Robot Lab Project | Page 32 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Financial Markets Observatory Lab 1,095 replies

Indicators Lab 236 replies

Rel's Lab 103 replies

Nikola's Trading Lab 247 replies

Spud's Lab Accident 22 replies

Attachments: Robot Lab Project
Exit Attachments

Robot Lab Project

  • #621
  • Edited 4:04am Feb 21, 2015 3:35am | Edited 4:04am
Sorry felixis for the late reply, as I just came back from holiday.

Quoting felixis
as for basket trading several issues are to be considered

1) history made by the script generates high and lows that may not be the real ones. you base your computing on High and Low of each pair the high and low may not occur at the same time on each pair hence the high and low is virtual good luck for backtesting and drawing RS on this
This effect is minimized since we are using M1 to generate the basket history. However after completing the history, the basket is updated every 2 seconds which is very close to live Tick Date.

Quoting felixis
2) the position size of buy and sell order on each pair are the most important, you are limited to 0.01 lot increments
by consequence the weighting of the pairs cannot be arbitrary they must be computed such as when a basket order occur
the lot size are exactly a multiple of 0.01 lot and greater than or equal to 0.01 lot
Lot size can be normalized by using the Tick Value for each pair.

Quoting felixis
3) if you trade the chart like a standard currency you expect when it goes up or down that your equity will follow the same
hence the case 2)
if not you are seeing an indicator that has nothing to do with the real trade such as the rsi, or the sma for example
Yes. Equity should go up or down following the basket chart, but not in exact pips.

Quoting felixis
4) by corrolary of case 2) trading a basket put you at a disadvantage

- the spread of the virtual basket is higher

- the execution on this prehistoric platform is a problem as MT4 cannot process more than one order at a time

- you cannot trade 1 microlot
-> if you have say 10 currencies in the basket the minimum size tradable is 0.1 lot = 10 * 0.01
-> if you use coefs differents from 1 the minimum size will be higher

- the leverage used will be reduced

- the margin required to trade will be higher

- you are exposed to spread swipe in any currency that could endanger the all trade
All these are normal conditions for trading multiple currencies and hence they apply to the basket. No harm in that.

Quoting felixis
taking the inverse of the pair is dangerous
you are trading in lot and if you invert the pair the variation for one pip is not the same
EDIT: the variation of money for the pair is linear 1 pip = x amount of $ for a fixed amount of lot
if you take the inverse of it
the variation is not anymore linear in terms of money
Inverse of the -ve directional pairs seems to be the only available solution so far. Pips (or Tick values) should be incorporated in lot size calculation.

Quoting felixis
the advantage to trade something virtual is dubious as nobody else than you see the instrument
the validity of SR are compromised
I don't see this as a Disadvantage, as long as your Basket does make sense, and not just a mix of random instruments.

Quoting felixis
be aware that SR and all other patterns appear in a random walk too
I don't think so, again , as long as your Basket does make sense, and not just a mix of random instruments.

Quoting felixis
demo platform is always fast in the execution but on real condition the chaining of the orders and the non transactionnality of it can make your order takes 20 second to complete with the slippage damage you can imagine
This is normal conditions when trading multiple currencies and hence they apply to the basket. No harm in that.

Quoting felixis
and the final 5) if you think you can mitigate the chaining by putting limit/stop order you cannot put limit stop order on the chart as one value in the chart can equate to an arbitrary combination of values in the pairs hence the impossibility to put on each pair a stop limit order you can only trade with market orders
I know. Simulation of Stop or Limit Orders can be applied to the Basket Chart as LITEchild suggested HERE, and when triggered (Using the EA) will open market orders on the Basket pairs.

Thank you felixis for raising such concerns, they've been helpful to me.

