Wondering why no trade was placed. My user inputs match what's on the chart.
Attached Image (click to enlarge)
Vegas Algorithm auto-trader by Steve and Squalou
- #124
- Nov 22, 2010 4:45am Nov 22, 2010 4:45am
- | Commercial User | Joined Apr 2007 | 8,331 Posts
Steve
Runnig this on demo at mo and i still get the order modify failed with error(130) Invalid stops.
Ive checked settings and Criminal ECNs is set to true
I repeat: I have not been able to replicate this problem on my Alpari UK demo; the problem is at your end. It doe not matter how many times you tell me about it. There is nothing I can do.
- #130
- Nov 23, 2010 4:13am Nov 23, 2010 4:13am
- | Commercial User | Joined Apr 2007 | 8,331 Posts
This was originally post 1 in this thread. The system proved to be unprofitable and most people have abandoned it. Everything you need is here for those who still want to try Vegas' system.
_____________________________________________________________________________________
Vegas
We are indebted to quibernau for bringing this strategy to my attention. Without this, the robot would not have been born.
Thanks to Squalou, we can forget the four indicators mentioned throughout Vegas's zip; sq created the universal indi that replaces them all.
There are two versions of this indi attached. The only difference the auto-trading version and the manual version is that it does not show its values on the screen; this would clash with V's screen display and cause flicker. I am going to try to bring the code the indi uses into V, but that is for later.
The manual-trader version is for those of you who want to trade this system manually. Ignore the Alerts input; there is no code to implement an alert (there never was) so I have asked sq to add this when he has time.
V shows the values it is working with on the screen. These are only so I can check that I have coded things correctly. You can ignore them and I shall probably remove them at a later stage.
Default tp, sl, be and js inputs are guesswork; time will tell which of these work best. Watch the lot size; it is set to 1 because I am developing this on a £50,000 Alpari UK demo just for fun.
I cannot guarantee the PSAR closure code is correct. The logic caused me pain, so it could be faulty.
A note about the H4 trend: you will sometimes see the trend described as 'None' and Vegas will not be trading. The bot both checks which side of the dark goldenrod the yellow line occupies and the direction it is pointing. If the direction is wrong, V regards the trend as untrustworthy.
DO NOT USE THIS ROBOT ON LIVE ACCOUNTS. Aside from V being at the testing stage, we have no actual guarantee that the strategy it trades is a good one.
No robot close
Some contributors feel that Vegas is closing down trades too soon. The 'no robot close' version turns off Vegas' ability to close trades and relies on stop loss, take profit or the various mptm-style management features to close trades.
Have fun.
_____________________________________________________________________________________
_____________________________________________________________________________________
Vegas
We are indebted to quibernau for bringing this strategy to my attention. Without this, the robot would not have been born.
Thanks to Squalou, we can forget the four indicators mentioned throughout Vegas's zip; sq created the universal indi that replaces them all.
There are two versions of this indi attached. The only difference the auto-trading version and the manual version is that it does not show its values on the screen; this would clash with V's screen display and cause flicker. I am going to try to bring the code the indi uses into V, but that is for later.
The manual-trader version is for those of you who want to trade this system manually. Ignore the Alerts input; there is no code to implement an alert (there never was) so I have asked sq to add this when he has time.
V shows the values it is working with on the screen. These are only so I can check that I have coded things correctly. You can ignore them and I shall probably remove them at a later stage.
Default tp, sl, be and js inputs are guesswork; time will tell which of these work best. Watch the lot size; it is set to 1 because I am developing this on a £50,000 Alpari UK demo just for fun.
I cannot guarantee the PSAR closure code is correct. The logic caused me pain, so it could be faulty.
A note about the H4 trend: you will sometimes see the trend described as 'None' and Vegas will not be trading. The bot both checks which side of the dark goldenrod the yellow line occupies and the direction it is pointing. If the direction is wrong, V regards the trend as untrustworthy.
DO NOT USE THIS ROBOT ON LIVE ACCOUNTS. Aside from V being at the testing stage, we have no actual guarantee that the strategy it trades is a good one.
No robot close
Some contributors feel that Vegas is closing down trades too soon. The 'no robot close' version turns off Vegas' ability to close trades and relies on stop loss, take profit or the various mptm-style management features to close trades.
Have fun.
_____________________________________________________________________________________
- #131
- Nov 23, 2010 4:26am Nov 23, 2010 4:26am
- | Commercial User | Joined Apr 2007 | 8,331 Posts
The focus of this thread has changed, so I have re-done page 1.
This thread is now about developing PH's ideas. Keep 'em coming guys. Come to some firm conclusions and either I or sq will code them.
