100% Win Math Grid Ea | Page 58 | 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

  • #1,141
  • May 29, 2010 11:13pm May 29, 2010 11:13pm
Dear TV, FF and others,

I'm done reading through the complete thread. It's been a pleasure to see that past members reappeared over time back and then. This shows ppl interest.

My thinking up to now:
- FF you started a great thread
- TV you provided non-coders here a "professional support and coding"
- There are way too many times the same questions on "I can't compile or attached the EA .... OrderReliable .... blabla" .... come on guys ... the only thing you have to do is "read" ... great ppl are doing the rest for you.

That said:
- The present code does not handle non-uniform grid but that is fine considering it does not allow to do so. Specifically, adding the "extra-increment" beyond the original TP would directly impact lot-size calculation .... at every level ... or it better be extra-pips you wanna suck in.
- I'm 100% on FF side when he said there were potential on "intermediate exit strategies"
Quote
you should work out the math to solve this problem, and that is something like 3 main exit strategies with each having 6 substrategies for loss mitigation.....so, total of 18 pure math exit strategies!
, this is exactly what I meant when saying
Quote
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.
- up-to-now, (using ViperEA .. and yes the latest version with backtesting improvements), and my own coding (with my own lot calculation), I'm not quite sure of the "starting from the outer band inward" to calculate lot size .... this does "over-weight" the last level ... however I do understand that it is statistically the most unlikely level for a reverse. Nevertheless, you'd better check ahead that this last level is not a support/resistance, pivot and else once you think of it ... obvious no
- I think what would work best is some kind of precalculated tables as we do in "distance transform calculation" to speed up the propagation of "a-priori known distances".... I have to find what entries would be best for the table though (level, increment, equity, nb of orders ....) and how to automatically "feed" the table to generate every cases with associated risk.

I'll keep you up-to-date with these concepts.
Great work.

XmPh.
  • #1,142
  • May 30, 2010 7:28am May 30, 2010 7:28am
Quoting xmph
Dear TV, FF and others,

Nevertheless, you'd better check ahead that this last level is not a support/resistance, pivot and else once you think of it ... obvious no

XmPh.
I like where you're going with this.. . Could you explain the logic behind your thinking more? I'd like to hear what kind of ideas you have for dealing with the concept.
  • #1,143
  • May 31, 2010 6:10pm May 31, 2010 6:10pm
Quoting luiboy
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...
Can you please post your set file?

Thanks very much.
福 子 Victor V
  • #1,144
  • Jun 1, 2010 1:30am Jun 1, 2010 1:30am
Hi camvcvoo,

sorry but what is a set file?
  • #1,145
  • Jun 1, 2010 1:51am Jun 1, 2010 1:51am
I've been back testing and forward testing this EA for the past almost month right now. and i have been getting quiet good results.

I have to thank you and FF for both your efforts and all the memebers here. I have read almost each and every line of this forum and almost learn the code by heart trying to evaluate and the number of sets that i have for back testing and implementing on forward testing ad getting out of control.

First of i totally agree that if we find better entries and make them very accurate we wouldn't need to rely much on low and high watermark strategies.

Now this is my request, and please take this into consideration.

Many times, we have seen that the EA would for example

Buy first level, and hit the second level & 3rd level but will not reach the TP.

Or even hit the first level or 2nd level, and wont reach the 3rd level and then start going down to reach first and 2nd level of sells, and would just keep lingering between these 2 levels for hours and days, and all we need to start thinking about right now is how to exit this strategy with the minimum losses.

Now if in the case of the trading started and hit B2 and above, move your stop losses to B2, and if it hits B3, move the SL to B3, if people who have more levels such as 5 or 7, that i've read before in the forum. why don't we move the SL to that level.

I know we are trying to take the maximum of the market, but increasing the ratio of being in the market with the Martingale system, multiplies our odds of being stuck in a ranging market.

Please take a look into this consideration, I really want to test this.

I have been running 2 forward tests for the past almost 2-3 weeks now,

first test = Aggressive with 70% up now and 30% DD $25k capital 0.1 lot
second test = Conservative test $100k capital with 0.1 lot (44% up with DD 13%)

now the problem with both the EAs, that i would have been in much better situation if i have the option to bank my profits, all of my positions right now have hit B3, but reversed back to the pivot / initiation point. and this is what caused the previous draw downs.

