100% Win Math Grid Ea | Page 57 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Grid - New indicator that replaces chart grid 60 replies

Grid (not grid system) or horizontal lines indicator 6 replies

100% win grid EA 45 replies

Grid, Burst Grid, Grid Risk and Grid Profit 1 reply

Double Time Grid and Pip Grid 0 replies

Attachments: 100% Win Math Grid Ea
Exit Attachments

100% Win Math Grid Ea

  • #1,121
  • May 13, 2010 9:55am May 13, 2010 9:55am
I have thought about re-writing the OrderReliable suite to allow two additional defaulted paramters to be passed in. The first, errorCode, by reference so the client code can get a handle to the error and the second, a verbose variable, that when false, the library functions don't make a peep unless there is actually a problem. Now they chatter too much - "trying this", "apparently successful", blah blah lol

Thanks for pointing that out though.

Viper

Quoting FrozenAir
Just FYi, in your code, you cannot call GetLastError(). Here is a note in the OrderReliable code:

// OrderReliableLastErr()
// Returns the last error seen by an Order*Reliable() call.
// NOTE: GetLastError() WILL NOT WORK to return the error
// after a call. This is a flaw in Metatrader design, in that
// GetLastError() also clears it. Hence in this way
// this library cannot be a total drop-in replacement.

-Frozen
  • #1,122
  • May 13, 2010 11:52pm May 13, 2010 11:52pm
Viper,, Thanks for your reply.

Attach the theory,, seem for me it will survive and will gain lots of profits (ranging & trending).

If you think these theory can be realise in a smart EA, please advice me more.

Thanks again.
Attached Image
  • #1,123
  • May 14, 2010 1:01am May 14, 2010 1:01am
Hey Nu, Any theory (with logical rules) can be made into a smart EA. Unfortunately, the picture you provided isn't descriptive enough to reveal all the rules of the system. Do you have a text document describing it, maybe supported with pictures ? Explain when the grid opens, order placement size and spacing, when to close... etc.

Thanks
Viper

Quoting NuCom
Viper,, Thanks for your reply.

Attach the theory,, seem for me it will survive and will gain lots of profits (ranging & trending).

If you think these theory can be realise in a smart EA, please advice me more.

Thanks again.
  • #1,124
  • May 14, 2010 1:43am May 14, 2010 1:43am
Let say we SELL 0.1 Lot and set the martingale level at 20. The market goes up and we got -20 so we open new Sell 0.2 lot. Then if the market still goes up and the second order is -20 then close all order and open new BUY 0.7 lot.

Then if market goes down reach -20, we open new BUY 0.8 lot. Then if market goes up and we have +20, we close all order.

Set Stoploss 20, Takeprofit 20

Lot 1 = 1 lot
Lot2 = 2 lot
Lot 3 = 7 lot
Lot 4 = 8 lot
Lot 5 = 31 lot
Lot 6 = 32 lot
Lot 7 = 127 lot
lot 8 = 128 lot and so on. it also can be 1,1,4,4,16,16,32,32,64,64...

We can start either BUY or SELL and the lot open will follow.

We have to choose the best martingale level depending on which graf we traded.

Thanks
  • #1,125
  • May 14, 2010 1:32pm May 14, 2010 1:32pm
Quoting tradingviper
I have some additional ideas for the high-water mark also. Just haven't had time to implement, test, implement, test...

If we can figure out a way, once we detect trouble forming, to scale out for a small profit, break-even, or even a small loss most of the time and never a wipe out, we will have truely tamed the beast...

I am very interested in your ideas so please share some of your ideas here.

Viper

When reading through this forum, post # 374 really stuck out in my head. The idea of breakout vs non breakout got me thinking about switching to a non-breakout scenario after the liquidation/high water mark scenario has occurred. The high-water mark code you developed follows this idea by hoping that the fluctuation in price will naturally close out each position as it becomes positive, hoping for a net positive gain.

What I was thinking was:
1.) Pull the plug on all future positions being opened ( switch into liquidation mode, your code deletes all pending orders. I left them all open and simply switched into liquidation mode. Personally I think it doesn't matter what is done here. If you close all positions, one side will be stronger than the other, possibly netting a gain or loss. The same possibility occurs if pending positions are left open, except the potential gain/loss is larger in both directions)

2.) Based on which side of the grid is weaker ( which side will result in a loss due to breakout with no more pending orders being placed), calculate the amount of pips required for the weak side to break even.

3) Trade on a non-breakout scenario until the projected weak side gains are =0. If it breaks the stronger side, we have even larger gains. if it breaks the weak side we end with 0 profit/loss. The risk we assume, is that while trading a non-breakout position, it reverses and breaks out of the grid, netting us a loss on that position.