This thread is now about developing PH's ideas. Keep 'em coming guys. Come to some firm conclusions and either I or sq will code them.
Thanks SQ. What about SuperSignal? I watched it on strategy tester and didn't see any problems, but woke up this morning to a lot of arrows pointing the wrong way.
Mike
SuperSignal is just like any zigzag: it has a big weakness when a strong trend develops like in the past days on EU/GU, or the past few hours with the "fantastic" 200 pips drop on EU:
just like zigzag, it will leave you with a long stream of wrong-way signals because retracements were too short (did not have enough bars) to trigger the opposite (good) direction signal/arrow.
This is because looks for the highest high or lowest low over "dist2" bars, and will throw a signal as soon as the current price IS the highest high or lowest low of that period...
and it will signal (and alert) as long as this keeps being the case.
During a trend like today, almost every single M5 bar had an "up" arrow below it... and it kept doing this all along the 200 pips drop
It is really a price exhaustion indicator, just like zigzag, it is not a trend indicator.
If you pull a Stochastic(21,1,1), (or even better: a stochasticLow(21,1,1))
(21 is the default "dist2" value of SuperSignal), you'll see that it remained in overbought zone (even nearing 0) all along, meaning that we ARE in a strong trend.
Stochastic will give you BOTH swing points AND trend indications.
When it crosses below 20 or above 80 (you can play with those limits), it is commonly interpreted as a price exhaustion signal ("overbought/oversold" condition), but this is only to tell you that you should PREPARE for a reversal;
it is only after confirmation, when stoichastic re-enters inside the 20~80 zone, that you can take it as a reversal (SWING) signal.
It tells you that you are in a TREND when it STAYS in the OB/OS zone (below 20 or above 80).
The best is to use a StochasticLow and StochsticHigh (thanks Mr. JohnnyBSmart) rather than the regular Stochastic: when StochasticLow is nearing or at 0 and stays near 0, (or StochasticHigh is nearing or at 100 and stays near 100), then you are in good trend (lower lows or higher highs).
However, when tight ranging conditions start, stochastics will start to get "noisy": it will jump up and down quickly because its high-low range is reducing, and price is diving into its "background noise".
That's the limit of the stochastics.
SQ
The focus of this thread has changed, so I have re-done page 1.
This thread is now about developing PH's ideas. Keep 'em coming guys. Come to some firm conclusions and either I or sq will code them.
Trying to get the EMA right is like playing Whack-a-Mole. How about a version with an 'OnlyTradeLong' or 'OnlyTradeShort' option in the user inputs, based on each user's own interpretation of trend direction, since it's so subjective anyway, and eliminate the EMA altogether?
Mike
- #135
- Nov 23, 2010 1:46pm Nov 23, 2010 1:46pm
- | Commercial User | Joined Apr 2007 | 8,331 Posts
Be careful what you ask for.
Trying to get the EMA right is like playing Whack-a-Mole. How about a version with an 'OnlyTradeLong' or 'OnlyTradeShort' option in the user inputs, based on each user's own interpretation of trend direction, since it's so subjective anyway, and eliminate the EMA altogether?
Mike
Thanks.
Hey Peter. Man, those charts look so familiar . . . . hah, found it. Look here:
http://www.forexfactory.com/showpost.php?p=4161627
Squalou and I tried to make Alf & Alma work together but it just wasn't happening. The problem: whipsaws.
Alma follows price so closely, that it became sort of redundant. ATM, I'm trying to get around weak crosses by setting the MinimumCrossPips higher. I've got it at 10 right now. Default is 1. I consider Lou G an expert on Alf, and one of his signals is a strong move through Alf. So if we get a candle crossing Alf with some energy and then enter only after a subsequent candle surpasses its high/low, we might be closer to solving our head fake issues. Of course, this idea is based on my understanding of the MinimumCrossPips input. It's quite possible I'm confused.
Mike
http://www.forexfactory.com/showpost.php?p=4161627
Squalou and I tried to make Alf & Alma work together but it just wasn't happening. The problem: whipsaws.
Alma follows price so closely, that it became sort of redundant. ATM, I'm trying to get around weak crosses by setting the MinimumCrossPips higher. I've got it at 10 right now. Default is 1. I consider Lou G an expert on Alf, and one of his signals is a strong move through Alf. So if we get a candle crossing Alf with some energy and then enter only after a subsequent candle surpasses its high/low, we might be closer to solving our head fake issues. Of course, this idea is based on my understanding of the MinimumCrossPips input. It's quite possible I'm confused.
Mike