Martingaling - Looks so good until the inevitable

This is probably because account longevity using a Martingale system correlates with the win rate. It speaks for itself that the higher your win rate the longer it will take to be wiped out, statistically:

Let one round be defined as a sequence of consecutive losses followed by either a win, or bankruptcy of the gambler. After a win, the gambler "resets" and is considered to have started a new round. A continuous sequence of martingale bets can thus be partitioned into a sequence of independent rounds. Following is an analysis of the expected value of one round.

Let q be the probability of losing (e.g. for American double-zero roulette, it is 10/19 for a bet on black or red). Let B be the amount of the initial bet. Let n be the finite number of bets the gambler can afford to lose.

The probability that the gambler will lose all n bets is qn. When all bets lose, the total loss is

http://upload.wikimedia.org/math/b/f...072ba02201.png

The probability the gambler does not lose all n bets is 1 − qn. In all other cases, the gambler wins the initial bet (B.) Thus, the expected profit per round is

http://upload.wikimedia.org/math/6/e...62b6a9f5fa.png

The probability that the gambler will lose all n bets is q^n, where q is your loss rate and n is the number of bets. I have made a quick graph, it is really easy you can do it yourself too, just replace the value of q by the loss rate of your strategy:
http://fooplot.com/#W3sidHlwZSI6MCwi...ItMCIsIjIiXX1d

As you can see the value of y approaches 0 but never actually reaches it unless q is 0 (win rate is 100%): this would be the chance of you getting a margin call with a Martingale system. 7 losses in a row with 0.5 loss rate still has a 0.78% chance of happening: which is 1 in every 128 trades!

I never stated that I see no valid reason to use a Martingale system, I was merely advising caution. We are all risk takers, the decision lies with us if we want to use it or not. I will summarize that however Martingale should not be used to turn a losing system into a winning one: it is just a dangerous tool which becomes exponentially better with win rate. In other words if you want to use Martingale, make sure you at the very least have a win rate to support it.
Edit: to expand upon this 7 losses in a row with 77% win rate has a 1 in 29370 chance of happening, compared to my previous stated 1 in 128 for 50% win rate.

To engage in constructive discussion about a subject, we must first agree on the definition. Since you cannot attack the argument itself, you change the definition of Martingale? Please elaborate on how you see its definition then. You also bring up MM, which I presume stands for Money Management. I would like to hear from you what you understand as a good MM to support this strategy.

whist i agree martingale is usually destine to failure

it really depends where the starting points are
if one starts from the bottom or top of a large range the chance of losses are diminished somewhat
and especially if you are trading an instrument that usually always ranges on the larger TFs and you know where the ranges usually always start and finish

if you start shorting at the bottom of a range - or longing at the top of a range obviously the chances of disaster
are much increased,
so martingale should (imo) be used with larger TF ranges, and never on its own merit

someone posted an example of a loosing martingale system the other wk for the GBP/USD from myFXBook
(or whatever its called)
but it had nt actually lost at the time - it was only about 25% down
those that were following possibly from nr when the GBP was starting to selloff had been shafted
where as the person who's system it was and whose who had followed from the start were still hugely in profit

i think i'll patent the idea as Dynamic range martingale
(no equations i'm afraid have nt done those for a long long time)


just a thought

100% agreed.

Full disagreement there , there are many kinds of martingales, and multi-martingales and some of them don't need a good entry signal. A martingale can have the edge on other aspects of the trading.


I would define a Martingale as any system that adds positions increasing the volume when you are losing.
How you increment the volume, when you add positions, how you close some of all positions, how you get the ride when you start winning with a series of positions, how you hedge,.. ... ...
For me they are all martingales and all very different from the basic definition. I dont really think anybody with some trade experience would run a basic martingale that just doubles after a trade is lost.

My current setup was described earlier in this thread: http://www.forexfactory.com/showthre...56#post7679956
It's one of the possible ways, and I can't say it works but I'm fairly sure that the key for Martingales is a sound MM (call it gut-feeling if you want).

My point is that things are far too complicated to be able to dismiss just with some maths.

t.

tacticat, if you have a martingale EA, have you back test it?
Run it in several pairs with at least 10 years of history data, look at the result and show us if you don't mind.

