100% Win Math Grid Ea | Page 51 | 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

  • #1,001
  • Dec 10, 2009 12:02am Dec 10, 2009 12:02am
Quoting tradingviper
That is exactly what the low water mark will do for you. If you want to bail at the first sign of profit after 60 orders is reached set the low water mark for 60 if using 1 level, 30 if using 2 levels or 20 if using 3 levels. After it broaches each low water mark multiple, it also lowers the profit goal a little which theoretically slows the accumulation of new orders.

Viper

Viper, I see how this is similar. Making less aggressive orders, etc...But doesn't the low-water mark feature still require the orders to hit their TP?

I'm thinking having all orders closed once the set hits a proft value set by the user (activated after a certain number of lots or orders are opened)...

I don't like to hear of someone losing $5k due to a spread issue - I thought it was more robust than that. Is this because once one order hits a TP, all other orders close and this results in a possible loss (because the grid is no longer aligned/incremented properly to have all TP's at the same price?)?
  • #1,002
  • Dec 10, 2009 12:12am Dec 10, 2009 12:12am
Hi Ogi,
Your point number one is not quite true. The implementation I did does require the slope to be opposite to the penetration, it just does not take the steepness of the slope into consideration. Since I now have the code to I-Regr grid, I can probably add that fairly quickly. Maybe I can get it done this weekend.

What input values are you using to iRegr ?

Point number is also not quite correct. ViperGrid begins compensating at the outer level and continues down through all levels if needed to achieve the desired profit goal. If it only has to place an order at level 3, then it stops there. This is actually better because by not putting new orders at level one, there is less chance that we'll whipsaw opening even more orders... I have watched this many many times and most people with a lot of experience with this concur. However, if you can come up with a better model that remains profitable, lets explore that. Seriously.

Thanks for all your efforts and contributions,
Viper

Quoting ogidimi
1. Viper grid enters when price is out of the iRegr channel, at points 1,2,3,4,5,6,7 and 8 regardless of the slope of the channel. I-Regr Grid enters only at points 2, 4 and 6, hence the lower DD....
  • #1,003
  • Dec 10, 2009 12:24am Dec 10, 2009 12:24am
Hi Jay,

No once the low water mark is triggered, the profit goal is reduced (eventually to zero) and as soon as that profit amount is achieved in the current set, it will close the entire set for that reduced profit. Every multiple of the low water trigger (number of positions) will lower the desired profit for that trade and exit as soon as that is hit. I think that is exactly what you were asking for. :-) I'm just a few months ahead of you :-)

ViperGrid is much more robust than the old grids. Keeps the grid in alignment when placing new orders regardless of the spread but if the spreads are crazy at the time of closing you will suffer reduced profit or even a loss. It is highly unlikely that you can lose large sums by the spread.

The only other defense is to use stealth mode so a stop doesn't get triggered when it shouldn't and use wider increments. Again, if the spread is wide at the time when the set closes you will lose some of your profits. That's just life.

Viper


Quoting optojay
Viper, I see how this is similar. Making less aggressive orders, etc...But doesn't the low-water mark feature still require the orders to hit their TP?

I'm thinking having all orders closed once the set hits a proft value set by the user (activated after a certain number of lots or orders are opened)...

I don't like to hear of someone losing $5k due to a spread issue - I thought it was more robust than that. Is this because once one order hits a TP, all other orders close and this results in a possible loss (because the grid is no longer aligned/incremented...
  • #1,004
  • Dec 10, 2009 12:55am Dec 10, 2009 12:55am
Quoting tradingviper
Hi Jay,

No once the low water mark is triggered, the profit goal is reduced (eventually to zero) and as soon as that profit amount is achieved in the current set, it will close the entire set for that reduced profit. Every multiple of the low water trigger (number of positions) will lower the desired profit for that trade and exit as soon as that is hit. I think that is exactly what you were asking for. :-) I'm just a few months ahead of you :-)

ViperGrid is much more robust than the old grids. Keeps the grid in alignment when placing new orders...
Cool, Viper. I always appreciate your thorough explanations. I'll try backtesting with the low-water mark again. Previously it appeared to be a hinderance to the backtesting results so I stopped using it - I will take another look.

