematic trading 1,576 replies

Trade another day through Systematic Trading 203 replies

Anyone use Systematic from Alpari? 1 reply

A Systematic Approach to Markets 68 replies

Systematic Trading vs. Discretionary Trading 14 replies

Attachments: Nonparametric Association Measures in Systematic Trading
Exit Attachments

Nonparametric Association Measures in Systematic Trading

  • #21
  • Jan 18, 2018 9:03am Jan 18, 2018 9:03am
Quoting alphadude
The challenge of course is to find those weights
The challenge is not really here. 7bit solved this part here https://www.forexfactory.com/showthread.php?t=262827
IIRC there is some MT4 stuff to download.

The currency pairs aren't co-integrated. 7bit's pragmatic approach was to rebalance often while keeping the trading fees acceptable.
No greed. No fear. Just maths.
  • #22
  • Edited 11:13am Jan 18, 2018 10:51am | Edited 11:13am
Quoting FXEZ
{quote} Hey Ultimate, I came to similar conclusions a while back but also feel (as you appear to) that further investigation of the concept could yield something worthwhile. If you or any other experienced developers with R experience want to form a work group to compare scripts and test ideas, please send me an email through this forum. Cointegration is built on the concept of regression of two or more time series in order to find a long-term linear combination that can yield stationary residuals that may be suitable for pairs trading. At the heart...
You can summarize the result of 7bit thread by the above quote

I personally already dwelt into this subject years ago; and gave up. I find the moving average to be the best detrending method to get a perfectly stationary timeseries, stable over time, and guaranteed to revert !

now who said that moving average based systems dont work ????

attached is a chart example of a SMA(5m, 20) of EUR/USD. Look at the spread; and see how easy it is to trade the spikes
Attached Image (click to enlarge)
Click to Enlarge Name: sc_20180118-165352.png Size: 255 KB
  • #23
  • Jan 18, 2018 11:30am Jan 18, 2018 11:30am
Quoting alphadude
{quote} You can summarize the result of 7bit thread by the above quote I personally already dwelt into this subject years ago; and gave up. I find the moving average to be the best detrending method to get a perfectly stationary timeseries, stable over time, and guaranteed to revert ! now who said that moving average based systems dont work ???? attached is a chart example of a SMA(5m, 20) of EUR/USD. Look at the spread; and see how easy it is to trade the spikes {image}
Can you explain how you build the spread? I mean if you buy the asset how do you "sell" the moving average to hedge the trend?
No greed. No fear. Just maths.
  • #24
  • Jan 18, 2018 11:45am Jan 18, 2018 11:45am
spread = price - moving average

for long position; the rules are (mirror for short):
buy EURUSD when spread < -10 pips
sell EURUSD when spread >= 0

obviously when the market is trending; you may loose money depending how early you entered the trade

this system is also called "moving average crossover".

although the concept is very simple; but it is hard to implement it as a profitable system. the nitty gritty details is what really matters
  • #25
  • Jan 18, 2018 1:23pm Jan 18, 2018 1:23pm
Alphadude,

Can you shoot me the subwindow indicator ,z transform of the subwindow indicator and RSI of the subwindow indicator in csv file?
I try to create buy-sell formula with the software I am using.I attached a sample csv .
Attached File(s)
File Type: csv sample.csv   30 KB | 455 downloads
  • #26
  • Jan 18, 2018 1:26pm Jan 18, 2018 1:26pm
Aha => spread = price - moving average

Wait a minute,l think I can do this.
  • #27
  • Jan 19, 2018 4:24am Jan 19, 2018 4:24am
Quoting alphadude
spread = price - moving average for long position; the rules are (mirror for short): buy EURUSD when spread < -10 pips sell EURUSD when spread >= 0 obviously when the market is trending; you may loose money depending how early you entered the trade this system is also called "moving average crossover". although the concept is very simple; but it is hard to implement it as a profitable system. the nitty gritty details is what really matters
You clearly don't understand what spread trading is! The goal is be become MARKET NEUTRAL under the assumption of mean reversion. When the spread returns to the zero line you're supposed to be in PROFIT.

Quoting alphadude
this system is also called "moving average crossover".
It is actually the opposite of MA-crossover since you go against the direction of the cross .