What would that prove? I can provide you not one but any number of martingales that would run for 10 years in a symbol and have spectacular results.
And, to be honest, I would recommend you NOT to run that EA with money.

Curiously martingales are the best EAs to provide nice curves on backtests (and that says how good backtesting is )

t.

Nothing to prove, I just want to study and analyze it.

I am against any kind of martingale systems so I won't run it even with demo account.

By this definition even Dollar Cost Averaging would be a Martingale system, but it's not...it is the average of a position closer to the market price from where you started. Martingale betting system is doubling up on a losing hand, which would approximate to you using hard stop loss and then reversing position in order to win the initial bet.

That would assume: Position A) 2 Lot Entry: 100.00 TP: 101.00 SL: 99.50

Position B) 4 Lot Entry: 99.50 TP: 98.50 SL: 100.00

This is Martingale...everything else where you leave a position open is not true Martingale because true Martingale requires position reverse and strictly speaking equal bets, which are not when you DCA...using martingale theoretically you always win until you run out of capital, using DCA you either go for a small loss, brake even or small profit.

I started using DCA without stops, and I always use it to brake even or go for a small profit if I can. Never to win, because with DCA I assume that my previous bias is wrong or poorly timed. I will usually allow myself some degree of DCA if I can pull something out of it, but if the market starts a reverse frenzy against my positions you have to cut it unless you have mega amount of capital. My cut off is at 50% of the account.

You must differentiate a Martingale philosophy vs Dollar Cost Averaging philosophy....

This is my rudimentary take on adjusted and not pure Martingale (attached in Excel):

Let me try to explain: You have 3 chances to be right and 1 to be wrong and take a 500 pip loss if using standard lots...

If you assume a position I) you have two outcomes: a TP or activation of 2nd position which activates 3rd position and int the end activate 4th position and you either brake even or suffer a loss if the price bounces 5 times between your prices, which would tell you something right

Now you can see that in this mine interpretation I was actually combining DCA with Martingale because of smaller lot size positions and in regards to margin required. I never tried it on real account, only once on demo but you have to factor in spread and commissions, and I always like to do everything manually, I trade manually so it gives me a headache in calculation so I never did it.

Anyway like I said this is what you want if you want to do Martingale, not DCA..

With ya, on everything you've said.

I started working with what was available for EAs, running hundreds of backtests, forward testing in accounts to see that most of the logic that was put in to code, was total garbage.

Then, I read...

People who code, typically cannot trade.
And people who trade, typically cannot code.

I found this to be very true. There is a grand disconnect from what somebody who can code really well, to those who do very well in trading...typically.

There are exceptions to all rules, but these are just that...exceptions.

After making my way through a few EAs, I worked with a "Coder", to implement my ideas on a particular strategy...found out, that my words never translated well into the code and on my first real martingale runs, my EA that was hitting 10-12/per trade, all of a sudden was 800 in the shit, on a 1300 account...after a couple of hours of this thing massively cost averaging, I escaped the trades with a $35 profit, shut it down and never have run the same cost average logic on a live account since.

I started to learn about making the changes to the code in the way that I wanted to see changes made and have evolved not only in a manual trade strategy, but have been also able to evolve the code.

Thing is...I had actually come to a point where I'd given up. Said the same thing as has been said throughout this thread, in that Martingales are shit, and they will only blow up your account...shut down my metaeditor and didn't look back.

I can trade profitably...have even done so as an account manager, why bother pissing around with these??

It's like this...

I happened on to a thread in another forum, where somebody asked a question about the best martingales that others had seen and used...and one member had posted one of the EAs that I had thrown together and shared for the house. His claim was that he started it and within almost 4 months, he was near double on his account.

LIVE

$10k start.



It took him to about 4.5 months to fully double the account...and others cheered and celebrated with him, as if they too were sharing the profits. What was even better, is that he shared his complete strategy, and set files so that others could too, take advantage of the same offering.

Now, ya see somebody shake out a few hun in a month on an EA...who cares.

You see somebody double an account and a light goes off.

I started working with him and opened up my editor again and haven't closed it since.

