100% Win Math Grid Ea | Page 5 | 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

  • #81
  • Edited 1:00pm Oct 5, 2008 12:49pm | Edited 1:00pm
Hanover thank you for that...but Im not a patient man

Today I started to code addition to ea....I added a S\R to it.....
Now ea trade only on suport, and exit from all trades when price hits resistance.

Im also working on another version, but without martingale element!
Ea will trade only buystops levels on suport, and sellstops on resistance.
Each level will have a TP on beggining of next level, and same SL.
It will execute like 10 trades in one direction, but with
small increment...something like 10 pips, picking all the pips in that direction,
and then when S\R gets hit, all trades will be closed, and then ea opens
a new set in direction of price.
This could be a good "scalping" tool with really small losses when price reverses.....but that is only a theory for now.

I will post 2 new versions here in a few days.
forexflash
  • #82
  • Oct 5, 2008 3:21pm Oct 5, 2008 3:21pm
I've been following this thread since it started. I started testing the original EA I think on 9/29 and had really good results until friday. It ended the day down $4100 for the day on a $6200 acct balance, but it was trading mini lots and not microlots like i would have been using if it was a live acct. I'm impressed so far and appreciate all the input everyone is giving. i almost have a hard time keeping up because my computer/forex study time is limited as of late. So keep up the good work and thank you very much.
Is it Break Time yet?
  • #83
  • Oct 5, 2008 7:11pm Oct 5, 2008 7:11pm
i was testing the ea and noticed that when 1st level was hit in the setup which was a sell didnt add a lot to the other sell orders pending but the buy levels looked like they were accurate
  • #84
  • Oct 6, 2008 1:14pm Oct 6, 2008 1:14pm
See inline
Quoting hanover
The ideas that I’d like to implement are –

1. See post #51. I believe that the order generation algorithm can be made more efficient. As orders at each price level are triggered, only enough buystop/sellstop orders need to be placed to bring the net result to +1, when either TP/SL is subsequently reached. It might also be possible to re-design the concept so that, instead of adding pending orders in one direction, we simply delete previously created pending orders in the other direction. But I need to give this a lot of thought.

David, as I understand the code, there only one side of the pending orders used in calculations for each side. Thus decreasing the count of opposite side pending orders won't do the trick.

2. Use TA-based concepts to have the EA enable itself when volatility appears to be increasing (along the lines of those I mentioned back in post #36, and/or like El Salon’s in post #78), and then disable itself until these conditions re-appear. I will code in such a way so that the “cascading orders grid” becomes an MM plug-in for any entry method.

Complete switching on and off the grid will add unnecessary problems due to errors in prediction of volatility explosion. I think changing/adjusting the grid increment according to the current situation would be more valuable solution.


David
  • #85
  • Oct 6, 2008 1:41pm Oct 6, 2008 1:41pm
Here is a recoded version with implemented S\R.....
now ea executes grid only on suport and resistance levels.
Use 30min timeframe, EU seems to be the best pair!
Pisara thanks for the tips.....maybe some day.

Respect
Attached File(s)
File Type: mq4 SR_mGRID_EA.mq4   13 KB | 7,311 downloads
forexflash
  • #86
  • Edited 10:39pm Oct 6, 2008 3:28pm | Edited 10:39pm
I'm really high on this EA because I believe it has the ability to be profitable, So I've started some very basic backtesting results:

9/1/08-9/30/08:
0.01 Lots
$500 account balance

EURUSD
M5 +$700.73
M15 +$670.70
M30 +$671.70

EURGBP
M5 +$34.84
M15 +$34.84
M30 +$34.84

EURCHF
M5 +$138.37
M15 -$132.37
M30 -$132.23

EURJPY
M5 -$74.19
M15 -$209.70
M30 -$166.31


GBPUSD
M5 +$123.51 - Drawdown of 51.46%
M15 +$134.19
M30 +$148.81
  • #87
  • Edited 7:01am Oct 7, 2008 4:47am | Edited 7:01am
Hi guys,

I have been backtesting this system since last night. But the only problem is that I don't have enough historical tick data. I used the tick data 90% from this website. But that is only about 3 months. Anyone got more historical tick data for metatrader?

Maarten
  • #88
  • Oct 7, 2008 5:05am Oct 7, 2008 5:05am
Quoting Maarten999
I used the tick data 90% from this website. But that is only about 3 months. Anyone got more historical tick data for metatrader?

Maarten
I have used Alpari 1min data recently. Those large zip files contain several years of data. If you only see 3 months worth of data I doubt any of it has been imported from the zipped Alpari *.hst file.

Make sure the broker you are using is compatible with Alpari. No non-standard currency name extensions. Also check your import procedure. When you click import you should see what dates are in the *.hst file.
  • #89
  • Edited 9:36am Oct 7, 2008 5:31am | Edited 9:36am
So guys. Here is a little update. Been today busy with backtesting this EA on all timeframes with different kind of settings.