Quoting patara
Aha => spread = price - moving average Wait a minute,l think I can do this.
No you can't unless you find a way to somehow "sell" the moving average.
No greed. No fear. Just maths.
  • #28
  • Jan 19, 2018 4:45am Jan 19, 2018 4:45am
hi PipMeUp, nice observations, and welcome to the real world of attention to details in trading. that's what all matters.

you are trading the spread timeseries; which is stationary; i.e. it is guaranteed to go back to zero on daily basis (depends on the period of the moving average).

since you cannot trade the moving average; you only take one side of the pair; which is the price series itself.

this strategy is explained in the book "Trade like a hedge fund" by James Altrucher, in chapter 2. It is as simple as I indicated above.
  • #29
  • Jan 19, 2018 6:25am Jan 19, 2018 6:25am
If you do this and keep considering it as a spread it is like you force the weight of the second asset (the MA) to be 0. The asset price becomes a "rough approximation" of the spread which can become very far "off target". Are you going to correct it by continuously "re-balancing", that is adding more and more to a losing position?

Look on the chart you posted above I can see at least 8 big losers. I doubt the tiny winners will balance.
No greed. No fear. Just maths.
  • #30
  • Jan 19, 2018 2:55pm Jan 19, 2018 2:55pm
Quoting PipMeUp
If you do this and keep considering it as a spread it is like you force the weight of the second asset (the MA) to be 0. The asset price becomes a "rough approximation" of the spread which can become very far "off target". Are you going to correct it by continuously "re-balancing", that is adding more and more to a losing position? Look on the chart you posted above I can see at least 8 big losers. I doubt the tiny winners will balance.
hi PipMeUp, in pairs trading concepts, when you go long the spread; it means you buy one of the legs, and sell the other; and the opposite for going short.

trading unilateral pairs; means you only take one side of the trade; which is trading one leg only. In our case; we trade the EUR/USD price series, and do not trade the moving average; as it is not tradeable.

to make this system profitable; you have to decide when you trade the spread; i.e. when the spread has moved too far away from the zero line. You can base this decision on historical spread of the last x bars. for example; if the spread is averaging about 5 pips away from the zero line; then a wise decision is to trade it when it is more than 10 pips from the zero line.

you can be creative on when to enter; in order to turn it onto a profitable system.

I advice that you do not add to losers. meaning; you only have one position open.

also; do not use stop losses; it will not work properly with pairs trading. Simply enter when the spread is wide enough; and close the trade when it crosses the zero line.

to spice up the conversations; here is a live trade I took today; based on this concept. Yep; I am trading this live; and it is profitable so far.
Attached Image (click to enlarge)
Click to Enlarge Name: sc_20180119-205750.png Size: 120 KB
  • #31
  • Jan 19, 2018 2:59pm Jan 19, 2018 2:59pm
Here is another one I took this morning on EUR/USD
Attached Image (click to enlarge)
Click to Enlarge Name: sc_20180119-205905.png Size: 117 KB
  • #32
  • Jan 20, 2018 2:37am Jan 20, 2018 2:37am
Quoting alphadude
Here is another one I took this morning on EUR/USD {image}
Alphadude, so are you short or long the EUR/USD? The US government shutdown may throw statistical relationships in the air for a while. My experience agrees with yours: stop losses do not quite work with FX pairs trading. However, I assume you are using some maximum duration of a trade, are you not?

In general, EUR/USD may not be the best cross for expressing your ideas. It's a regime-switching beast, which would go from dynamics A to dynamics B all of a sudden and return to dynamics A only after 1-30 weeks... Over the last year, I have done some machine learning analysis on how "well-behaved" various crosses are. I define a cross as "well-behaved" if its dynamics can be understood relatively well with various non-trivial models. In this competition EUR/USD falls in the bottom 40% of 25+ crosses. For example, AUD/CAD or AUD/NZD are much better behaved. One just needs to choose the right broker, with relatively tight spreads... Of course, my analysis is not final.
  • #33
  • Jan 20, 2018 4:34am Jan 20, 2018 4:34am
Quoting staassis
{quote} Alphadude, so are you short or long the EUR/USD? The US government shutdown may throw statistical relationships in the air for a while. My experience agrees with yours: stop losses do not quite work with FX pairs trading. However, I assume you are using some maximum duration of a trade, are you not? In general, EUR/USD may not be the best cross for expressing your ideas. It's a regime-switching beast, which would go from dynamics A to dynamics B all of a sudden and return to dynamics A only after 1-30 weeks... Over the last year, I have...
i trade on 1minute or 5 minute bars only. positions are open for few minutes to hours and closed at end of day. the positions mentioned in the charts previously were closed within 20 minutes

your analysis on behavior are interesting. it will all depend on the timeframe chosen.