Regards,

Khalil
No guts, no glory
  • #622
  • Feb 21, 2015 4:07am Feb 21, 2015 4:07am
OK. So were do we go from here:

  1. Test different basket calculation approaches
  2. Apply the best approach to the Create_Basket_Script
  3. Finalize the first version of Basket Trading Robot

No guts, no glory
  • #623
  • Feb 21, 2015 4:55am Feb 21, 2015 4:55am
Abokwaik Thank you for your explanation. I myself am a newbie to robot systems and robot NBB_XBB I was very pleasantly surprised by their work. I use 1H TF and I will try different settings and TF. I'll let you know how I was satisfied with the system. Excellent job from you abokwaik
  • #624
  • Feb 21, 2015 8:27am Feb 21, 2015 8:27am
Quote
This effect is minimized since we are using M1 to generate the basket history. However after completing the history, the basket is updated every 2 seconds which is very close to live Tick Date.
1) in your create_basket_script

while(!IsStopped()) // computation of current real time bar
{
...
low+= NormalizeDouble(iLow (pairs[j],0,0)*pair_dw[j],i_digits);
high+= NormalizeDouble(iHigh (pairs[j],0,0)*pair_dw[j],i_digits);

Sleep(50);
}

you are taking the iLow, iHigh of the pair every 50milliseconds
summing the iHigh of the pairs still gives you a virtual high for the basket as the iHigh is the high of the 1minute bar of each pair
you can sample every 1nanosecond it will change nothing

if you want to have the real bar you must sample from the Bid value of MQL4 and create the bar with
high = max(high, Bid)
low = min(low, Bid)
but because you are sampling every 50milliseconds you could miss a Bid tick and thus the bar would be wrong
moreover because your history has been made by iLow and iHigh, the history high of the basket is still virtual
thus you would have a mismatch between the history and the current bars because the method of getting the high would be different




2) your script is lagging
prehistoric platform has been hacked by
PostMessageA(hwnd,WM_COMMAND,33324,0);
I am impressed by the hacking made by the own developer of this prehistoric platform
the threading model of prehistoric platform is subpar

you'd better make an indicator of it and send the refresh window command
at every OnCalculate event by putting it in an active pair
  • #625
  • Edited 8:56am Feb 21, 2015 8:29am | Edited 8:56am
Quote
Lot size can be normalized by using the Tick Value for each pair.


if you have a basket of 10 pair
you must enter a position in each of these 10 pairs
thus you must enter with 0.01 * 10 = 0.1lot at a minimum

for 10 pairs in basket minimum lot size = 0.1 lot and minimum increment = 0.1 lot
for 20 pairs in basket minimum lot size = 0.2 lot and minimum increment = 0.2 lot

weights cannot be arbitrary

they must be integers 1, 2, 3 if not you will have lot size you cannot trade
of course you can scale your basket by arbitrary real constant but that will change nothing in the shape of the chart

if you take weights differents from 1 the minimum lot size is worse
example:
3 pairs
weights: 1, 2, 3
minimum lot size tradable:
(1 + 2 + 3) * 0.01 = 0.06 lot
  • #626
  • Edited 9:12am Feb 21, 2015 8:33am | Edited 9:12am
Quote
All these are normal conditions for trading multiple currencies and hence they apply to the basket. No harm in that.


the trading condition of the basket is more harmfull than a trading for single pair:

1) order can take forever to complete
2) average spread is higher due to trading exotic pairs with crazy spread

most importantly

you will never have a trending higher than the highest trending pair
because you add pairs that do not trend or trend weaker or trend the opposite

because of this

the amplitude of the variation of your basket is reduced
but the spread you pay is higher

you must have a hell hedge to fight this adverse condition
better apply it to single pair
  • #627
  • Feb 21, 2015 8:35am Feb 21, 2015 8:35am
Quote
I don't see this as a Disadvantage, as long as your Basket does make sense, and not just a mix of random instruments.
you mix a random instruments

or

you mix at random real instruments

mixing at random real instruments (arbitrary weights = mixing at random)
will lead to a basket that converge to a random walk
summation of stochastic process always lead to a random gaussian walk
  • #628
  • Edited 9:14am Feb 21, 2015 8:50am | Edited 9:14am
Quote
Yes. Equity should go up or down following the basket chart, but not in exact pips.