I threw this code onto your 2.6.2 version of code and it worked well, but I find it inconclusive because of a lack of testing. I found that I typically captured 1-2 of these non-breakout scenarios before the grid went and broke out in either a positive or negative direction. I have been using a 5% watermark on a $1000 test account to trigger these scenarios as much as possible. I have a feeling that if we were actually in a "death trade", it might trigger more non-breakouts to allow us to recover potential lost pips.


Ideas I had to enhance this scenario:

1.) I also have not come up with an algorithm for determing pip size in the non-breakout scenario. I have been using a fixed number (3* minlot). A variable lot size may be able to capture more pips in the beginning, and lower the lot size towards future non-breakout positions.
2.) I have only been trading non-breakout between B1 and S1, the first line in each direction after the pivot. I am wondering if trading between B1 and B2, or S1 and S2, would add further profit potential versus the risks of a reverse swing.
Attached Image (click to enlarge)
Click to Enlarge Name: Non-Breakout Theory.jpg Size: 68 KB
  • #1,126
  • May 16, 2010 9:05am May 16, 2010 9:05am
As no one has responded to my last post about my findings on the Initial pivot price variable, I thought I would just put it out there again as it seems there is some new activity on the thread and that Viper is about again.

Anyhow as I previously mentioned, Iam finding that when the price is moving about quite quickly it can jump over or gap across the initial pivot price that has been set and not trigger the start of that grid. I have looked into the code and have tried to edit it so that if the ask price = initial pivot price, + or - X pips then open the grid based one the initial pivot price. Iam using ver 2.6.1 and I think the line that I need to edit is 1370

" if( InitialPivotPrice > 0.0 && Ask != InitialPivotPrice ) "

But I don't know how to write this in the correct way for it to work as every think I try wont compile.

Is there anyone out there who could help me out with this or perhaps if Viper is reading this could you please consider this in any future updates.

Thanks
Darran
  • #1,127
  • May 20, 2010 10:27am May 20, 2010 10:27am
Hi Dazz,

I have seen this happen before and thought about adding an "anywhere close" rule to enter. But frankly, it doesn't seem to happen that often because on a micro scale, price is always jitterring, so it is bound to hit. (I use a 5-digit criminal as well)

Another issue to consider is that gaps and spreads are not friendly to this system. So if price is moving that fast, perhaps it is better to sit on the sidelines for a bit.

I have mixed feelings about it. The code is not difficult, so if/when I do an upgrade I will consider this one of my potential mods.

Thanks for the input and continued brain power applied :-)

Viper


Quoting Dazzafx
As no one has responded to my last post about my findings on the Initial pivot price variable, I thought I would just put it out there again as it seems there is some new activity on the thread and that Viper is about again.

Anyhow as I previously mentioned, Iam finding that when the price is moving about quite quickly it can jump over or gap across the initial pivot price that has been set and not trigger the start of that grid. I have looked into the code and have tried to edit it so that if the ask price = initial pivot price, + or - X pips...
  • #1,128
  • May 20, 2010 10:47am May 20, 2010 10:47am
Thats interesting. I am not saying your idea does not have merit, because it does, but I would be worried about it. After a prolonged sideways movement, causing the death trade in the first place, the market is very prone to strong breakouts. In fact it seems (Murphy's Law?) that everytime we give up hope and begin taking corrective action, the market breaks strongly right after !!

In the current High-Water emergency procedures, after closing all pending orders, it tries to liquidate the profitable orders on the strong side. Based on the observation that the explosion of orders that puts us over the edge often occurs by hitting a swell of pendings sitting on the outer level. So the inner levels should have profitable positions.

We have to weaken the strong side to prevent a reversal from killing us. Another thought I had was to pull in the TP/SL line of the weak side. If we have more buys than sells, as soon as price moves down past the last sell level, every pip down does more damage than good (more buys than sells), so why not just pull it in to the last level ?

Anyway, great thinking. Keep it up. It may be impossible to ever make this "safe", but it certainly doesn't hurt to think about it.

Viper

Quoting FrozenAir
When reading through this forum, post # 374 really stuck out in my head. The idea of breakout vs non breakout got me thinking about switching to a non-breakout scenario after the liquidation/high water mark scenario has occurred. The high-water mark code you developed follows this idea by hoping that the fluctuation in price will naturally close out each position as it becomes positive, hoping for a net positive gain.