mean reversion work best on shorter timeframes
  • #34
  • Jan 20, 2018 1:37pm Jan 20, 2018 1:37pm
Quoting alphadude
{quote} your analysis on behavior are interesting. it will all depend on the timeframe chosen. mean reversion work best on shorter timeframes
My analysis is based on 1-minute data taken from Dukascopy (https://www.dukascopy.com/swiss/engl...tch/historical). I am not interested in higher frequencies because I do not want to trade purely algorithmically at this point. I have experimented with various specifications of the decision window (N1 minutes) and maximum trading window (N2 minutes). Transaction costs have been accounted for but not necessarily in the Dukascopy fashion, since they offer suboptimal spreads on selected crosses. All major (and most liquid) crosses have been tested. For each cross, the "good behavior" has been quantified by several objective numerical measures... The analysis is not quite a justification of my wishful thinking. For example, I had big expectations of GBP/AUD based on my discretionary experience. But it did not perform that well.
  • #35
  • Jan 21, 2018 12:43pm Jan 21, 2018 12:43pm
I use dukascopy data too for my analysis. it is great service.

what do you think about AUD/NZD? in theory they are cointegrated but i did not find success in trading it. perhaps due to high transaction cost
  • #36
  • Edited 1:32pm Jan 22, 2018 11:37am | Edited 1:32pm
Quoting alphadude
what do you think about AUD/NZD? in theory they are cointegrated but i did not find success in trading it. perhaps due to high transaction cost
Alphadude, I never tested / tried pairs trading involving AUD/NZD. Inefficiencies in AUD/NZD were identified using a different kind of screening. Having that said, it does not mean that some form of flexible cointegration is not possible... I thought that AUD/CAD and AUD/NZD might be of particular interest to you since you are based in Australia and since there is much "juice" in there relative to EUR/USD. Being awake during the liquid hours is important as well. On the other hand, trading EUR/USD might be tough since you have to miss the New York session... Btw, in your opinion, what are the best brokers to trade AUD and NZD crosses?
  • #37
  • Jan 22, 2018 5:34pm Jan 22, 2018 5:34pm
Quoting staassis
{quote} Alphadude, I never tested / tried pairs trading involving AUD/NZD. Inefficiencies in AUD/NZD were identified using a different kind of screening. Having that said, it does not mean that some form of flexible cointegration is not possible... I thought that AUD/CAD and AUD/NZD might be of particular interest to you since you are based in Australia and since there is much "juice" in there relative to EUR/USD. Being awake during the liquid hours is important as well. On the other hand, trading EUR/USD might be tough since you have to miss the...
most ECN brokers are good for the crosses. try Pepperstobe, ICMarkets, Axitrader or GlobalPrime
  • #38
  • Jan 22, 2018 11:36pm Jan 22, 2018 11:36pm
Quoting alphadude
{quote} most ECN brokers are good for the crosses. try Pepperstobe, ICMarkets, Axitrader or GlobalPrime
Thanks!
  • #39
  • Jan 25, 2018 11:50am Jan 25, 2018 11:50am
Another live trade today on EURUSD using SMA(60)

trade lasted 15 minutes; and made 10 pips profit
Attached Image (click to enlarge)
Click to Enlarge Name: sc_20180125-174915.png Size: 114 KB
  • #40
  • Jan 26, 2018 1:18am Jan 26, 2018 1:18am
Quoting alphadude
Another live trade today on EURUSD using SMA(60) trade lasted 15 minutes; and made 10 pips profit {image}
Ok, great, but how did the strategy perform over the whole of [Jan 25 European session + Jan 25 US session], with EUR/USD going from 1.24 to 1.253 and back to 1.237. I would imagine the strategy to perform better in calmer waters.
0 traders viewing now