Does a safe Martingale exist ? - page 5

Bob Davids  
iGoR:

So I think david and I reapeat I think that the way I calculate is a safer martingale if we first look to losses and risk and only then to possible profits.

Do you agree ?....if not we are going to go further on this. If yes then lets move to the next step.

Looking forward to your reply.

Friendly regards... FXiGoR

Hi iGor,

Thank you for your reply. Here let me clarify, we have some disagreement earlier when we're discussing about 10point3. Can't remember where is the content, but you have disregarded my point and we're standing our own ground for sometime. I must apologize that I have make such a big deal for such small matter. I have not deal with you that long for those silly argument. So, no problem now.

Let's talk a little bit about martingale concept. Mr. Martingale was a gambler and came up with this idea. If he continue betting with a series of double down stakes, as long as he win, he'll get back everything he loss with an extra dollar in pocket. Now, the fact is, the ERROR. When Mr. Martingale thought of this idea, he hasn't really go through his pocket. By logic, this system will NEVER loss you anything provided your capital is huge enough. By logic, this system will NEVER win if you have a series of losses of more than 40 folds. Because if you lose 40 times Martingale bets, the amount of money will have to stack on your foot all the way to the SUN. If you goes 41 folds, it will be heading out from the solar system. People out there, this is a simple method I threaten people when they wanted to convince me they're master in martingale. Get a piece of A4 paper, measure the thickness of the paper. Fold it from the middle, and measure the folded A4 paper again. Then fold it again, you'll have 4x of the thickness from the original thickness. Keep folding the A4 paper until 5th times, you'll figured the damn thing is too thick. You wouldn't have the strength to fold it anymore. So, off you went to your garage and proceed to 6x by using your G-clamp. After you folded 6x, your G-clamp probably bent because the amount of force you applied to fold it. By the time you measure the thickness of the folder A4 paper, it should be around an inch thick!

So, what is all about the grandmother story I told. The force you applied on folding the paper is exactly the force you applied in trading when you're facing a losing streak. After 6 consecutive losses, the amount of stress in your head is exactly the same as somebody trying to apply 6x folding a piece of A4 paper on the G-clamp. You will lose your logic thinking when you see abnormal lot size. You will misjudge other trading opportunity. Your lot size calculation and risk management all gone. You're mentally unstable and not fit for having sex ahem. Pardon me. I'm kidding, but don't be surprise if you couldn't get your dick up. That's what I heard when people stress, they tend to have erectile dysfunction.

To conclude this, there is only 2 possibilities.

1) You walk away after you win a series of betting. The fact is:

a) Your reserve fund is just enough for you to win this series of bet.

b) Your emotional sustain does not vary your betting position

c) You walk away JUST BEFORE you lose your ass off(we call it luck)

2) You lose your ass off after a series of betting. The fact is:

a) Your reserve fund does not compete the series of losing streak.

b) Your emotional does not sustain you to the extend of losing your ass off.

c) You did not walk away BEFORE you lose your ass off(simply call it bad luck)

So, iGor. Regardless of how you wanted to make up the TP or SL, you either win or loss, provided all the facts/criteria/requirement meet at the same time. I have read your spreadsheet. When you talk about take profit of 60 and stop loss 20, that is exactly the way to put your accuracy down to... may be below 33%? Meanwhile, you reduced your progression down to... may be 1.3? or 1.6? Doesn't matter, because your take profit is 300% of your stop loss. As long as your 33% accuracy works, otherwise you'll be hitting your maximum lot size very frequent.

There is no right or wrong in betting. I am trying not to offend any one here. There is no low risk martingale. Only thing exist in martingle is, profit before its too late. Walk away before it lose your ass off. Let me simplified my thoughts and share with you again:

1) You have huge TP, small SL - Lower your progression to sustain more losses

2) You have low TP, huge SL - Lower your progression but increase your starting lot size, so you profit from your winners more before you sustain a huge loss after progression.

3) You have low TP, huge SL - Lower your starting lot size, but increase your progression multiplier. When you suffer a series of loss, thats when you make huge money.

4) When you have similar TP/SL - You have no other choices. You can only do 2x multiplied progression.

I don't mind to spend time in finding other possible method to delay the wipe out incident. If you have other idea, let's hear it. Meanwhile, iGor's idea can be simply put into an EA and let it run for weeks on demo.

Regards,

David

p/s: Again I would like to emphasize, safe Martingale does not exist. Changing TP/SL/Progression/D'Alembert/Cost averaging is only repackaging. Sooner or later, you'll still get it. Mr. Martingale does not think that much because you only do BINARY when you go blackjack. You place a dollar bet, you lose a dollar. You place 2 dollar bet, you win 2 dollars. You replace your initial loss and you have gain $1 extra.