Also, I am still getting funky backtests with version 2.5 and on with all the "extras" you've add turned off...previous versions work great. I know I already posted this "complaint" earlier, but you definitely didn't change any of the entry logics, right?

Thanks again for all your hard work. This still seems like a great EA, especially if you were to use it combined with other setups or price action; such as:

1. when a pinbar appears on a swing high or low
2. play a breakout when there is consolidation of price
3. some strategies play a weekly breakout (Ace's 4H breakout)
4. initiate on stong S/R
5. play the 123-setup and initiate a grid on a retrace/pullback (take a look at an EA I helped build: http://www.forexfactory.com/showthre...59#post3293259). Actually, this one I would be very interested in!

These ideas may be better than relying on lagging indicators?? For example, it seems just as the auto-ATR is narrowing volatility explodes (which is good for our grid), but if ATR is expanding and price narrows then we are in trouble! Same with volatility and trend filters - as soon as those filters say it is good to enter a grid, it changes! That's just a fact of life being a speculator!

In summary, I think the grid will fail if used randomly or depending on price movement alone - which is why Hanover left this thread long ago. But the grid offers some "fail-safe" if you are mistaken in your technical analysis! Phew, long post...
  • #1,005
  • Dec 10, 2009 12:56am Dec 10, 2009 12:56am
Haven't heard from Plasmapants in awhile...he must be off making millions somewhere with floating multigrids!
  • #1,006
  • Dec 10, 2009 4:32am Dec 10, 2009 4:32am
Hi Viper

I do all my backtesting and live trading thru TadawulFX in Cyprus. I just download the historical data from the history centre, and run the backtester on max speed (32 on the slider). I get no mismatched chart errors, as long as the code is good. I did backtest another EA this morning, and got lots of mismatched errors.

Cheers
Designmax
  • #1,007
  • Dec 10, 2009 10:58am Dec 10, 2009 10:58am
Quoting tradingviper
Hi Ogi,
... However, if you can come up with a better model that remains profitable, lets explore that. Seriously.

Thanks for all your efforts and contributions,
Viper
OK,
Now that Viper took me up on the challenge I spent the time to give you some more details and hopefully we can make this to be the best EA ever made (I love team work – it diffuses the blame).

 

  1. Look at the table in file “ Compensation logic for Grid EA”. The table is compiled from the testing of both iRegr Grid and ViperGrid. The results in the table are taken from the files “Viper Comp” and “iRegr Comp” (attached). The table summarizes the results from both EA and makes it easy to compare. Also I have my suggestions on the most right of the table. I have tried to make it straight forward to avoid e-mailing back and forth.


The table shows what happens when price hits every level for both Viper and iRegr EA. It shows the levels and lots the EA compensates with for both EA. As you can see iRegr is overcompensating from the very beginning but the lots are more evenly distributed over the levels and the DD is not very big. At the very same time Viper compensates at levels 6 and 7 only, which is OK if someone can guarantee that price will hit the TPS after it hits level 7S. How about if it does not…That is why the big DD and blowing the account.

At the very bottom (row TPS) I have shown the profit the EA will make if price hits the level indicated with X and reverses. I guess it does not get any clearer than that…

 

  1. I keep singing about the entry points and I will keep doing it until someone with programming skills hears me.

In their current condition both EA are rather dangerous to use in automatic mode. I have been using Viper Grid by assigning a pivot price, based on my analysis.Viper Grid (compared to iRegr Grid) is a very solid and robust EA.

Recently I have been turning ON the EA manually on a signal from my manual trading system – instead of entering in a position I just turn on the EA. So far I have not seen more than couple of hundred dollars DD. Keep In mind though that when I turn it on manually I consider all factors I would consider if I enter the normal way.

