100% Win Math Grid Ea

I agree with you, leverage is a big factor for this method.....
I use fxpro, 1:500

The thing I like is a big return..
The best thing to do is to risk(like with any other method),
but as soon as you make the 100%, you withdraw money from acc,
and let the ea do its job...that is my MM

MM doesn work, compounding is a aditional risk.......

Guys sorry for not talking more but it is a hard task for me..Im from europe.

David,
We get very different results. I have to sleep on it and then I might explore what is going on. I tried your settings on USDCHF and crashed on July 17th.. It is 3 in the morning for me so maybe I am overlooking something obvious.

I added a screen shot and PDF of how many trades were open on July 17th.. I reduced the lot size to .10 to get this far into the time period.

Sorry, if I overlooked what time frame this EA should be ran on? Is it M1?

Also, what were your settings for the $500 account, if you don't mind sharing? I look forward to testing this ea today and the rest of the week. :-)

Timeframe doesnt matter becouse this is indicatorless ea!!!
Do not use this on real account.

Nvwine you are using to small increment....the smaller increment is, the more will price whipsaw
between levels! I tested this so many times, and the best increment is from 50 to 150 pips!
I never use increment smaller than 50. Trust me on this one. Also using more levels can really
make a difference, but it must be more than 3.

FF, I’m afraid that what I must say will sound negative. I don’t mean it to be, I sincerely hope that you can somehow use my (and others’) earlier comments constructively, to improve your EA. I’m simply trying to be realistic. And I'm not criticizing the idea behind the EA - IMHO, it's original and ingenious, like you say, a ‘Martingale on steroids’.

I don’t believe that it’s possible to completely eliminate the possibility of the ‘death trade’, no matter what parameters (INCREMENT, LOTS, LEVELS, etc) are used. Given enough time, price must eventually move in such a way as to create the ‘not enough money’ scenario. Like Pisara, I think it’s very possible that risk could be reduced by making entries only when market conditions appear more favorable, and that this offers better hope of surviving longer, than by simply trying to optimize the MM parameters. But I can’t see how there can ever be any absolute guarantee. The bottom line is that to enjoy a massive win rate, one has to take a massive gamble (i.e. be lucky enough to reach one's financial goal, and then quit, before the ‘death trade’ occurs).

Withdrawing money from the account does not solve the fundamental problem. Firstly, the less money there is in the account, the greater the probability of the ‘not enough money’ scenario occurring, so we’re actually compounding the problem. Secondly, while I agree that any money withdrawn is safe - and also that once the initial deposit is withdrawn, we’re playing with the market’s money - but the point is that the ‘death trade’ is equally likely to occur at any time, i.e. very possibly before we have the opportunity to make the withdrawal. So no matter how we slice it, it’s still a gamble, as to how we time any withdrawal(s), and how much we withdraw. This topic has been discussed on other Martingale threads.

Another possibility that’s been discussed is to take nominal losses when the situation appears to be ‘getting out of hand’ (i.e. set the MAX_LOTS parameter to a significantly lower value, to ensure that every block of trades is closed well before the margin is exceeded). But this is likewise a gamble, because there’s no way of knowing which (if any) of these situations will end up as the ‘death trade’, so we would end up incurring many ‘unnecessary’ losses along the way.

Nvwine, it appears that the EA parameters we used are identical. However, my the first trade in my back-test is at 07:55, yours is 00:02. I didn’t put a date range on my test, my test data simply starts at that point. The way the EA operates, a new block of trades is opened immediately the previous set reaches its TP/SL level, so a different starting point will result in a completely different sequence of trades being taken thereafter. (Also my data is sourced from Alpari, and has much poorer modelling quality, and mismatch errors.) However, it’s interesting how such small differences can cause a radically different result. IMHO, it further highlights the precarious nature of Martingale (or pseudo-Martingale).

I recall reading somewhere that Martingale is the product of wishful thinking, and the natural human desire to avoid loss.

The money markets have been around for decades, and they’ve attracted many millions of traders. I just can’t help but think, that if there was a MM-based formula that guaranteed success without risk, some brilliant academic would have discovered it long before now. Gaming theory tells us that MM in itself can not offer a real edge.

David

Do you think that 50 increment and level 5 would be okay for a $500 account at .01?

do Not Use This On Real Account!!!

Flash, how to activate this ea on the chart?
i tried to put it on my chart but it didn't appear anythong....

I'm testing it on a DEMO today not a REAL account... :-|

You said that it was able to survive on a $500 account and I wanted to test it on IBFX demo.

Agreed with David on that.