Now the problem: in the short and mid term this EA fails. Sometimes it very profitable for days, weeks or sometimes even months but in the end, it always blows itself up. I have used all diffirent kind of increment level and even the stop loss -100, -250 etc. But in the long term it doesn't work.

The good news: in the long term (with increment 100 pips level) this EA can be profitable (see chart).

Attached Image (click to enlarge)
Click to Enlarge Name: pic2.GIF Size: 46 KB


(the backtesting was with tick data from alpari)

The only problem is, is that there are very weird temporarily drawdowns.

Now that I am thinking about this EA, I must say that I really like this system. Because when the long term trend changes (like it did lately @ 8-08-2008), this EA is still profitable.

Maarten
  • #90
  • Edited 12:04pm Oct 7, 2008 11:54am | Edited 12:04pm
Quoting Maarten999
Sometimes it very profitable for days, weeks or sometimes even months but in the end, it always blows itself up. I have used all diffirent kind of increment level and even the stop loss -100, -250 etc. But in the long term it doesn't work.
Maarten, you're correct. Given enough time, the EA will always blow itself up. That is due to the way that the MM works (which is explained in the earlier posts in this thread). The equity curve that you posted will eventually blow up. It might happen later in 2008, or perhaps some time in 2009, or 2010, or maybe not until 2025. But, as soon as price zigzags sideways for a long enough period (relative to the grid parameters), it will blow up.

The best that can be done is to reduce the probability of a blow up, say from 1% to 0.1% to 0.01% etc, by doing any or all of:
-- optimizing the MM (grid) parameters as much as possible
-- making the order placement algorithm more efficient (see my earlier posts)
-- using market analysis to time the entries when a large, clean price movement is most likely to occur

The "weird drawdowns" that you describe are also caused by the way the MM works, but they are always immediately followed by a new equity high. That's because the MM ensures a 100% win rate across every block of trades....... until a situation occurs where the account margin is exceeded, causing it to blow up. It has nothing to do with the trend changing. It doesn't matter whether the market is trending up, or trending down, or even moving sideways (depending on the grid parameters), the account equity will keep on rising....... until the EA blows itself up.

David
  • #91
  • Oct 7, 2008 12:02pm Oct 7, 2008 12:02pm
Quoting pisara
See inline
Quote
David, as I understand the code, there only one side of the pending orders used in calculations for each side. Thus decreasing the count of opposite side pending orders won't do the trick.
I was talking about re-designing the algorithm, but you're probably correct. Like I said, it would need a lot of thought.

Quote
Complete switching on and off the grid will add unnecessary problems due to errors in prediction of volatility explosion. I think changing/adjusting the grid increment according to the current situation would be more valuable solution.
You might be correct. Depends on whether it's easier to forecast an impending increase in volatility, or a future price target (to adjust the grid TP/SL target accordingly).

David
  • #92
  • Edited 1:37pm Oct 7, 2008 1:16pm | Edited 1:37pm
Quoting hanover
I was talking about re-designing the algorithm, but you're probably correct. Like I said, it would need a lot of thought.
David, maybe injection of tdion's Blackbeard ideas might somehow help?

Quote
You might be correct. Depends on whether it's easier to forecast an impending increase in volatility, or a future price target (to adjust the grid TP/SL target accordingly).
Not only that, but we can also simply reduce the block's lot size depending on hourly statistics of the pair!

Quoting hanover
But, as soon as price zigzags sideways for a long enough period (relative to the grid parameters), it will blow up.
...
the account equity will keep on rising....... until the EA blows itself up.

David
I'd like to politely disagree on that one... Imagine we manage to implement gradually switching to taking partial TP after specified margin reservation %, and, as it will do in most out of luck situations, closing with controlled/limited loss like FF already tried to implement (although IMO it should be strictly the current block equity stop limit).

MM for this idea should be quite interesting, as I see the bunch of possibilities to implement rather original approaches with use of margin used, lifetime of the block, current volatility and flexible time-equity based SL/TP. Also, lot size (and an individual block's calculated risk %) might be varying quite dramatically, from say 0.01 to 0.2 lots.
  • #93
  • Oct 7, 2008 2:21pm Oct 7, 2008 2:21pm
Quoting pisara
I'd like to politely disagree on that one... Imagine we manage to implement gradually switching to taking partial TP after specified margin reservation %, and, as it will do in most out of luck situations, closing with controlled/limited loss like FF already tried to implement (although IMO it should be strictly the current block equity stop limit).
Pisara, you may be right. I guess that ultimately the only way to find out would be to code and test your idea.