For my manual entries I am using an indicator I have named StoPSar (short for stochastic/ParSar) The indicator gives a signal when all indicators built in give a signal at the same time. The indicators are Stockastic, ParSar, Accelerator and Awesome Oscillator. For example, let’ take the entry at point 1 on the screenshot: The arrow appears when:

-Fast Stochastic(8/5/3) is UP
-Price has just crossed ParSar
-Accelerator is Blue
-Awesome oscillator is Blue

In addition to that I am taking signals ONLY in the direction of the medium and slow stochastic (on the bottom of the screen) and ONLY if the signal is outside the iRegr Channel AND in the direction of the iRegr Channel. Yes it is true I do not enter in a position often, but when I enter I rarely hit stop loss.

How about we apply this very same criteria to Viper Grid and have the EA enter only in the best of the moments (sounds romantic, doesn’t it)? Coupled with a better compensation logic it should be very safe to use. It may not make as much money on one currency pair but at least you know that whatever it makes is not going to take it back when it blows the account. And with a low DD you can always plug it in on more than one currency pair.

Ogie D

PS. I posted my idea of a Rolling Grid sometime ago. It went unnoticeable, but in my next posting I will post some more details why I think it will not blow the account and will be the best thing after canned beer.

Cheers

Attached Image (click to enlarge)
Click to Enlarge Name: StoPSar entry.jpg Size: 213 KB
Attached File(s)
File Type: doc COMPENSATION LOGIC FOR GRID EA.doc   79 KB | 1,242 downloads
File Type: doc VIPER COMP.doc   263 KB | 1,367 downloads
File Type: doc IREGR COMP.doc   264 KB | 998 downloads
  • #1,008
  • Dec 10, 2009 12:08pm Dec 10, 2009 12:08pm
I guess I must just have bad historical data. I have tried to download new data and it tells me I already have the data. Maybe I should move and rename my current data and try downloading again.
Viper

Quoting designmax
Hi Viper

I do all my backtesting and live trading thru TadawulFX in Cyprus. I just download the historical data from the history centre, and run the backtester on max speed (32 on the slider). I get no mismatched chart errors, as long as the code is good. I did backtest another EA this morning, and got lots of mismatched errors.

Cheers
Designmax
  • #1,009
  • Dec 10, 2009 12:13pm Dec 10, 2009 12:13pm
PS. I clicked on the button too quickly and now just realized that there are more things i have to say:

If the entry strategy suggested is implemented in the ViperGrid Ea there are many parameters to be optimized:
- all parameters of iRegr Channel
- slope of iRegr Channel
- number of bars over which to measure slope
- all parameters of Par Sar
- all parameters of stochastic - fast, medium and slow
- Time frame could be included in the optimization - different currency pairs have different results and different time frames.

I personally never use "true / false" in the EA - always use 0 and 1 because I can just optimize it and this way i know which time frame to use, whether to use some parameter or not.

This is my two cents

Ogie D
  • #1,010
  • Dec 10, 2009 12:14pm Dec 10, 2009 12:14pm
Hi Ogie, If you want to post your special indicator combo thingy, I'm sure I can transfer that into ViperGrid in just a couple hours (hopefully this weekend). I really am trying to get this as safe as possible to be run fully automatically with no babysitting required.

I will look at your data. But consider this with both models. Price bounces from from b1 to s1 to b1 to... 10 times. Then runs up to b3 and reverses. The old compensation model is screwed. ViperGrid is still happy as a clam. Also, the "new" model is how the old one was designed, there is actually just a bug in it...
Viper

Quoting ogidimi
OK,
Now that Viper took me up on the challenge I spent the time to give you some more details and hopefully we can make this to be the best EA ever made (I love team work – it diffuses the blame).

[list][*]Look at the table in file “ Compensation logic for Grid EA”. The table is compiled from the testing of both iRegr Grid and ViperGrid. The results in the table are taken from the files “Viper Comp” and “iRegr Comp” (attached). The table summarizes the results from both EA and makes it easy to compare. Also I have my suggestions on the most...
  • #1,011
  • Dec 10, 2009 1:45pm Dec 10, 2009 1:45pm
Quoting tradingviper
Hi Ogie, If you want to post your special indicator combo thingy, I'm sure I can transfer that into ViperGrid in just a couple hours (hopefully this weekend). I really am trying to get this as safe as possible to be run fully automatically with no babysitting required.