After going through the code that he was playing with, I realized how poorly it was put together and figured if he could do this by accident, why couldn't I actually improve on this same logic...

Since then, he has doubled that account a couple of times and spun off other accounts with my new breeds...With each development, he sends me a profit report and his ideas on how it can be either improved or most effective.

Lemme tell ya...somebody that uses this logic and has churned over 40k in a single year, is somebody you listen to. Not these fucking wankers who wouldn't understand the concept if they watched it live.

So, from there, we continue...

I have restarted my accounts, and actually have 7 accounts trading different strategies, all of which have been funded by a single account...LIVE.

And then I have also run several demo accounts doing the same.

My application doesn't come from an idea that something cannot be done. My application comes from the idea that I have not only been witness to others, but I have done so myself.

The little group of like minded at my forum are a great group of people that too, are a lot like the FF members, and in fact a lot of them have migrated over to the new house with me.

They don't have 10k to start, but they do have a passion to learn and an idea of the risks involved.

Why do I run the contests...

To show what can be done.

To encourage others to TRY.

And if in the end, they tap out of market or tap out of the ideas...they at least have a solid foundation to build a bias...not on words that somebody else said.

Look at that fucking wanker who quoted Einstein as a keynote speaker on Martingales...

Of all of the things I have ever come across on Einstein, not a single fucking word has been spoken about his attributes as a Forex Trader...or any other market.

The EA that has brought me back into the game...is posted here at FF.

HERE.

All honesty, I've never ran it for more than a single week, and never on a live account.

Absolutely right!

This is my experience too. Unfortunately.
We worked with a great coder. Some of the functions and buttons on the various scripts were excellent, and stand alone from any comparative scripts.
He released some MT4 management scripts to the market, and they are top quality. I use them daily, and they have excellent functions.
Our personal EA's were equally smooth in terms of functionality, but I noticed one major thing which started to make things difficult.

The coder was exceptionally skilled in his profession....but he was not a trader. I mean, he had a good understanding of the functionality of MT4, and a basic understanding of the principles of trading. But there was a distinct lack of trading logic that an experienced trader would automatically set into the code. This made the development of the EA's a very a trial and error process, with long times spent demo testing the various features.
Sometimes this was costly because once applied to the live account, we found buried issues in the coding logic that literally cost $100's in unnecessarily lost trades.
My coding skills are non existent so i was helpless in this area. Also the coder was getting more and more frustrated having to fix reported errors that he had not encountered when he demo tested it himself. And that was the disconnect in my experience. For the management scripts it was fine, but when it came to programming a trading system, a certain amount of knowledge is needed. In most cases this is absent.
It's another reason that I take all freeware downloadable codes with a pinch of salt. Could be that it's great code, with little trading knowledge, or its great system with poor code...as you said....the best of both worlds are exceptions.

The guy you mentioned with the $10k start is a brave man! Honestly I would never apply any EA to that amount in my account....anymore! But, fair play to him, and it was clearly successful. And if it was done with poorly constructed code, then it does give some hope that there is a way to make it work.

As I stated previously, for me I have reservations about letting any EA, martingale or otherwise run unchecked on my live primary account. I believe there are certain instances where human discretion is required, and intervention is either helpful or essential. We programmed in protections from gaps/slippages, and many other perceived issues in trying to build the EA, but invariably, it still had to be monitored to make sure it was not exposed to obvious issues. Even then, it hurt eventually. For me, it's not worth the stress.

I am happy that there are those still working with EA's, like the members of your forum. And as i said it's refreshing to see people leave links to the ones that were unsuccessful. It shows progression in my opinion. I hope one day I get inspired to start playing with them again, because in principle, I like the idea.

However, there is another part to the equation, that you touched on previously. In truth, a persons aversion to one idea or another could be more down to personality and interpretation.
I think you said in an earlier post, "if we all had the same information, would we all trade the same?"
I believe this to be at the core of the issue.

Some people will never trust EA's.
Some people ridicule systems with low R:R ratios.
Some people INSIST lower time frames cannot be profitably traded.