What I was thinking was:
1.) Pull the plug on all future positions being opened ( switch into liquidation mode, your code deletes...
  • #1,129
  • Edited 9:54am May 21, 2010 9:52am | Edited 9:54am
Hello to all,

I was trading live using the williama's trading system, the EA coded by AutoFX found in forex peace army. Seems good on the first few weeks until I got hit by the whip saw of eurjpy May 19, 2010. I stopped the EA and did some research and I found this thread. I backtested this EA from May 17 to May 21, 2010 eurjpy and was also hit by the death trade but with minimal loss compared to the EA that I was using. Is there anyone trading the ViperGridEA live? Would you mind to share your settings? Thanks.

Thanks Viper for these great improvements.

While there's an existing grid, is it possible for the EA to open another set of grid when the last level is triggered (e.g. the level 3 for 3 levels) instead of waiting for the sl/tp to get hit? In this case, as soon as the first grid hit the sl/tp, the level 1 of the second grid is already triggered. But when the price reversed without hitting the sl/tp level of the first grid, the second grid will take profit when the price hit the first grid's level 1 of the other side of the grid.

Or maybe I can manually open a second grid by attaching the EA to another chart of the same pair. Is it possible for the EA to enter the market only when a certain price of a currency pair is reached?

Thanks and sorry if my questions were asked before. This thread is more than two years old and I just can't wait to find the answers.
  • #1,130
  • May 22, 2010 6:22am May 22, 2010 6:22am
I am trading live, I started with just under 2200 about 2-3 weeks ago. I have not been the market everyday, as i took a couple breaks here and there. The big move ( occuring in the beginning of the curve), was because i had .40+ buy lots open on a friday close. When the market jumped on sunday, i made 300 dollars on the trade. Other than that you can see its a slow climb. I have a custom version of vipers grid, with slightly different watermark characteristics. I also coded it to keep margin calculations specific to the individual currency/graph. That way if one of the 10 currency pairs i am trading hits a watermark level, it does not trigger the other currency pairs to go into liquidation or low water mark mode. With that graph, most of the trades until the very end were using the original code where, when a liquidation mode was hit, it would take all my positions on all 10 graphs into liquidation mode. I definitely lost some decent money doing that. So hopefully that graph should be more positive after some of these changes. my watermark levels are 5%/15% and the IRegr filter is turned on ( i used the trend filter for a bit, so most of the graph is using the trend filter. I found better results on backtesting with iRegr).

Hope that helps. It still needs a lot of work to run profitability and unwatched. The watermark strategy has to be perfected. I have a couple mathematicians I went to college with working on it this weekend. We'll see what they come up with. I am trying to get something parabolic from them, more accurately reflecting the oscillating nature of the market. But i'm sure it can be solved exponentially.



Gross Profit:3 180.81Gross Loss:2 575.58Total Net Profit:605.23
Profit Factor:1.23Expected Payoff:0.73
Absolute Drawdown:164.25Maximal Drawdown:308.80 (10.16%)Relative Drawdown:10.16% (308.80)

Total Trades:831Short Positions (won %):389 (71.72%)Long Positions (won %):442 (75.79%)
Profit Trades (% of total):614 (73.89%)Loss trades (% of total):217 (26.11%)
Largestprofit trade:44.43loss trade:-61.70
Averageprofit trade:5.18loss trade:-11.87
Maximumconsecutive wins ($):49 (149.95)consecutive losses ($):12 (-139.03)
Maximalconsecutive profit (count):569.79 (19)consecutive loss (count):-139.03 (12)
Averageconsecutive wins:5consecutive losses:2


Quoting luiboy
Hello to all,

I was trading live using the williama's trading system, the EA coded by AutoFX found in forex peace army. Seems good on the first few weeks until I got hit by the whip saw of eurjpy May 19, 2010. I stopped the EA and did some research and I found this thread. I backtested this EA from May 17 to May 21, 2010 eurjpy and was also hit by the death trade but with minimal loss compared to the EA that I was using. Is there anyone trading the ViperGridEA live? Would you mind to share your settings? Thanks.

Thanks Viper for these great...
Attached Image (click to enlarge)
Click to Enlarge Name: graph.gif Size: 7 KB
  • #1,131
  • Edited 6:03pm May 24, 2010 5:48pm | Edited 6:03pm
Quoting tradingviper
Anyway, great thinking. Keep it up. It may be impossible to ever make this "safe", but it certainly doesn't hurt to think about it. Viper
Dear Viper and all,
I've gone through post 1 to 41 so far and I'm really interested in the concept, implementation and ideas I have read.

However, (and that is with version 2.1 and previous), the grid strategy involving doubling lots at opposite side upon pending trigger, or opening 1-2-3 units at respective opposite levels works ONLY if the grid is uniform .... and ONLY without taking into account the spreads ... AND choosing a grid increment that will counter trend the cumulative effect for spread..