I'll be glad if you can look into my request


In addition, I want to see if we all can do an optimization for the entry conditions

1 - Trend
2- BB Filter
3 - I-Regr

can we all run the same pair, same period same broker, and try to optimize the settings and see what could be the best for as many pairs as we could.

I have been into this business for many years now, and i always beleived in formulated based strategies. Have been across many, but i feel this has the potential to go somewhere, with all the brains in here. Im sure we can do something or even this might be the opening for something else.

Thanks,

Best Regards,
  • #1,146
  • Jun 1, 2010 8:12am Jun 1, 2010 8:12am
Quoting FrozenAir
I like where you're going with this.. . Could you explain the logic behind your thinking more?
There are 3 points in my line of thought so far:
1. Determine "Incomplete Exits"
2. Use the grid wisely, that is to see what the future might be.
3. Determine Better Entry.

I'm investigating these in this order of relevancy.

1) is basically similar to what Wael117 mentioned earlier (see quote). The most simple "incomplete Exit" is, (from my point of view, slightly different than what Wael117 presented): if you hit three level in a row, start trailling from in between second and third level, of increment the price action happening as third level is hit. From the calculation I did, this should make sure profit is always achieve. Not true if you move your stoploss to second level. Currently, I'm working on finding the more complex ones.

2) see what the future might be: what does this mean: We don't predict the future, however, we know what we should do if price go down two levels and up three down (sapiens sapiens right .... you don't throw the stick cause you know it might be useful tomorrow) .... so you can think ahead to see what would cause a large DD or a margin call .... you can even determine whether or not you still have lots of chance to end up in profit ... if very little .... you'd better plan your way out right now and make sure you limit your loss .... "thinking of a Basket Strategy Alarm".

As well, why would you keep "surcharging" level 3 if after three days this is a support/resistance .... if that's the case, eventually it would be better planning an "incomplete exit". To me ... this line of thought is the next step of thinking, the step after the ones presented by FF. It is also more complex. (so this one is number 4) on my list)

3) better entries .... well .... no need for explanation. FF already gave us some clues .... I manage to have 18mths backtested without Margin Call on an account starting at 2500 and finishing above 8k simply by running the strategy at the Frankfurt opening and allowing only one set.

Quoting wael117
Buy first level, and hit the second level & 3rd level but will not reach the TP. [...] Now if in the case of the trading started and hit B2 and above, move your stop losses to B2, and if it hits B3, move the SL to B3, if people who have more levels such as 5 or 7, that i've read before in the forum. why don't we move the SL to that level.
Quoting wael117
In addition, I want to see if we all can do an optimization for the entry conditions
I would say this is not useful, at least at this point in time. Optimization at this time would only "if lucky covers up points 1) and 2) ". Additionally, the point of the strategy is to not require optimization as it handle naturally the price action.

That's it for me. The rest is .. my trying to figure out 1) ;-)
Take Care.
  • #1,147
  • Edited 9:24am Jun 1, 2010 8:50am | Edited 9:24am
This is what i meant with this for further clarification.

Please be this is only in the case where the market moves in 1 direction

as in B1,B2,B3 or S1,S2,S3

in which if market moves B1/S1

Do nothing

When reaches B2/S2 move SL to B1/S1

When reaches B3/S3 move the SL to B2/S2

I also like the idea, if after a certain level we can start trailing profits.

For example if after B2 is crossed we can start trailing with the size of the Increment, until it reaches the TP level.


I have been in trades on my forward test since Friday, I'm trading 7 pairs. on my standard settings. All the 7 pairs reached between my B3/S3 and TP and moved back on market opening on Monday back to pivot /initial point.

Now im done with all of them banked in profits, but if the market reversed further and came back to the B3 as was on Friday. That would be a huge margin and possibility of Margin Call.

I would like to take the easy way out. IM not even using large increments, my whole setup runs on 15 pips increments. Therefore i would prefer to make sure i get my 30 pips and 15 pips from the first 2 levels than going in all this deficit especially im being very considerate running with 0.1 % loss.

Can someone please explain if they are using % of ATR increment. I did some back tests with amazing results.

And which pair is everyone using? and found more comfortable with?


In addition with optimization, doesn't have to be back testing, but forward testing. I am very interested in the ATR increment, I do not believe much in the back testing, i agree with many people who think it is very poorly setup. Many of the back tests that i ran all blew out. and thats only 1 pair. but when i run a forward test, im doing amazingly good with much higher risk and lower equity size.