Bob Davids  
iGoR:
Hi Davidke,

Again let me repeat that I am AGAINST martingale. So I am not here to defend martingale or that I like martingale. That was why we had in the past some issues between us about 10point3.

Again when I opened this topic it was to see if there are safer ways of "martingale". Because like you say Mr Martingale was looking to his system with the fact that loss and profit are equal.

Like flipping a coin or going to a casino with no limits and no zero on the roulette table.

I know very well the effects of a S/L that is maller or bigger then the T/P. That is what I have been trying to explain already for a long time here on this forum that markets are 100% random. Because one will never find a system that over a LONGER period of time and on different pairs, will produce a hitrate bigger then 50% with a S/L that is equal or smaller then the T/P. The moment you can make yourself such a system you found the Holy Grail and you proofed that markets are not random ( where up till know nobody succeeded in doing so-- they are convinsed that markets are not random but they can not proof it with facts--like I would say that God really exists but I can not show a picture from him).

BUT...

When we place a system on a pair or index or any financial instrument and on a certain time frame, we know that depending on the system we use, we will see that our system is making money in the trending periods and is losing in the ranging periods. Or vice versa that an other system will make nice money when markets are consolidating but is losing a lot of money in the trending markets.

The reason I say this is that there is a difference between random price action on a pair or a market and random flipping a coin or random that we can have black or red on the casino.

In the casino or in flipping a coin, the money that we loos is equal to what we win.

But when we place a trend following system on a pair or market then we will see that the average losses are many times smaller then the profits one will make if the price is in a trend.

In other words: smaller losses but profits bigger then the average loss with a hitrate lower then 50%.

Do you agree on this ?...if no then we can get deeper on this if yes we can go a step further...(if we go further I will explain more what I want to do with the type of martingale I have in mind.)

Looking forward to your answer.

Friendly regards... FXiGoR

PS. When I say safer ways of martingale we are not using martingale anymore:

Martingale:1,2,4,8,16,32,64,128,256,512 x1$ = -1023$

My idea: 1,1,2,3,5,6,8,11,15,20 x1$ = -72$

Hi iGor,

Thanks for clarifying. I totally understand and agree with you. So, what are we going to do?

Thanks and regards,

David

iGoR  

Hi Davidke,

Again let me repeat that I am AGAINST martingale. So I am not here to defend martingale or that I like martingale. That was why we had in the past some issues between us about 10point3.

Again when I opened this topic it was to see if there are safer ways of "martingale". Because like you say Mr Martingale was looking to his system with the fact that loss and profit are equal.

Like flipping a coin or going to a casino with no limits and no zero on the roulette table.

I know very well the effects of a S/L that is maller or bigger then the T/P. That is what I have been trying to explain already for a long time here on this forum that markets are 100% random. Because one will never find a system that over a LONGER period of time and on different pairs, will produce a hitrate bigger then 50% with a S/L that is equal or smaller then the T/P. The moment you can make yourself such a system you found the Holy Grail and you proofed that markets are not random ( where up till know nobody succeeded in doing so-- they are convinsed that markets are not random but they can not proof it with mathematical facts--like I would say that God really exists but I can not show a picture from him).

BUT...

When we place a system on a pair or index or any financial instrument and on a certain time frame, we know that depending on the system we use, we will see that our system is making money in the trending periods and is losing in the ranging periods. Or vice versa that an other system will make nice money when markets are consolidating but is losing a lot of money in the trending markets.

The reason I say this is that there is a difference between random price action on a pair or a market and random flipping a coin or random that we can have black or red on the casino.

In the casino or in flipping a coin, the money that we loos is equal to what we win.

But when we place a trend following system on a pair or market then we will see that the average losses are many times smaller then the profits one will make if the price is in a trend.

In other words: smaller losses but profits bigger then the average loss with a hitrate lower then 50%.

Do you agree on this ?...if no then we can get deeper on this if yes we can go a step further...(if we go further I will explain more what I want to do with the type of martingale I have in mind.)

Looking forward to your answer.

Friendly regards... FXiGoR

PS. When I say safer ways of martingale we are not using martingale anymore:

Martingale:1,2,4,8,16,32,64,128,256,512 x1$ = -1023$

My idea: 1,1,2,3,5,6,8,11,15,20 x1$ = -72$

iGoR  

OK David lets continue our search...