According to my math, changing any of the parameters and creating a non-uniform grid will screw the "holy grail math" behind it. This means every ideas and implementation up to post 41, that creates a non-uniform grid is creating a flawed system .... which explains why sometimes the system performs a complete set and looses money. As well, the cumulative effect of Spread Versus Increment Size may quickly add odds against the strategy if not taken care of properly.

To my math, most of it is handling correctly Lot Size Calculation for pending orders (and this is not simple). Eventually you will have discovered this after post 41, in which case I'm sorry to come up with oldies ... if not, there is some work to be done to satisfy the "holy math".

On the other side, the watermark is a great concept to free margin, that does not interfere with the strategy as long as new orders are still created. However, playing with "pending orders" is not the right way to go if you are willing to extend the "watermark" concept in order to further reduce free margin. - eventually something else that you may have said after post 41.

I code for myself and do not share my work. However, I like to share concepts and ideas to build upon and improve strategies. If you are interested, I'm willing to share further these limitations. If not, thank to every one for the great work.

PS: Using Viper version 2.1, preventing the limitations I mentioned and using the Time parameters I get a system that passed 2009 up to now and turns 2500$ into 8k$ ... not bad ... and quite reliable if you have 500:1 leverage at 0.01 lots with no limits as per amount of open lots.

Regards,
XmPh.
  • #1,132
  • May 24, 2010 6:06pm May 24, 2010 6:06pm
Quoting FrozenAir
I have a couple mathematicians I went to college with working on it this weekend. We'll see what they come up with. I am trying to get something parabolic from them, more accurately reflecting the oscillating nature of the market. But i'm sure it can be solved exponentially.
FA, do you and your mathies corroborate my above statements ?
  • #1,133
  • May 26, 2010 9:57am May 26, 2010 9:57am
Thanks. I did try to remove the variability that would result in a non-uniform grid. That was the original enhancement of ViperGrid.

Please share your Time parameters settings that contributed to this stability. Why are you using version 2.1 instead of a later version ?

Viper

Quoting xmph
Dear Viper and all,
I've gone through post 1 to 41 so far and I'm really interested in the concept, implementation and ideas I have read.

However, (and that is with version 2.1 and previous), the grid strategy involving doubling lots at opposite side upon pending trigger, or opening 1-2-3 units at respective opposite levels works ONLY if the grid is uniform .... and ONLY without taking into account the spreads ... AND choosing a grid increment that will counter trend the cumulative effect for spread..

According to my math, changing any of the...
  • #1,134
  • May 26, 2010 10:25am May 26, 2010 10:25am
Quoting tradingviper
Thanks. I did try to remove the variability that would result in a non-uniform grid. That was the original enhancement of ViperGrid. Please share your Time parameters settings that contributed to this stability. Why are you using version 2.1 instead of a later version ?

Viper
Hi Viper,

I used v2.1 as, up to page 41 (that is post 601) that is the version I'm comfortable with for my testing .... anyway, I'm recoding mine for the reasons/limitations I stated earlier. I believe (and I may be wrong) that pending orders is not the way to go (although it prevents potential broker lag time issues - that is the interest I see ... so I may mix pending and others)

I did not see the part of the code that tried to manage the non-uniformity due to spread (or other factors). Maybe my mistake or presented after post 601. I have to continue reading to be totally up-to-date. To speak correctly, I should say that non-uniform grid should be handled correctly considering they are appropriate as well, however every lot management I've seen so far in this thread are design for the specific case of uniform grids.

For the parameters:
Open at Frankfurt, and no more than 3hours after London Opens (I chose 5-7am UK time). 3 levels, 40 increment. That's it, and yes, I'm trying to get the edge of a London Breakout ... stupid but quite efficient.

Now, for this to work you need leverage 500:1, initial deposit 2500$, minimum lots of 0.01 (I did not try %equity) and no maximum limits in terms of open orders. (no FXCM, no Alpari ... check your brokers)

Still, in this configuration you'll have some "loosing sets" but overall a nice return.