Quote
Inverse of the -ve directional pairs seems to be the only available solution so far. Pips (or Tick values) should be incorporated in lot size calculation.


i give you a counter example that falsify this assertion

say we have a basket made of 2 pairs
EURUSD weight = 1
GBPUSD weight = -1

basket value = EURUSD + 1 / GBPUSD

at time t0 you buy the basket at the following values
EURUSD = 1 *********** buy 0.01 lot ********** pip value = 0.1$
GBPUSD = 1 ***********sell 0.01 lot ********* pip value = 0.1$

basket value = 1 + 1 / 1 = 1 + 1 = 2

at time t1 you have the following values
EURUSD = 1.75
GBPUSD = 4

basket value = 1.75 + 1 / 4 = 1.75 + 0.25 = 2

normally according to you you have 0 gain/loss

EURUSD gain(buy): 1.75 - 1 = 0.75 = 7500 pip = 750$
GBPUSD gain(sell): -(4 - 1) = -(3) = -3 = -30000pip = -3000$

total gain for trade = 750 - 3000 = -2250$

for the same basket value of 2 you end up at t1 to have a loss -2250$

you are not trading money
you are trading an indicator
likewise if rsi is at 80 that does not mean your trade is at zero gain
  • #629
  • Edited 10:01am Feb 21, 2015 9:39am | Edited 10:01am
Quoting abokwaik
{quote} Thanks for your input, But again, the discussion is on a Basket , a Generic Basket of Instruments, not an Index. Although some implementations of the basket might look like an index. Most brokers offer indexes already, which can be traded directly from Metatrader. And if we want to trade pairs related to the index (instead of the index itself) a simple EA can be developed and applied on the index chart. We don't need a an offline chart for that. A basket can be composed of a mix of instruments with relevant directional weights: for example...
Referring to the quoted post, and by using the Create_Basket_Script , a Commodity Currencies Basket with corresponding directional weights between parentheses:
USD/CAD (-1), AUD/USD (+1), NZD/USD (+1), EUR/CAD(-1), EUR/AUD(-1), EUR/NZD(-1), GBP/CAD(-1), GBP/AUD(-1), GBP/NZD(-1)

The latest H1 offline chart looks like this:

Attached Image (click to enlarge)
Click to Enlarge Name: CMDTH1.png Size: 52 KB


Can this chart be used to give a Signal to trade the Basket Component pairs ?

Check how the chart re-acted to fibo levels 23.6, 38.2 and 61.8, Amazing

PS: For people unfamiliar with commodity currencies, they are (AUD,NZD,CAD).
No guts, no glory
  • #630
  • Feb 21, 2015 9:58am Feb 21, 2015 9:58am
Now suppose our system trades the 61.8 fibo bounce at close (orange horizontal line), targeting 0 fibo (or 61.8 expansion) lime horizontal line.
So we made our entry on the 9 pairs composing the basket at the close of H1 (2015.02.19 09:00), and Closed positions at the close of H1 (2015.02.20 09:00).

Net profit of 1099 Pips (in one day).

See screen shot:

Attached Image (click to enlarge)
Click to Enlarge Name: cmdt_pairs.png Size: 159 KB
No guts, no glory
  • #631
  • Feb 21, 2015 10:05am Feb 21, 2015 10:05am
there was no touch on a .618 at the basket.

lets say it was..one should pick several of this and see what happens overall. plus, how one determines where .618 or any other level is.

and most importantly as felixis said:

"you will never have a trending higher than the highest trending pair
because you add pairs that do not trend or trend weaker or trend the opposite

because of this

the amplitude of the variation of your basket is reduced
but the spread you pay is higher"


but still, i would dig forward and try to see something. at least few occurrences to see what is there
  • #632
  • Feb 21, 2015 10:26am Feb 21, 2015 10:26am
Quoting 21vs7
there was no touch on a .618 at the basket. lets say it was..one should pick several of this and see what happens overall. plus, how one determines where .618 or any other level is. ..
This is just an example on how a basket can be used as an indicator to trade component pairs. Each system will use its own method for entry and exit.