If we now concentrate on the trend following systems, for many years I have been testing trend following system and a good and robust trend following system gives a hitrate of +/- 37-40% and a maximum consecutive losses between 13 to15. (I have done this testing for many years with Metastock professional with hundreds of different indicators and different rules and filters).

If we reach this maximum of 13 to 15 consecutive losses we will see that price is indeed in a very strong consolidation period. But the moment it breaks out of that consolidation then we see strong trending moves.

Those trending moves are many times bigger then the average losses we made during that consolidation period. By many times I mean many times bigger then 3 x times the avg. loss or stoploss.

I posted a picture. On the picture you can see buy and sell arrows. An arrow apears ONLY at the close of a current bar and once an arrow apears it will never disapear (So these arrows do not have any repainting elements what so ever).

Each buy arrow apears after a bottom was made and each sell arrow apears after a top was made.

Lets now say that we take the highest high of each top as our stoploss for the sell signals and the lowest low of each bottom as our stoploss for the long signals. Lets call them strategic stoplosses and not a fixed stoploss of for ex. -10 or -20 pips.

The problem is now that the value of each stoploss will be quite different. Because we have tops and bottoms that are made in a slow way and they give very small stoploss values. But we can have very fast and sharp tops or bottoms and they will produce big stoploss values.

So with the help of a backtest we check the average stoploss value of all tops and all bottoms. Lets now say that the average value would be -20 pips. Then in our system we will allow the system to take only a trade when the stoploss value is -20 pips or LESS. So all tops and bottoms that would give -25 or -30 or -40 pips we do not take those trades.

By making this first rule we get ourself a stoploss that is placed on a strategic place (above top or under bottom) but that has also a fixed value which we need if we want to work in a mathematical way. But important also is that sometimes our value will be smaller then -20pips because we had a small top or a small bottom. And this is going to be an additonal edge we will get in our profits.

If we have a maximum stoploss value of -20 then according my idea of contracts sizing then the profit target needs to be 3 x times bigger then the stoploss in total +60pips. Now we have 2 facts that we know when we will trade that is that a losing trade will be -20 pips or smaller and a winning trade will be +60 pips (in reallity I would work in a different way by looking to the market volatility because 60 pips today can be very reasonable but in very low volatility in Apr-June 2007 then that becomes to much--but I will explain this in further steps how one is solving this--now I just want to stick to the principle of my "safe martingale")

If we look to the chart we will see many signals. Many of those signals will not be corrrect to trade from because the stoploss would be to much (More then -20pips so they do not come into play and have no influence on the result). We see also many signals that that can be traded and that will indeed trigger their stoploss. But we can also accept that many signals don't hit the stoploss and that move more then 60pips in profit.

Lets for a moment go back to your Micro lot account and lets have a look to what would happen if we would have a losing streak of 15 consecutive losses: If you start off with 0.01 lots (1 Micro lot) then after 15 consecutive you would be sitting on 0.81lots and the 16th trade you need to take in 1.08lots and your total loss would be on that moment ONLY -64$ . Your 16th trade you hit the T/P. That means that you recovered your loss completaly and that you added a little profit on top of it.

Even if we push it in an extreme way: Lets say that we have 20 consecutive losses. Then after 20 losses we are sitting on 3.42Lots with a total loss of ONLY -272$!! Trade number 21 is taken with 4.56 lots and recovers all the losses and adds a little profit on top of it. Remember also that we will have severall losses with bigger contract sizes but with S/L's that are smaller then -20pips, because the stoploss value is -20 pips or smaller to take in a trade (meaning that the 19th trade could have had a stoploss of only-15pips). So this last trade will not only cover the theoretical losses of -20 pips but is going to cover the practical losses that are smaller and this way give more overall profit then the theoretical overall profit.

Important is that you can also clearly see that only arrows apear according "the trend". If trend is "so called up" then only long arrows apear. When trend is "so called down" then only short arrows apear. And not taking in shorts and longs all randomly whatever price is doing.

Also important is to realize that when you see a group of arrows (lets say all long arrows) and that the 60pip T/P was not hit then then NOT all of the arrows are losing trades. Because it is possible that the 1st arrow is a correct entry then severall other arrows followed but the 1st stoploss was not hit so it stays in its first trade till that 1st stoploss would be hit. So all other arrows that followed were no more entrys to take.

Interesting element that I would like to add to all of this: If one does a backtest and sees that he could have on a very regular base 10 or more consecutive losing trades but lets say with a maximum of 15 consecutive losses, then one can build the system is such a way that the system only starts to trade after it has 10 theoretical losses. So the system would not have the first 10 initial losses. The system would of course trade a lot less but the risk would again be smaller.