I checked the log file and no margin call occurred (that was the important thing)

regards,

PS: Are there still active ppl in the thread beside you .... I mean brain of coding active ?
  • #1,135
  • May 27, 2010 4:02am May 27, 2010 4:02am
hello, been looking up this type of trading recently and its seems interesting.
what if you have a secondary system for choosing direction of price. i am looking for winning streaks. pyramiding is fine but i would like to actually double the lot size after every winning trade. up to how many levels i want. i have backtested some systems and the highest winning streak i get is 8. that equates to 12800 profit with 0.1 lot start.

applying this grid when we want with our other system. i beleive it makes it a winning system.

an easy system i have backtested is buy if open > last close and sell is open < last close

it gave me some nice streaks. a system with nice streaks makes money, right?

can someone backtest this? help in coding an ea?
much appreciated
  • #1,136
  • May 27, 2010 4:18am May 27, 2010 4:18am
im backtesting as is grid ea. it s equity curve rises but at the end it loses everything and more than initial amount deposited. why does it do that at the end?????
  • #1,137
  • May 27, 2010 7:14am May 27, 2010 7:14am
I think you guys are getting carried away in the wrong direction. Please read every page of the thread before further commenting.

Forex Flash has the key with a purely mathematical implementation of an exit strategy. With a proper exit strategy, and some entrance strategies that make sense, this system does work. As long as the gains> losses and the gains are better than a market return in a safe security.

I have found that the .01 lot/ $10,000 is a little too conservative. Without a proper exit strategy, the more margin the better. With the incomplete exit strategies i've been developing, i've been profitable with a .01 lots on $1000. Once over 10% of the account is being engulfed, a proper exit strategy will salvage most of that. The problem with running an exit strategy with an amount as high as $10000, is that if we have to take small losses in our mitigation strategy, those losses are astronomical compared to the returns gained from .01 lot profits. on $1,000 , this strategy can easily return 5-10% a month per currency being traded with death trades mitigations kicking in once 10% of margin is consumed.

Again i'll keep everyone posted with my progress, I am almost done coding up a new mitigation strategy that I have been working on the past two weeks.
  • #1,138
  • May 27, 2010 7:22am May 27, 2010 7:22am
Quoting andreasp70
im backtesting as is grid ea. it s equity curve rises but at the end it loses everything and more than initial amount deposited. why does it do that at the end?????

There are 70+ pages on this forum for a reason --> to prevent that.
  • #1,139
  • May 28, 2010 7:06am May 28, 2010 7:06am
Quoting andreasp70
what if you have a secondary system for choosing direction of price. i am looking for winning streaks. pyramiding is fine but i would like to actually double the lot size after every winning trade. up to how many levels i want. i have backtested some systems and the highest winning streak i get is 8. that equates to 12800 profit with 0.1 lot start.
.... remember: Pigs get Fat, Hogs Get Slaughtered!

.......
Quoting FrozenAir
I think you guys are getting carried away in the wrong direction.
.... Indeed.

Andre.., why would you martingale a martingale ? Why would you add a "trend notion" in a system build on a "no trend notion" ?

However, as FrozenAir mentionned, Incomplete Exit strategies are a valid option and I do understand why FrozenAir was able to make this work on a 1000$ starting account ..... however I bet you have L=500:1 to manage that.

My analysis and coding up to now have led me to "mitigate" through different whipsaw scenarios ... specifically, some scenarios are "killing/risky ones" depending on your available margin. I believe "mitigation" must evolve with available margin ....

My current mindset is to consider the grid as a predictor, that is you can calculate ahead what will happen .... so better handle now the worst case scenario based on your margin and think ahead to strategize "incomplete exit strategies" based on your current margin.

Are you following such principle FrozenAir ?

As well, I will post my results and keep ppl updated as soon as I've a good implementation of my concepts.

regards,
  • #1,140
  • May 28, 2010 11:12pm May 28, 2010 11:12pm
Hello,

This is my past week live trading result (24th to 28th of May) for using the latest ViperTrading EA. I'm using only one currency pair (eur/jpy), so far it never activated the low water mark mitigation that i set to 5%.


Gross Profit:8 611.57Gross Loss:6 955.97Total Net Profit:1 655.60
Profit Factor:1.24Expected Payoff:11.19
Absolute Drawdown:927.14Maximal Drawdown:1 222.17 (4.88%)Relative Drawdown:4.88% (1 222.17)

Total Trades:148Short Positions (won %):69 (72.46%)Long Positions (won %):79 (82.28%)
Profit Trades (% of total):115 (77.70%)Loss trades (% of total):33 (22.30%)
Largestprofit trade:111.06loss trade:-591.24
Averageprofit trade:74.88loss trade:-210.79
Maximumconsecutive wins ($):16 (1 255.09)consecutive losses ($):2 (-514.41)
Maximalconsecutive profit (count):1 255.09 (16)consecutive loss (count):-591.24 (1)
Averageconsecutive wins:4consecutive losses:1
Attached Image (click to enlarge)
Click to Enlarge Name: DetailedStatement.gif Size: 6 KB
1 trader viewing now