Well it was an idea, maybe the combined effort of a couple of people can present something that we cannot. in the end of the day many people based results on back tests. although i tested back tests and forward tests do not match at all, even with 90% quality modeling.

Thanks everyone
  • #1,148
  • Jun 1, 2010 12:24pm Jun 1, 2010 12:24pm
Quoting wael117
I would like to take the easy way out. IM not even using large increments, my whole setup runs on 15 pips increments. Therefore i would prefer to make sure i get my 30 pips and 15 pips from the first 2 levels than going in all this deficit especially im being very considerate running with 0.1 % loss.
From my point of view, the sooner you want to bank profit, the less chance you'll have for this strategy to work. Basically, if you bank profit after Level 1 ..... what's the point of using a grid ?

Again, from my point of view .... banking at Level 2 ... or moving a SL at level2 would defeat the fundamental idea of this strategy .... and it defeats the math as well ....

However, placing a trailing mid-Level2/3 once price hit Level3 does not defeat the math for a single 1->2->3 and maintains most of the logic for any reversal prior reaching a straight 1->2->3 which still preserve the fundamental interest in this grid strategy ... and hopefully the design of any other "IE" (incomplete exit)

I believe that securing too much is just departing from the proposed strategy, and from its interest (that is handling crazy reversal).

X.
  • #1,149
  • Jun 1, 2010 2:07pm Jun 1, 2010 2:07pm
Thanks X for your reply,

Although i want to keep the room for the swings as this would be in the scenario of a 3 level incremental, the trailing would start once the price reaches the 3rd level. with the I value,

Still you are giving the chance of Level 1 , and Level 2 to hit and reverse, in which in Vgrid, the increasing in lot size wouldn't be much, and exactly in reverse S1 & S2 will be both open with minimal lot sizes, the definite increase would be on the S3.

Also think about 5 level incremental. at that time i only want the triggering of the Trailing to start on reaching level 5. I have experienced this so many times. prices is almost there just 2-5 pips away, and starts a big whipsaw for the many times i have seen this in the back test and for the many times i've seen this in forward test, i think this time would be used for better purpose than just sitting there hoping that the market makes the 2 pips to close your low margin positions with big profits, than waiting for your account to start hitting more and more positions and start thinking about low level and high level mitigation.

I am not saying that I FOUND IT!!!.. or not making claims that the EA doesn't work.. i think it works.. and really does if u know how to tame it.. but just trying to reduce the risk levels for the account and the worry levels for me sitting trying to think the best manual exit strategy...

it wont harm if we add this feature to test it, and if it doesn't work.. we give another shot to something else...

Any one else please give us your thoughts..

Thanks

W
  • #1,150
  • Jun 2, 2010 8:08am Jun 2, 2010 8:08am
Quoting xmph
- I think what would work best is some kind of precalculated tables as we do in "distance transform calculation" to speed up the propagation of "a-priori known distances".... I have to find what entries would be best for the table though (level, increment, equity, nb of orders ....) and how to automatically "feed" the table to generate every cases with associated risk.
XmPh.
As anyone any thoughts on what kind of "Table" we may use? What type of input/output?

There is a couple possibilities and I'm struggling to find the best approach so far:
- input: DD, Nb position opens, Current Level .... I also thought of Frequency price reverse at Level X since grid-set started ...
- output: Level, lot size ... but how far ahead ... only for next shot, five next shot ... how many scenarios in advance .... might be based on input=Frequency of Reversal at different level.

To me these are important notions and I did not find what would best suit our needs. I would appreciate some insights.

This is pushing way ahead of simple "Incomplete Scenarios" but I believe this might be also what FF did without clearly stating it.

Best of all,
  • #1,151
  • Jun 2, 2010 8:12am Jun 2, 2010 8:12am
Quoting xmph
As anyone any thoughts on what kind of "Table" we may use? What type of input/output?

There is a couple possibilities and I'm struggling to find the best approach so far:
- input: DD, Nb position opens, Current Level .... I also thought of Frequency price reverse at Level X since grid-set started ...
- output: Level, lot size ... but how far ahead ... only for next shot, five next shot ... how many scenarios in advance .... might be based on input=Frequency of Reversal at different level.

To me these are important notions and I did not find...
I think this is a great idea. Lead and tell me what you want me to do.. i can try to help as much as i could. now im running and monitoring a couple of esettings with different risk and increment levels and according to the results that i have might seem to have lots of potential.

now the thing is that, again today i was caught in almost a a death trade, and again the market was 4 pips away from closing and the reversed and hit S1, S2, then B1, then S1, B1 and S1,S2,S3 and exit. on profit, but all this panic. couldh ave been easily avoided. with the trailing idea.