When people become fixed, for the most part this view will not be changed, despite any evidence presented to the contrary. It's a feature of human nature unfortunately.
In trading, we see it all the time. Many threads are started with a concise, if not simple methodology laid out.
For every 10 traders that view the rules, and apply the idea, 7 or more of them will then add or subtract features according to their taste. And each in different ways. It comes to the point where the trades listed are barely recognizable from the original idea.
The point is you can give 10 people the same tool, and nearly all of them will find a different way to use it! Some people will throw the tool away because the can't, or don't want to learn how to use it. The same applies here.

Experience has stung me with martingales in the pure sense, but I encourage others to experiment. And I watch with interest. As far as quoting Einstein to disprove it, that tickled me a little too. I can understand your frustration, but i feel getting angry about this type of thing is energy wasted. Keep trying, and encouraging others. Energy far better spent But always keep in mind you will never prove anything to someone who doesn't want to listen.

I experienced something similar in another thread...I cannot understand someone who insists on aggressively dismissing something. Telling traders what they should do is not our business. Let anyone experiment. But I found one guy so bitter his thread title was telling people not to trade a certain system. I found this ridiculous, and actually baseless. More about a disgruntled individual than anything else.

It's something I believe is quite common across many forums. I am definitely not referring to specific individuals in this thread, or others but in general terms;

There are people who come to genuinely learn and share ideas, people who come to sell their wares, and an unfortunately large population of people who have some narcissistic need for praise and recognition.

Those ingredients are listed in ascending order of quantity

Read most of what is in this thread and there were few things that encouraged/discouraged people from using a martingale system.
1.Drawdown is the first thing people are afraid of and it is normal to not want to lose money, but at the same time you should not take risks that you cannot afford to take. What that means is that you shouldn't over leverage your account, shouldn't push it to make 15% return a day and watch the system all time while it's running.
I cannot give anyone advices on how to trade and simply explain how i trade using this method called Martingale.
I scalp 5m timeframe with it and i use a high leverage of 1/1000 because i aim to have a very tiny profit from a single order. With a leverage of 1/100, what people would normal choose with a starting 5k account, an order of 0.05 would bring me 5 eur with a tp of 10-12 pips, but having 1/1000 lev i take much less.
Why would i want to take less ? Well because with this system you may need to open 10 positions at a point and i want those 10 positions to mean nothing for this account.
10 orders of 0.05 = 0.5 lot and on this account a 0.5 lot = 30 eur
30 eur x 5000 means i am investing in bad market conditions less than 1% of the capital and intend to close all together with a small profit. Why i said "bad market condition" is because i don't open 10 orders unless the market is moving against me.
2. This system fails in time and you lose one day everything you put on the account in few weeks or months.
Let's have a look at the account i'm using martingale on and see it's results.
http://prntscr.com/4hkpp8
1% a day growth with 35% a month and DD 19%
As you'll say the DD is 20% of the account and that is too much but for me, that 20% came after the first month when i had already 35% growth and in bad market conditions when news spiked this currency.
It fails in time... How much time ? 100 trades ? 500 trades ? I have over 3000 trades closed and 600 lots in total have been traded.

95% of traders are losing money and maybe one day i'll be part of them but untill that day, let me trade using this unsuccessfull system that produces money daily and has shown consistent growth on my account.
Looking at the equilty curve i don't see a day when this system has closed in loss, and while this system doesn't work i have 10% growth this week and all month so far a 33.3% growth.

This system is not for sale as only crap is sold these days by scammers.

Maybe it was quite the contrary - from 10000$ to 1000$ and the trader forcedly ceased his flight of genie ?

From my experience Martingale is cause-a-disaster strategy, not because you put wrong trades, but because the common sense hazed with confident gamble-feeling that you are determined to win at the end. You are making stakes with no doubt, fear. MM? - outta of my way! .

It is true on paper, but it's not true when you have spreads, market, moving out of the 50/50 odds, limited lot size and many more factors ..

Never ceases to amaze me how much can be written about such a simple concept. Martingale, like any other approach (and I do mean ANY other) is a just another system or strategy.

Those that know how to use it will make money. Those that don't, won't. Just like ANY other approach, it's only ignorance of it that will cause it to fail.