I will look at your data. But consider this with both models. Price bounces from from b1 to s1 to b1 to......
Sorry Viper, I was thinking of attaching it and forgot. The indicator is attached to this message.

As far as the models are conserned - yes, you are right, BUT (a big BUT here):

1. Using the EA with 3 levels only is simply a suicide. I have done thousands of tests and experience shows that less than 5 levels will give you a big DD for couple of reasons:
a - it starts compensating as soon as it hits level 1 or 2
b - number of compensating lots grows very quickly and is very close to the pivot. As we all know price tends to be attracted to the pivot lines
c - if you use it with 7 or more levels it starts compensating from the 3rd level and up. So this scenario with price bouncing between S1 and B1 10 times is of course likely, but it will do no damage if you have more than 5 levels even if your "IncrementAdjustFirstLevel" is = 0

I have never used the EA with less than 7 levels. I have even made 7 levels default in iRegr on my VPS.

2. That is why you have "IncrementAdjustFirstLevel". I usually set this to at least 50% or 100% the increment - it works great.

Viper, you have both ViperGrid and iRegr Grid EA. Do the testing with both EA with different levels and see by yourself. Also compare the DD for ViperGrid and iRegr Grid. I know you will agree that the proposed compensation logic is the the safest one.

The situation where price hits levels 7S(see previous posting) is not going to happen ofter. All it takes is to happen once...

Viper, currently iRegr Grid enters on a different criteria that the one described in my manual strategy - you will see it in the code. But if you do not provide an option for optimizing all those parameters of the indicators, especially the slope of the iRegr it will be a waste of time.

Also, TP is something that will influence the compensation logic too, especially considering "IncrementAdjustFirstLevel". I would suggest TP to be the value set by the user in the filed "EntireSetTakeProfit" or if that value = 0 - the calculated TP if price would go from pivot line to Tp with no compensation

Let me know if you have any questions

Thanks

Ogie D
Attached File(s)
File Type: mq4 StoPSar Alert.mq4   4 KB | 937 downloads
  • #1,012
  • Dec 10, 2009 2:17pm Dec 10, 2009 2:17pm
Hi Guys
Very interesting what you are doing with Viper Grid. I have been running it on a Alpari UK demo account for about 4-5 weeks now and like what I see. I have changed versions as Viper posts new ones and I have run them all on the defaults. I have gone back and forth a few times between EURUSD and GBPUSD a few times, so I haven't been consistant in my testing. I just used a $500 acount at .01 lots and it took the $500 up to $812 as of today. Maybe just lucky so far that I have had no big DD. Looks very promissing, if I was a better programmer I would like to help, but that I am not. So hope you don't mind my borrowing your hard work.

Thanks
Gene
  • #1,013
  • Dec 10, 2009 3:47pm Dec 10, 2009 3:47pm
Viper, one more thing.

if I had to do the Auto Increment I would probably use the width of the I-regr channel instead of ATR and take only about 40% of it(optimizable). Problem with ATR is you never know when you are going to enter in a position. If you enter at the beginning of the day at the day's high or low, then you fine, but what are the chances of that happening...?

Iregr is similar to BB and we all know that the middle line tends to act as a support/resistance. Once you enter outside the channel (or at the channel line) then you know that the only way the price will go is toward the middle.

This is how I have been determining the levels/ increments when i turn it on manually and it has not failed me so far...

Just another food for thought...
Ogie D
  • #1,014
  • Dec 10, 2009 4:27pm Dec 10, 2009 4:27pm
Quoting tradingviper
Hi Desi,

Running a second grid does not prevent a blowup at all. What I said or meant to say is that if a second grid is spaced right, it can't contribute to a blowup. Everytime one grid swings across the the pivot to accumulate more positions, the other grid closes for a profit.