The system could also have a time filter so that it only takes in trades during UK and US session to make sure there is a bigger possibility to reach 60pip profit targets then to take in trades during the assian session that have not such a big potential of moving 60pips in a certain direction.

Up till now do you agree on this ?...if no we go deeper into this otherwise we move over to the next step.

Looking forward to your answer.

Friendly regards... FXiGoR

PS. Spreadsheet is at the bottom of this posting (under the images)

Files:
taz_1891  

Thanks IGor for starting the thread.

My 2 cents....

The main problem i have with martingale strategies and alike is not the lot sizes people use to average down. Mathematically, it is proven averaging down is NOT the way to go whatsoever (infact mathematically, averaging up is). It is the re-entries after a losing trade is noticed or situated.

The re-entries are what kill most strategies cause there is VERY little thought that goes into it. Most of the time people just use random pip distances thinking this is optimal. Obviously it isn't... not even close.

Have mentioned it a few times (ofcourse on deaf ears) more time and effort should be applied to work on re-entry methods. Be that strong support resistance/psych levels/MA cross... bla bla bla whatever. The list is endless.

My point is if people treated re-entries more seriously rather than just thinking "oh its got to turn around soon - doesn't it" there EA's might last longer.

iGoR  
taz_1891:

.............

My point is if people treated re-entries more seriously rather than just thinking "oh its got to turn around soon - doesn't it" there EA's might last longer..........

Hi Taz,

Thanks for your posting.

Personally I would never trade systems that are speculating on the element that it will turn around or that price will or should turn around.

If you look to the images I posted then you can clearly see that entrys are taken in based on signals that are according "a trend" and each trade needs always to have a clear acceptable stoploss.

But to tell you something that suprised me very much is the following: A couple of months ago I had a long talk with a lawyer that makes the contracts between investors and institutional traders and investment banks.

In those contracts he needs to stipulate very precisly what and how the investment bank or institiute is going to trade. So even for a lawyer he is very much awere how the institutional traders are trading.

It was nearly a shock to hear from him that for the institutionals it is absolutly normal to use "avering down" techniques. Meaning if they buy or sell something and later in time they have the possibility to get that same instrument they traded, for a cheaper price they will do so. and they do that severall times on that same instrument. So a pure averaging down. Unless price suddenly drops because of bad news and that can lead to a total melt down ( for example bad news that can lead to a company to go bankrupt).

Personally I would not work this way. I would prefer to close a certain position and then re-enter with a higher volume if a re-entry signal is given but not adding postions purely on price dropping by a certain %

Friendly regards....FxiGoR

taz_1891  

Interesting...

I personally don't mind averaging down. I was specifically mentioning that statistically it is a mathematical no no. If you are median trading, then it does make sense. Buying cheap expecting the market to return to a 'centralised' figure.

But still, how do you make it safe? If you have a runner, then you are kind of screwed.

There is a brilliant man on the Forexfactory forum..... 'Nanningbob'.

He has a trading scenario which can help with corrections without martingaling into oblivion. His 'recovery' strategy is call the 2.4.2.

I won't go into it here (best to read it yourself) but it is interesting to say the least.

As we all know, you can't negate the effect of a runaway trade when your martingaling/averaging down. If you continue to build upon a losing situation you will lose your pants eventually.

So how do you get around it? Where/when exactly do you take the loss or know when to pursue the recovery..... or both.

I see what your saying with Risk/Reward and i do agree with it. With averaging down and removing a 'smallish' SL you can greatly improve your odds. It's just that there must be a point where you have that runaway trade and you know you should stop without losing your pants. If you already had a built up opposing position then that can save you but how/when to place the trades. And ofcourse it is hedging (US brokers wont allow it).

Sorry, starting to dribble. Just typing and thinking at the same time. Anyways, lets continue this discussion....

Anyone..... bueller.... bueller??

iGoR  
taz_1891:
................. I see what your saying with Risk/Reward and i do agree with it. With averaging down and removing a 'smallish' SL you can greatly improve your odds. It's just that there must be a point where you have that runaway trade and you know you should stop without losing your pants. ......

Hi Taz,

Read this posting carefully again https://www.mql5.com/en/forum/180202/page3

In the example I explain there is no runaway trade or trades that you could loos your pants.

In my example each trade has a stoploss of -20pips or less.

Even in the spreadsheets I posted it is very clear to see. S/L's of -20 pips and T/P's of 60pips.

Nothing scarry or out of the ordanary. Just a plain trend following strategy.

Friendly regards....iGoR