Just trying to understand...if you are keeping the same lot size, then really it's not a martingale in the true sense of the word is it?
I don't know your system, but the overall principle in martingale is some kind of position size increase for each new position in the grid? Most double, but i fully understand not everyone.

So, if you don't increase the position size, then how can you maintain a 10-12 pip TP distance, because if you are at the 7th or 8th level, closing them all at the last level's TP will not bring the total trade bucket into positive? Sorry, but this statement leads to more questions than answers.

What most people are apprehensive about is the position size increase. Because simple maths says that (in the doubling martingale case) if you start with even 0.05 lots, after 10 steps you have to put an order of 25 lots into the market, with 25 lots already in negative.
Most accounts can't survive that.
I'm not saying this is what everyone does, It was not my exact approach, but I'm trying to point out why people turn away when they see the word martingale.

And, the number of trades is absolutely irrelevant to any systems longevity. Most martingales have frequent entries, depending on the grid size. In this case longevity is measured in time, and with changing of market conditions.
I noticed your screenshot is only the last 2 months.
Sincere congratulations on the success so far, but 2 months is nothing. And, moreover, the conditions in the market during this time have been low ADR's, low volatility.....perfect for martingale systems.
Longevity will be established after 2 years of consistent results.

I hope you are still as confident after this time.

Sorry, I fail to see how you can prove that. Unless you refer to very simple martingale techniques.

Look at this:


These are the equity evolution of two systems. operating quite frequently on 32 pairs.
- 1st one (red) is a system whose entry signal is received when a combination of up to 7 indications (price against a couple MAs, Relative position of MAx, MACD,...) detects that there's a stablished trend in H4 but a retracement in M15. We open an order in the long term trend.
- 2nd one (blue) is exactly the same BUT we apply some martingaling (opening up to 4 extra orders with bigger volume and closing everything if the 5th level is reached)

In this case, at least, it seems that martingaling turns a negative system into a positive one.

t.

Being a Martingale i don't chase pips but profit therefore i'll calculate profits here and will take a buy order and how i make profit with it.
Deffinition of martingale was to double the position everytime and i can calculate that also on how/when to increase but here's a strategy similar that. By increasing every 4th position we limit the open orders and not risk having 25 lots open in 10 orders.
Martingale can run without having 25 lots open on 10 orders and you just have to change the ecuation for that.

The position that will go for a loss will have tp (in loss) where the winning position will close. It is necessarily to have tp and not sl because of how the broker will calculate to close your position and not to have one closed at bid and one at ask.
All orders will have same tp at all times and that will be manual adjusted.
If you are using tp and sl than the ecuation will change and this may lead to a loss.

A 0.05 tp 10 pip = 0.5 eur (if 10 pip = 0.5 eur than 2 pip spread = 0.1 eur and 0.5-0.1 = 0.4 Eur profit) - If a losing than
B 0.05 open 10 pip from A with TP in A (A =0 B=10 than profit is 0.5 eur -2 pip spread for 2 orders . 0.5-0.2 = 0.3 Eur profit) - If losing than
C 0.05 open 10 pip from B with tp in 1/2 AB ( A= -5 B= +5 C= +15 A and B are canceling eachother leaving with C as profit and spread to be taken off from C .
C=15 pip/ 0.75 eur - 3x2 spread = 15-6 = 9 pip profit or 0.45 eur
D 0.05 open at 10 pip from c with tp in B ( A= -10 B=0 C= +10 D= +20 profit 20 -2x4 spread = 12 meaning 0.6 eur
Martingale starts working at E as first four were only hedging for small profit
E 0.1 open at 10 pip from D with tp in 1/2 BC ( A=-15 B=-5 C=+5 D=+15 E=+50 since A+B=C+D we have E with 50 left and we substract 2x6 spread leaving 38 profit = 1.9 eur. Only 5 orders open but we substract spread for 6 since E is double or it could easily be 2x 0.05
F 0.1 open at 10 pip from E with tp in C (A= -20 B= -10 C=0 D=+10 E= +40 F= +60/// Profit 80 - 2x8 spread = 64 or 3.2 eur
G 0.1 open at 10 pip from F with tp on 1/2 CD ( a=-25 B=-15 C=-5 D=+5 E=+30 F=+50 G=+70/// Profit 110 - 2x10 = 100 or 5 Eur
H 0.1 open at 10 pip from G with tp in D (A=-30 B=-20 C=-10 D=0 E=+20 F=+40 G=+60 H=+80/// profit 130 - 2x12 = 106 or 5.3 Eur
Every 4 open orders we double position.
I 0.2 open at 10 pip from h with tp at 1/2 DE (A=-35 B=-25 C=-15 D=-5 E=+10 F=+30 G=+50 H=+70 I=+180/// Profit 280 - 2x16= 248 or 12.4 Eur .
we calculate I as 4x A or 4x0.5 therefore spread is calculated 4 times.
J 0.2 open at 10 pip from I with tp in E (A=-40 B=-30 C=-20 D=-10 E=0 F=+20 G=+40 H=+80 I=+160 J=+200/// Profit 400 - 2x20= 360 or 18 Eur