Now I have to ask. How do you get your backtester working so well ?? Very nice. I always have chart mismatch errors and everything else under the sun. Maybe I have poor historical data ?

Viper

TV

Go to 'tools', then 'options', then to the 'chart' tab.

In 'Max bars in history' and 'Max bars in chart', enter 99999999999999 in both.

It will default to: 2147483647

Then try and re-download your 1 min data from the history centre.

That should help.

OT
Luck is where preparation meets opportunity.......
  • #1,015
  • Dec 10, 2009 8:15pm Dec 10, 2009 8:15pm
Quoting optojay
Haven't heard from Plasmapants in awhile...he must be off making millions somewhere with floating multigrids!
Here when ever i need a break from my playboy life style. Just havn't had anything useful to add. Kinda happy with they way vgrid is as it stands. Running on gu with an 80-90 pip take profit target.

I read somewhere, i think its art of war stuff that 'if you know yourself and you know your enemy, you need not fear the result of a hundred battles'. Well i know how much equity i have, and i know my enemy the unenviable 'Death trade' - and the result of the battle will be decided when gu refuses to go that distance from the daily open.

Now about entry filters..... and i reserve my right to be a cynic even after what i say next....

It seems the purpose of entry filters is to find a price point from where price will not range but tend to move off in one direction. I've been lurking on a few threads here at ff and a few strategy's have caught my attention which may be useful entry points.

The first one is ichimoku kinyou hyoo heeya waya where the kangen sen and tangen sen cross over above the cloud but underneath the kinky sen. (if your not familiar with itchy check out the thread, i dont have the imagination to make this stuff up.)

The second is a simple adx entry; its just a range trend indi.

And lastly, if the ea initiates on a candle bunched up zone breakout.
bull in a china shop
  • #1,016
  • Dec 10, 2009 8:54pm Dec 10, 2009 8:54pm
Quoting ogidimi
Using the EA with 3 levels only is simply a suicide. I have done thousands of tests and experience shows that less than 5 levels will give you a big DD for couple of reasons:

I've only ever used 3 levels, had a few close calls but never a death trade or dd resulting from 3 levels.

Suicide is using a pip increment which is either too small so that you get stuck inside 4 soldiers standing side by side on an hourly chart or too great so you are going to be sitting around a long time waiting for swings to complete so your tp can be hit.

I understand what your saying about maxing orders at level 5, but if you've got a 15 pip increment with 5 levels, and i'm using 23 pip increment with 3 levels, your level 5 is at 75 pips from pivot and my level 3 is at 69 pips from pivot, it hardly a significant difference. With enough swings the inevitable is equally likely for both setups.
bull in a china shop
  • #1,017
  • Dec 11, 2009 4:20am Dec 11, 2009 4:20am
Quoting plasmapants
I've only ever used 3 levels, had a few close calls but never a death trade or dd resulting from 3 levels.
.

There is a first time for everything
  • #1,018
  • Dec 15, 2009 6:12am Dec 15, 2009 6:12am
Hi ForexFlash,

It's a long time since you last posted here. If you are still lurking around can you post details on your update with your EA?
What are you doing with it, do you plan to sell it, do you still work on it, etc ?

Bests.
  • #1,019
  • Dec 17, 2009 4:47am Dec 17, 2009 4:47am
Hello to all

First I want to thank you for developing the new ea.

I have tried to backtest the viper ea but no matter what setting I am using. The account will be killed...

Please help me with the right set file - and the ea you are using.

Thank you

Regards
  • #1,020
  • Dec 17, 2009 11:50am Dec 17, 2009 11:50am
Thanks Ozzie,
I'll check out what it's set to now and try your suggested approach.
Where's Forex Saint ? He was very gungho and then disappeared..
Viper
Quoting Ozzie Taz
TV

Go to 'tools', then 'options', then to the 'chart' tab.

In 'Max bars in history' and 'Max bars in chart', enter 99999999999999 in both.

It will default to: 2147483647

Then try and re-download your 1 min data from the history centre.

That should help.

OT
2 traders viewing now