Besides, returning to the blown-up setup block shown in post #40, I can't see how the outcome could be profitable, even in the case the margin wouldn't be an issue and the setup block would be successfully closed on April, 16.
It's late and my math may be flawed, but here the situation in that setup just after all last sells opened. I roughly ignored slippage and swap, took the increment as 1 pips unit, the starting lot as 1 lot unit just for the shake of easy calculus. See attached image. Last row represents equal 1 lot unit orders on every Buy and Sell level (instead of default 1-2-3). In both cases the net results of the setup block that ended on April 16 are negative (!?)

Back in post #8, I suggested that somebody might try to make the EA's position sizing algorithm more efficient. By 'efficient', I'm assuming that the goal is to reduce the risk of the 'death trade' as much as possible, even if it means that the size of the gains are reduced in the process (but of course we still want to make a gain each time the TP/SL target is hit).

See the graphic below. This assumes that we're using 3 levels. The rows A thru I represent the different levels where the orders are placed, or profits are taken. The numbers in cols K & L are the number of increments from that point, to the profit taking level. The red arrow represents the path that price took in reaching either of the profit goals. The green arrows represent the new pending orders that are placed (that later get triggered) as a result of price reaching the corresponding green cell.

Col M is a 'straight thru' trade. The '1's represent the size of the orders that are placed, and the '6' in the gray cell is the profit (in units), i.e. the trade triggered in row D generates 3 units of profit, the one in C 2 units of profit, and the one in B 1 unit of profit.

Cols N & O illustrate how, when the first buystop order is triggered, the 3 corresponding sellstops are created, with the position sizes shown (1, 2, 3 units). The gray box shows that a net profit of 11 units is generated in this scenario. This profit is the net result of the gains made by all six sell orders, less the loss of 5 units (increments) caused by the single buy order.

Cols P & Q illustrate the scenario when 2 buystops are triggered. As far as I can tell (somebody correct me if I'm wrong), the EA doesn't create any pending sellstops when the second buystop is triggered. The net profit in this scenario is 5 units.

Cols R thru T illustrate the scenario when 3 buystops are triggered. The triggering of the first and third buystops generate the sellstops shown. The net profit in this scenario is 8 units.

Cols U thru W illustrate the scenario where price moves thru the zigzagging red line. In this case the net profit is 6 units.

My point is that the net profit (gray cell) needs to only be 1 unit in each scenario, for the system to make money on each block of trades. Reducing the number and size of the pending orders accordingly will significantly improve the longevity of the method, i.e. the number of trades before 'insufficient margin' occurs (the 'death trade').

Of course this exercise needs to be taken further, by considering every possible zigzag permutation. It's not necessarily a trivial task, but for those who seriously want to reduce the risk of ruin, I suspect that it would be a very worthwhile exercise.

However, because I don't believe that it's possible to eliminate the 'death trade' risk altogether (reasons in posts #31, #45), I would never have the nerve to trade this system live. Hence I will wish you all the best, and leave this thread (probably time I left the forum too, I just noticed that this is my 1,000th post - LOL).

David

[Edit - Footnote]

A couple of final thoughts:

1. I notice that when a TP/SL level is reached, sometimes the losing trades in the block get closed before the winning ones. Setting TPs and SLs relinquishes the control of how orders get closed to MT4. Explicitly coding the closes would give the EA that control. I don't know if closing the winning trades first would reduce the possibility of a 'margin call'; of course, the losing trades still need to be closed immediately after. It all depends on how 'available margin' is handled, but somebody might like to pursue this further.

2. Re the System Tester not halting on an 'not enough money' error, I believe that it's possible to explicitly code errors of this kind into the EA, by using functions like AccountFreeMargin(). Thanks to Tdion for reminding me.

David, I would just like to say thanks for your imput on so many threads, 1000 posts packed with usefull, interesting, and well thought out comments, just like what you have presented here. Whenever I see your monika, you will always be the first on my list to read. I just wish I had your gift for words.

I must admit I have never been a big fan of EA's. Even less about a martingale systems. For some reason ( not exactly sure why ) I have become interested in this EA and the way it operates. I played around with it over the week-end and was really surprised with the results. Not any massive gains or get rich quick system, just the fact that, for reasons I cannot explain yet, It does seem to have potential. I think in the back of my mind I have this idea that if one was to use it with a simple filter, such as only long trades if you believe the Daily is in an uptrending market, the overall performance would be increased by some degree.

Note. When you open this up it's already set to allow live trading, so caution needed.

Last thing, I hope I'm around to read your 2000th post....