For this to happen and to have 10 positions open we need to market to travel 90 pip against us without retracing enough to hit our tp.

Either way, profit comes from pips. So in your calculations you must have a predetermined number of pips in mind that will cause a TP.

"Deffinition of martingale was to double the position everytime and i can calculate that also on how/when to increase but here's a strategy similar that. By increasing every 4th position we limit the open orders and not risk having 25 lots open in 10 orders.
Martingale can run without having 25 lots open on 10 orders and you just have to change the ecuation for that."

Sure, I didn't say everyone doubles from step 1. I certainly didn't. I'm familiar with the calculations.
But regardless, the principle remains that you must increase the position size, and/or your distance to TP once the market goes against you by a few steps in order to have an overall positive result. And of course TP is used to close out all orders, this was a feature built into the code I used. Every new step opened, modified the TP accordingly for all open positions.

"For this to happen and to have 10 positions open we need to market to travel 90 pip against us without retracing enough to hit our tp."

This is the part where i feel you will come unstuck.

I followed a similar idea to the one you listed. However my way, I could survive 140 pips against me. I did used a capped system so as to avoid total account wipe out.
One day on my beautiful non trending, low volatility pair, without warning or news event, the market decided to run against me without so much as a 12 pip retrace in my favour (which would have closed the whole lot in positive, or at least BE) from a new step level. I lost that bucket

This is why I did not continue with martingales.

Sooner or later, it is inevitable that any pair will make a run. And due to the fact that we choose pairs for martingales that are less volatile, with lower ranges, it stands to reason that all movements on that pair, in particular retracements, will be of smaller magnitude also. So, when it runs, you are left hoping for a retracement that may never be enough. if you are lucky, you will only be at step 1 or 2 when a run happens, and you may survive. But if you are at step 5, and as in the example you used, are opening new steps at 10 pip intervals, then sooner or later you will be caught out.
This can take various forms. Most commonly, your account size is not large enough to take the new positions required, so you get margin call. Otherwise, like me, you predetermined a maximum loss and you took it. Either way, this risk is a seemingly unavoidable feature of martingale systems, because by definition you are adding to losses. In most cases you can come out positive, as it will come for you. But....it only takes one case to catastrophically deplete your account.

Sure, you can apply larger grid steps etc. but doing that reduces the frequency of trades, and therefore profits. So in my view, if you have to play it safe with martingale, and rake in a few dollars here and there...is it really worth it? Why not trade less open, and more lucrative systems...if you can?

I'm not saying it's impossible, but i am yet to see any variant or calculation that eliminates the threat of huge losses. I am watching as people try with interest, but I'm afraid as yet, ive seen nothing acceptable for my personal style.

Because it is simple math. The lower your win percentage with a strategy the higher your chance of getting a margin call, which in practice means the faster you will get that margin call. More losses equals more tries equals more drawdown.

Yes, you're right. Simple math... for a simple martingale that I don't think anybody uses.
Martingala lovers (mostly using EAs) use currently very different martingala techniques, all sharing the increase-risk-when-losing but adding trailings, partial closures, smart entries, increasing opening gaps, multipair hedging,... a whole world out there.

t.