Quoting 21vs7
..and most importantly as felixis said: "you will never have a trending higher than the highest trending pair because you add pairs that do not trend or trend weaker or trend the opposite because of this the amplitude of the variation of your basket is reduced but the spread you pay is higher" ..
This is diversification, which is maybe the main purpose of using the basket in the first place.
No guts, no glory
  • #633
  • Feb 21, 2015 12:15pm Feb 21, 2015 12:15pm
hello
if anyone here is so powerful, and able to help me.

Can someone help me make ea.
LR Forex Freedom Extreme 4.4 is not difficult system for even a novice to understand. This trading system is equipped with arrows who report about the direction of entry into the market. This strategy is suitable for any currency pair, but better to give preference majors. Trade b.etter on time frames M5 and above


Long position:

  1. 1. When a red or yellow arrow pointing up (red arrow - stronger signal)
  2. 2. Indicator RSI (14) must be above level 50
  3. 3. Indicator ODST_Wave is green

Exit (2 versions):

  1. 1. Conservative - when the indicator ODST_Wave will turn red
  2. 2. Aggressive - when an arrow appears pointing down

Short position:

  1. 1. When a red or yellow arrow pointing down (red arrow - stronger signal)
  2. 2. Indicator RSI (14) must be below the level 50
  3. 3. Indicator ODST_Wave is red
  4. Exit (2 versions):
  5. 1. Conservative - when the indicator ODST_Wave turns green
  6. 2. Aggressive - when you see the arrow pointing up

Attached Image (click to enlarge)
Click to Enlarge Name: sddd.PNG Size: 43 KB
  • #634
  • Feb 21, 2015 12:21pm Feb 21, 2015 12:21pm
Quoting abokwaik
{quote} This is diversification, which is maybe the main purpose of using the basket in the first place.
Experience has shown me clearly that its easy to predict a certain currency is going to continue strengthening and/or weakening. The unpredictable aspect is usually figuring out against which other counter currency this strengthening and/or weakening action is going to be most pronounced. Thats how and why I started taking a serious look at Basket trading because when this 'expected' happens, it more than compensates for any 'unexpected' contrary/counter-directional bias action on any of the other pairs in the basket.
  • #635
  • Feb 21, 2015 12:25pm Feb 21, 2015 12:25pm
Quoting abokwaik
OK. So were do we go from here: Test different basket calculation approaches Apply the best approach to the Create_Basket_Script Finalize the first version of Basket Trading Robot
  • #636
  • Feb 21, 2015 12:55pm Feb 21, 2015 12:55pm
Quote
Experience has shown me clearly that its easy to predict a certain currency is going to continue strengthening and/or weakening. The unpredictable aspect is usually figuring out against which other counter currency this strengthening and/or weakening action is going to be most pronounced

and you need a basket to finally trade a single pair
  • #637
  • Feb 21, 2015 12:56pm Feb 21, 2015 12:56pm
Quote
Net profit of 1099 Pips (in one day).
1099 / 9 = 122 pip on average

if you trade only the EURAUD pair with 10 time the position you would have made 244 * 10 = 2440 pips
  • #638
  • Feb 21, 2015 12:58pm Feb 21, 2015 12:58pm
if you want more pip in one day isuggest you put all the instruments available in your mt4 platform
for sure you will end up with 20000 pip in one day

but i have a question, where is your stop ??

you earn 1099 pip for what risk ?

why you use pip ? pip is not important its the money that is
a scalper gain few pip but a lot of money anyway
  • #639
  • Feb 21, 2015 1:01pm Feb 21, 2015 1:01pm
you didnt reply to my post 628 which is the most important

I assume that you will keep going with your inverse calculation

then I enjoin you to give a warning to the guy that will use this script because he may think he is trading an instrument
and he expect to have 0 gain if the price return to the level he made a trade

its not the case with this script
you can end up with a gain or a loss when the price goes back to the entry
  • #640
  • Feb 21, 2015 1:13pm Feb 21, 2015 1:13pm
Quoting felixis
[b]{quote} 1099 / 9 = 122 pip on average if you trade only the EURAUD pair with 10 time the position you would have made 244 * 10 = 2440 pips
I/You wouldn't know in advance that EUR/AUD is going to be the Biggest Winner. Would we ?
No guts, no glory
2 traders viewing now