If someone knows any coding can please help..

Or Viper.. please I'm interested in this function.

Thanks,
  • #1,152
  • Jun 2, 2010 8:32am Jun 2, 2010 8:32am
This is another example of a trade i am in at the moment,

IF there was a triggered T/S at market open S3, and moved up, the order would be closed below S3 with -2/-3 pips

Or just break even everything at S3

and have the function to limit your Trailing at whatever price you want.

Trailing the stop after S3 at 5 pips or 10 pips would have exit all this in profit.

Save the logic

and if this was closed at that price the set after that would have exit in a straight profit B1,B2 & B3 as well.

Thanks
Attached Image (click to enlarge)
Click to Enlarge Name: 1.jpg Size: 381 KB
  • #1,153
  • Jun 2, 2010 9:19am Jun 2, 2010 9:19am
today eur/usd

8 pips from S3
Attached Image (click to enlarge)
Click to Enlarge Name: 2.jpg Size: 145 KB
  • #1,154
  • Jun 2, 2010 5:38pm Jun 2, 2010 5:38pm
I had a death trade yesterday. When my trade reached the high water mark, as expected, the EA deleted all the pending orders. Thus, there were more short positions than long. Then the price go up straight to the sl/tp. If the long pending orders are still there, I should end up in profit.



Gross Profit:14 487.53Gross Loss:26 613.92Total Net Profit:-12 126.39
Profit Factor:0.54Expected Payoff:-38.37
Absolute Drawdown:12 126.39Maximal Drawdown:15 520.90 (60.19%)Relative Drawdown:60.19% (15 520.90)

Total Trades:316Short Positions (won %):180 (42.78%)Long Positions (won %):136 (85.29%)
Profit Trades (% of total):193 (61.08%)Loss trades (% of total):123 (38.92%)
Largestprofit trade:433.60loss trade:-591.24
Averageprofit trade:75.06loss trade:-216.37
Maximumconsecutive wins ($):21 (1 895.45)consecutive losses ($):62 (-14 568.45)
Maximalconsecutive profit (count):1 895.45 (21)consecutive loss (count):-14 568.45 (62)
Averageconsecutive wins:4consecutive losses:2



This is just a cent account. All prices are divided by 10.
Attached Image (click to enlarge)
Click to Enlarge Name: DetailedStatement.gif Size: 5 KB
  • #1,155
  • Jun 3, 2010 12:41am Jun 3, 2010 12:41am
Quoting wael117
today eur/usd

8 pips from S3
This trade ended up in a death trade, wasn't there online to manage, and later i found out that i did hit a MC and some orders didn't open.

Wiped out 40% of my profits on this test.

All other test accounts yesterday were on very high margin. Thought about TP all of them, but isn't the optimum solution for all of these trades.

Plus TP is only in $ value and not in pips, cannot work it out with %risk.
  • #1,156
  • Jun 5, 2010 1:01am Jun 5, 2010 1:01am
Hi FrozenAir,

How's your eurjpy pair last week? Did you had a death trade? I only trade eurjpy in my live account and encountered death trade last week and lose almost 50% of my account. My demo which is trading 10 currency pairs was also hit by eurjpy death trade but overall is still ended profitable because of the other 9 currency pairs.

Quoting FrozenAir
I am trading live,...
Attached Image (click to enlarge)
Click to Enlarge Name: DetailedStatement.gif Size: 6 KB
  • #1,157
  • Jun 7, 2010 9:38am Jun 7, 2010 9:38am
Hello VT,

I have been asking for your help for the break even or a trailing stop feature to be activated after the trade reaches S3/B3.

Another example of why this model would succeed much better than the current one we are working with attached here.

Trade would have exit at B3 with profit and opened another one which would have closed again at TP or worst case S3.

Attached is the photo

Chart is from today's Forward test
Attached Image
  • #1,158
  • Jun 8, 2010 4:27pm Jun 8, 2010 4:27pm
Quoting wael117
Hello VT,

I have been asking for your help for the break even or a trailing stop feature to be activated after the trade reaches S3/B3.
Dear all,
I did use the trailing stop function as I stated previously (so a bit different than Wael).

The results in backtest are good ... and it allows to more than double the gains on some sets (explicitly, realizing more gains and eventually adding the gains of two sets but for the cost of a one-set spread), and prevent un-required exposure in other cases.