GL & GT

Roy

well, clearly MM alone will not make it alone. But maybe we just could make trades that fits in this kind of MM with better chances of being profitable. FF already mede a modification making entry hour custmizable. May be other parameters could help to improve entering trades. Maybe customizing entry price so we could choose certain fibo levels, or pivot points or support - resistance etc.

Just an observation, but looks to me like the main reason your account could be blown out is due from overtrading. Why not just start the EA before the London open, and if TP hits, stop the EA and start again next trading day?

Roy, I said I was leaving this thread, but I had to thank you for your kind comments. I must admit that I persevered with this EA longer than I would normally have, because I likewise thought that I saw potential.

However, look at the end result of the equity curve attached to post #40, which is typical Martingale. By progressively applying technical (or other) filters it might be possible to reduce the risk of this kind of occurrence to (for example) 1%, 0.1%, 0.01%, etc, but I don't think it's possible to eliminate it altogether, and it takes only one block of trades to exceed the available margin, and thereby undo months', possibly years', worth of prior profit. Granted, everybody has a different appetite for risk, but personally I just couldn't sleep at night if I was trading this way. Still, each to his or her own.

I would definitely like to see this thread go somewhere, if it's at all possible. IMHO, there is much merit in the sheer originality of the approach.

Thanks again, and best wishes,
David

my connection was down while i was posting before. thats what i was talking about, and FF alreay modified the EA so we can choose the time to entry. It also has the option to stop trading when TP is hit if you set continue to false. Another issue i was thinking about is setting increment and levels in a value that the profits are within a range lower than the average range of the currency. I mean, if you start trading for example gbpusd at 7 - 8 gmt when london opens an put profits between a 150 range (2 levels of 25 pips + TP) it s highly probable that you ll get it. And even more if we coud choose the entry point (the EA is not capable to do this up to now)

Hi ForexFlash,

I tested it on GBPUSD.
On June08 it gain 274% ROI
On July08 it loss the whole account.

Try put this command in teh program for backtester purposes

if(AccountEquity()<0)
return(0);

David, I enjoy your posts, please don't quit just yet
I think the phrase in bold is the key here. The problem with this EA is that it doesn't take immediate measures to preserve a profit gained between the shifts on each level, thus we are actually progressively increasing the risk in time until the block is either blown (together with our deposit) or succeeded. Thus seems illogical to shift all that increasing risk to overload available margin that has otherwise many constraints due the the sheer physics of profitability. What I'm trying to say that we need some mechanism to manage individual positions' loss level + be ready to make overall decision to close entire setup block at any time.

I see 2 merits behind the idea:
1st, it tries to manage trades to the end of setup block automatically, even when disconnected. That is unfortunately not quite true because we need watchdog to place new pending orders and close pending orders left in the end of block;
2nd, it seems it prefers trending setups vs ranging, that's alone is different from the rest in its class and may be advantageous.

Well, with big IF that somehow that risk will be dramatically reduced, why not. IF the risk to blow an account within one setup block is comparable with some long term trading system slowly but surely eating your account after some changes in market behavior. But IMHO looks like reducing risks THAT much is the non trivial task to achieve.

Did you think of that there is a hole in testing data?
Or some of the trades isnt executed due to unqnown problem!
I assure you that this EA does what it needs to do, I tested this so many times
that I got a bleedage in my brain
Try to test it with 1min timeframe...that is what i do....I already said that tester is using interpolation for candels, and there will be "abnormal" ups and down within candle....do not use any different TF than 1min to max 5min! I tested this again, and again, and again, ang again,and again, and again, ang again,and again, and again, ang again,and again, and again, ang again, AND IT WORKS!

Again sorry for my bad english!!!

FF, I did my test on M1 ticked in the first place. No, no holes in data. The gaps and spikes are the only major problems that can be with data, then we will not be filled at all or improperly. I don't see this on the setup block in question.

I have 2 problems. First attachment shows that 3 pending buys were triggered (all those on on level 1) and caused exact 2 pending sell order sets (0.1, 0.2 and 0,3 lots for each level) placed. Is that supposed to be so?

The second problem is my math. There are 2 different calculation paths shown in the yellow cells on the second picture. Just for that beated to death case. Providing there was enough free margin to open last buy positions thus it could complete the block. None of them show positive result! Maybe my both calculations are wrong (calc on the bottom row is cumulative one provided we will close the block on upper TP/SL level; the formula for the rightmost column shown in Excel Formula bar), or the idea has something open... I'm lost so far

Edit: I corrected the calculus on the bottom row. Now the both are -76, i.e. HEAVILY negative expectation for the finished block in this scenario even with infinite margin!