Replies

TradingDeveloper
15 Jul 2013, 17:52

They will tell you that this will be implemented in the future and you will wait forever :-)

But I heard about  some software that may be helpful:

http://www.scyware.com/

Good luck

 


@TradingDeveloper

TradingDeveloper
29 Jun 2013, 23:34

RE: RE: RE: RE: RE:
lec0456 said:
MrTrader said:
lec0456 said:
TradingDeveloper said:
lec0456 said:

.2 pip spread$

$30/million commission

The stop loss varies but starts at 100pips.  if price moves slowly against the position it can hit the stoploss but the robot uses an indicator that detects breaks against a position and then makes a corective trade.  Thats why you see interuptions in the graph pattern.

Hi,

Thanks for the info, what is your take profit exactly?

Based on my experience I see some potential problems in your backtesting:

1. $30/million commision is not a real value, the broker I know charge $65-$80 / Million, do you know any charging $30?

2. Depending on your take profit the spread may have an impact in your backtesting results, if you have a low take profit the results may be much different in a real scenario.

4. If your take profit is low I would say 100 pips stop loss is too high and involves some risk, and if you say your stop loss increases its even more risky.

I see your last trade in the Screen shot was about 10 pipsa profit, so I assume your take profit is low.

I would suggest you to adjust the commisions and place a random spread between 0 and 3 and check the results again.

 

I use 7.5 pip take profit for the initial trade.  But corrective trades will be executed in 2 varieties. I call them MAX and normal trasdes.  A MAX corrective trade uses 75pips and a normal corrective trade uses the loss of the origonal trade plus 7.5 pips(the stoploss would be the same as the TP for corrective trades).  There is criteria I use to decide if I go MAX or normal on the correction.  I don't use stop loss greater than 100

As far as the Brokers, yes there are brokers in the $30/million range.  Recently some have dropped commisions, here is a list I complied a few onths ago, but be warned it may not up to date:

Broker Location  Commission  Time  Cost 
Liquid Markets Cyprus  $             45.00 Cyprus  $    11.72
Trader'sWay Dominica  $             35.00 Cyprus  $       9.27
ICMarkets Sydney  $             30.00 GMT  $       7.90
FXPro Cyprus  $             65.00 Cyprus  $    17.14
Divisa NewZeland  $             30.00 GMT  $       7.90
RoboForex NewZeland  $             20.00 Cyprus  $       5.20
FiboGroup Cyprus  $             30.00 Cyprus  $       7.90

 

 

Hi lec0456,

 

What is this robot called? and will you make it available for download?

 

Thanks,

The name of the robot really doesn't matter.  But I have spent over a year developing the robot and custom indicators it uses, so I wouldn't be inclined to give it away.  However, if another devveloper had a robot that intrigued me, I would consider trading. 

lec0456, contact me at tradingdeveloper2013[at]gmail.com I have something you might be interested in


@TradingDeveloper

TradingDeveloper
28 Jun 2013, 15:26

3.*


@TradingDeveloper

TradingDeveloper
28 Jun 2013, 11:38

RE:
lec0456 said:

.2 pip spread$

$30/million commission

The stop loss varies but starts at 100pips.  if price moves slowly against the position it can hit the stoploss but the robot uses an indicator that detects breaks against a position and then makes a corective trade.  Thats why you see interuptions in the graph pattern.

Hi,

Thanks for the info, what is your take profit exactly?

Based on my experience I see some potential problems in your backtesting:

1. $30/million commision is not a real value, the broker I know charge $65-$80 / Million, do you know any charging $30?

2. Depending on your take profit the spread may have an impact in your backtesting results, if you have a low take profit the results may be much different in a real scenario.

4. If your take profit is low I would say 100 pips stop loss is too high and involves some risk, and if you say your stop loss increases its even more risky.

I see your last trade in the Screen shot was about 10 pipsa profit, so I assume your take profit is low.

I would suggest you to adjust the commisions and place a random spread between 0 and 3 and check the results again.

 


@TradingDeveloper

TradingDeveloper
23 Jun 2013, 00:45

RE:
cAlgo_Development said:

We've checkked your bars and can confirm that the reason is the minor difference in prices between live and demo account. To make it similar pleaseensure you do this:

  • Open cTrader on PC1
  • Login to Live account
  • Right click on a chart, press "Refresh" in context menu
  • Open cTrader on PC2
  • Login to Live account
  • Right click on a chart, press "Refresh" in context menu

Prices will be exactly the same after this procedure.

Hi,

 

I'm not being able to go back to that date (4/6/2013) so I can't confirm. (The minimum date I can go back with 1Minute is 21st June).

Something I found also is that the backtesting results are not the same as during Live trading. Not talking about differences because of spread or commissions but trades that are fired on Live but not while backtesting the those Live dates.

 

The robot I'm using uses Ticks for trading not Bars, Is this normal?

Thanks


@TradingDeveloper

TradingDeveloper
23 Jun 2013, 00:28

3 questions:
 
1. Whats the spread you are using in your back testing?
 
2. Whats the commissions per Million your are using?
 
3. Whats the SL you are using?

@TradingDeveloper

TradingDeveloper
09 Jun 2013, 16:48

RE:
cAlgo_Development said:

We have investigated this issue and most probable answer is that bars are a bit different because you use live and demo accounts. Theoretically prices are supposed to be the same or very-very close, but on practice prices for different types of accounts are processed on different physical servers. Although server settings are the same sometimes prices can differ. If you try to use the same account type (do not switch, or completly refresh charts after you switched account) prices will be the same.

Hi there,

 

Seems there is still some problem.


For the same times (04/06/2013 16:29 and 16:30) there are still some differences.

This time I cleared cache and I logged in with the same live account in different computers, this is what I have:

Computer 1

 

Computer 2:

There is still a difference... any ideas?

Thanks


@TradingDeveloper

TradingDeveloper
04 Jun 2013, 20:44

RE:
cAlgo_Development said:

Can you tell us the name of broker, exact symbol and bar time where such diiference appears.

I made some tests on 2 pcs today using cAlgo, here is what I got:

Symbol: EURJPY

Broker: FxPro

Bar time: 04/06/2013 16:29 and 16:30

Computer 1:

16:29

16:30

Computer 2:

16:29

16:30

We can clearly see some differences on these bars: note the close price and volume at 16:29 and the Open and Volume at 16:30 (Note there is no spike on this bar on Computer 1).

Both computers were running at the same time.

Do you have any explanation for this?

Could this have something to do with the capacity of the computer / Network activity? How are the price movements processed exactly by the calgo client app?

Thanks


@TradingDeveloper

TradingDeveloper
28 May 2013, 11:54

Hello,

I'm experiencing a simillar issue.

I have a calgo robot running live with a static signal. This signal fires on Live but when I backtest the dates where the signal was fired on live there is nothing on the backtest. (I also tried to clear the cache as mentioned but it didn't help)

Also in the backtesting the signal fires for some dates but nothing happened on Live.

I assume there is something wrong with the backtesting.

@cAlgo_Development please let me know if some sort of fix is going to be applied for this kind of issues, its kind of frustrating when backtesting is not exactly what it should be and its difficult to rely on this information.

Some possible useful information: 

- Code is executed OnTick event

- Executing at 1 Minute

Any ideias what could be the issue?

Thanks


@TradingDeveloper