But (without thinking through the possibilities in too much detail) here is my main cause for doubt: Costs make trading a negative expectancy game. Gaming theory postulates that it's impossible to beat a negative expectancy game by using only MM. (For example, there's no MM system that will beat the casino, long term, at Roulette).

If you do manage to succeed with this (the implications, if I understand correctly, are a 100% win rate with zero risk of a blow-up), then market analysis would become completely unnecessary. I expect that the big fund managers would happily pay millions for an EA that would allow them to make all of their analysts redundant, and simply (if MrGreen will allow me to use his turn of phrase) "click and grow rich". If everybody was to run such an EA, then I can only wonder what effect it would have on the forex market (given that, in order to function, all speculative markets ultimately depend upon a composition of willing buyers and willing sellers).

David
  • #94
  • Edited 3:07pm Oct 7, 2008 2:51pm | Edited 3:07pm
Keep on brainstorming guys, I love it!
Maybe this EA need to be coded in a way that after X cycles the ea either makes it "0" zero out once the partial profit target is hit or you actually know which buy and sell orders to delete at $$ so that the EA can keep ranging and freeing up your margin so that you can keep adding positions?
forexflash
  • #95
  • Oct 7, 2008 3:12pm Oct 7, 2008 3:12pm
Let's assume that it's possible, by using ideas similar to those described in posts #90 and #91 (and possibly others, also) to reduce the risk of blow-up to something tiny, e.g. 0.0001% per block of trades. So a blow-up might occur, on average, only once in (say) 80 years. If I was 'lucky' enough, I could make several million, and then quit before the blow-up.

Given that potentially anything can happen (even with 'conventional' systems whose TA provides some kind of edge, e.g. a sequence where losses exceed wins by 50, at 2% risk per trade, is going to seriously damage an account), then an EA offering a 100% win rate with only 0.0001% risk of blow-up might arguably be 'competitively' attractive.

Who knows, perhaps fund managers like Jim Simons have an EA like this somewhere among their arsenal of trading methods. LOL.

David
  • #96
  • Oct 7, 2008 3:33pm Oct 7, 2008 3:33pm
Quoting hanover
If you do manage to succeed with this (the implications, if I understand correctly, are a 100% win rate with zero risk of a blow-up), then market analysis would become completely unnecessary.

David
You are right, single MM alone will not do the trick due to the math and even a common sense! And of course not 100% win rate. As I wrote, we'll have the fuse to close 'hard luck' blocks with strictly limited (comparing to margin call) loss. Sure, such situations will happen. Such 80-95% win grid (can't say exactly until implemented & backtested) will still benefit from a prediction or an other edge. The beauty of the idea is rather small but steady comparable to conventional TS ROI. The price is still the same - the risk of quick downhill of the account equity by series of subsequent losing blocks due to prolonged sleeping/ranging market. The thing is we still need either TA, statistics or some other advantage to be able to limit the frequency of occurrence the blocks with margin deficit. But we don't have to determine the direction of the trend (although it still will make the difference - and you might guess the next improvement idea behind that already!), so that unipolarity is one of the key features/advantages of the system.
  • #97
  • Oct 7, 2008 3:42pm Oct 7, 2008 3:42pm
Quoting hanover
Let's assume that it's possible, by using ideas similar to those described in posts #90 and #91 (and possibly others, also) to reduce the risk of blow-up to something tiny, e.g. 0.0001% per block of trades. So a blow-up might occur, on average, only once in (say) 80 years. If I was 'lucky' enough, I could make several million, and then quit before the blow-up... LOL.

David
More seriously, hardly one willing to wait the account blowout 80 years...
Personally, I would be happy if backtests will show bad luck every 2-3 years, provided we still manage to achieve reasonable ROI.
  • #98
  • Edited 4:43pm Oct 7, 2008 3:46pm | Edited 4:43pm
EDIT: Stupid me!
  • #99
  • Oct 7, 2008 4:22pm Oct 7, 2008 4:22pm
Quoting FOREXflash
...or you actually know which buy and sell orders to delete at $$ so that the EA can keep ranging and freeing up your margin so that you can keep adding positions?
Freeing up margin by closing hedging orders on the certain level(s) depending where the current price is. This way we possibly might be able to push the emergency fuse (closing the block when loss limit has been reached) further away in time (if not eliminate at all!?), as closing hedged positions at a comparably smaller loss (due to double spread and the current price vs hedge asymmetry).
  • #100
  • Edited 3:46am Oct 8, 2008 2:03am | Edited 3:46am
Learning the ropes, I read your PM......ok, il code it for you, dont spend money on this "simple" tasks, there is many programmers here who would code this for you! I never try to code that, but I think that it is very simple to add such variable. Give me a few days, when I got some time Ill code it for you. No problem man. Just be patient.
Respect
forexflash
0 traders viewing now