However, considering the Trailing Stop is updated on a tick-basis, how do you think this would behave in a Live account that inherently has some lag ? Did you ever program or discuss handling trailing multiple orders on a potentially tick-basis with lag in a reliable way ?

I have multiple options, and no matter what you have the guarantee that the first trailing point will maintain a win situation, which is a good start, but I have the feeling that an "OrderReliableTrailingStop" would be a little more complex to do ?

regards,
  • #1,159
  • Jun 9, 2010 2:22am Jun 9, 2010 2:22am
Quoting xmph
Dear all,
I did use the trailing stop function as I stated previously (so a bit different than Wael).

The results in backtest are good ... and it allows to more than double the gains on some sets (explicitly, realizing more gains and eventually adding the gains of two sets but for the cost of a one-set spread), and prevent un-required exposure in other cases.

However, considering the Trailing Stop is updated on a tick-basis, how do you think this would behave in a Live account that inherently has some lag ? Did you ever program or discuss handling...
Hello XMPH,

Thanks for your feedback mate,

Now I have been running more than 5 forward test, results were very good, although in each of every case, i hit a death trade, now the times when i was able to monitor this manually, and exit, i had a similar way to do this as a trailing stop after B3/S3 and i was really successful in getting all of my trades out before reversing to the other set of buys/sells.

and managed to get the next cycle open and get an advantage of the reversal.

But when i haven't been able to monitor them, i got stuck in extremely difficult positions which resulted in MC the account.

Its a safety precaution which really works well. I kind of dont believe much in the Back testing results, back testing results just to show how will the EA behave, and if adding the trailing stop function will show pretty good behavior, but on the live forward test i think it would be very accurate and will save a lot of pain and stress thinking about the best strategy to exit your positions.

I have tried as much combination of low and high exit, and cannot seem to find one, although working on trialing stop manually really helps.

Automating the feature will be really a blessing to my strategy and will give it much higher accuracy. We just need to find someone to code this i hope viper trading can help us in this.
  • #1,160
  • Jun 9, 2010 5:07pm Jun 9, 2010 5:07pm
Quoting wael117
Hello XMPH,
We just need to find someone to code this i hope viper trading can help us in this.
See my code below (implementing my vision of the trailing as specify in the few lines of code comments). I do not know where this would go in VT code considering I recoded everything for myself. However, I believe my variables are self-explanatory.

Now Trailling is one thing if you are seeking a quick exit after a Straight price action from Level 1 to 3 .... Is this STILL profitable if that happens after multiple swings ?

Answering this allows to start investigating the different incomplete exit strategies. I would appreciate some ppl to provide us with their thought about this.

If no one is investigating these incomplete exits ... then my understanding is that we are not on the right track as a group.

regards,

Inserted Code
void TrailOrder()
{
   double PriceMovement = 0;
   double InitialSL = 0;
   int i=0;
   
   int cnt,total;
   total  = OrdersTotal();
   
   
   
   if(TrailingStop)
   {
/*
   THIS LOOP WILL HELP UPDATING THE STOPlOSS IN ORDER TO TRAIL FROM MID-LEVEL 2, BASED ON THE MOVEMENTS SINCE LEVEL 3
*/
      for(i=total-1;i>=0;i--)
      {
         OrderSelect(i, SELECT_BY_POS);
         if (OrderType() == OP_BUY)
         {
/*
   We check if there is anything to Trail by comparing the point movement from Level 3 with current SL.
   
*/
            PriceMovement = thisAsk()-GridLevelSetUp[1][LevelBuy3];
            InitialSL = GridLevelSetUp[1][LevelSell2] + (GridLevelSetUp[0][GridSpace]/2);
            
            if (InitialSL +  PriceMovement > OrderStopLoss())
            {
               OrderModify(OrderTicket(),OrderOpenPrice(),InitialSL +  PriceMovement,0,0,Green);
            }
         }
         else if (OrderType() == OP_SELL)
         {            
            PriceMovement = GridLevelSetUp[2][LevelSell3]-thisBid();
            InitialSL = GridLevelSetUp[2][LevelSell2] - (GridLevelSetUp[0][GridSpace]/2);
            
            if ( InitialSL -  PriceMovement < OrderStopLoss())
            {
               OrderModify(OrderTicket(),OrderOpenPrice(),InitialSL -  PriceMovement,0,0,Green);
            }
         }
      }
   }
}
0 traders viewing now