Martingale debunked by Gary

[Jeuro]

I do not know about others, Jeuro, but what you have written is way too complex for me to understanding.

I would rather be an un-clever trader; Buy if it is going up, Sell if it is going down.

Khalid

good one Khalid ... ... actually forex is that simple...at any point on time , no
one know the price of the near future... not even the Banks. ... the complexity comes when we traders try to use complex and not forex related tools and analisis.

I guess I was too complex to explain the simple...bottom line.. principles for stocks are not the same then forex.


J.

[Jeuro]

I think you understand a few thing in a different way..

Today, the real economy in foreign exchange transactions is down to 2.5% and 97.5% is now speculative.
http://www.twnside.org.sg/title/nar-cn.htm
Sorry JB... the article is old, and in spite a few things are true, mostly is pure non-sense. International commerce has not decrease a bit is the last 10 years. Cheaper and more expedite transport, communications , trade agreements and hunger of China for raw materials has help for the increase. I would bet if you contact the author of the article, he will not corroborate what he wrote in 1997. While I was trading with Deutch Bank , the Foreing Exchange manager of the time had it very clear... and that was the very first he related to us a few trader that met with him in London...speculative exchange did not account more the 15% of the totals. I would believe that more then any article I find.. after all that Bank in spite that is out of retail, still transact a good portion (biggest as individual bank) that is exchanged nowdays.


You cannot discard all studies from stocks, there are a few differences like currency tend to range while stock have more trends. You just have to adapt and read what those great guys said about trading ranging markets. Stocks have a huge tendency to bubble up than crash, with currency is different so a buy and hold may work nicely on stocks while an averaging will never work on stocks but may work nicely on forex. In the personal I disregard all studies for stocks. ... but I agree on the last part... that is what I said...Martigale is crazy for anything including forex... but averaging is NOT too crazy in forex

Another thing, what we are doing here is betting on exchange rates using some papers named CFD. So if I buy CHF and I don't like the rate I cannot take my CHF, go in Switzerland and have fun, I'm forced to sell them to the same crim. Same with NYSE. Doing exchange with the real thing are taxed with heavy commissions

You may be working with CFD... but . are you really sure? ..most of trader like me deal in spot forex.
It is real money today.. not a futures contract. ... major brokers offer CFD but I do not
work with them.. like most of us.

if my accoun is in USD and I buy 50000 euros (0.50 lot) it means I own them. I do not take delivery of them but I do own then..( obviously I bought them on credit) Someone "sold" me 50000 "real euros" and my brokers "have" to settle that with the bank and the end of the day sending them the equivalent in USD. (well they don't do it with my mickey mouse volume but they do it with the total of the day transactions). I can hold those euros forever (if my credit line allows it) or sell them whatever I want. No one knows when I will do that.

I have not never try, but I could take delivery of them if I wanted to. (If I had the money to settle the debt I incurred to buy them.) .. .. I am guessing my account would be switch to euro base and I would withdraw the euros and do anything I want with them.

J.
.

[garyfritz]

However, to me stocks, bonds, commodities, or foreign exchange pairs, are the same: mere numbers I trade risking real money, which I must protect by controlling and managing risk.

I understand your point, Khalid, and to some extent I agree with you.

BUT not all markets behave the same. I can prove that quantitatively with e.g. a trendiness measurement. Some markets trend, and some markets mean-revert, and it may vary between timeframes. This is why, for example, it's a bad idea to apply trend-following systems to daily DJIA or SPY -- they mean-revert. (That's why "buy the dips" [and conversely, "sell the peaks"] works on daily bars.) Other markets trend pretty well, and you'll have trouble if you try to fade moves in those markets.

So you can't treat different markets as interchangeable anonymous numbers. You want to understand the behavior of the market/TF you're going to trade, or you might be swimming against the current!

[garyfritz]

Sounds like we're on the same page. I want to know how the bars and numbers behave so I know how to trade them. I don't care what those bars and numbers represent -- currency, stock, gold prices, hemlines, matters not.

[cuzgeorge]

Thanks Khalid, nice post.
will try it out.
cuzgeorge

[TheLocustFund]

The best (and only) way to trade martingale is to take a small sum of funds you can afford to lose, find a market with low volatility and run a mean reversion strategy and martingale at each standard deviation from the mean. The least riskiest form of mean reversion is to trade cointegrated pairs/baskets. (Note** cointegration and correlation are COMPLETELY different). Think of martingale as a lotto ticket. Martingale is a not a model.

[hedgeitall]

The best (and only) way to trade martingale is to take a small sum of funds you can afford to lose, find a market with low volatility and run a mean reversion strategy and martingale at each standard deviation from the mean. The least riskiest form of mean reversion is to trade cointegrated pairs/baskets. (Note** cointegration and correlation are COMPLETELY different). Think of martingale as a lotto ticket. Martingale is a not a model.

That is smart! With Cointegrated Pairs you are condensing the Range. You would not have to double as with the standard Martingale, just add small positions until it tilts your way.
You could use that as a recovery method also. It is better then trying to add positions of the same currency pair.

[Iceworld79]

First of all, I know my following statement may get fire on myself. But since this is an open forum and I have seen so many strong opinions against the Martingale, I think I should bring my 2 cents and understanding to the Martingale and hopefully bring more discussion.

Bottom of Line about the Martingale:
1. After all, it is just a sizing method. Every time you increase your holdings, you increase your risk. At same time, generally you are increasing your wining ratio.
2. It would NOT change the profitability or risk/reward ratio as discussed in the post 1. A losing strategy can not be magically turned into a wining one with simply adopting Martingale method.

so, where would the Martingale be useful?
My understanding is that, in terms of psychology, it can change an existing system to match your personality.
For me, I prefer continuous small winnings and if I got hit, fine, just like a car accident. Just make sure you don't get killed and you can recovery from the accident.

But before you use the Martingale, you have to FULLY understand the risk of using such system. A hardstop, in my opinion, is a must. You cannot keep wining and getting hit is healthy. Keep the loss within your tolerance.

The ground of trading is all about probability, nothing else (still my own opinion). We just need to get 50.0000001% of wining chance and we shall see green.

[garyfritz]

In principle I mostly agree with you. In practice I very definitely don't.

The Martingale does basically just shift the risk from one place to another. It takes a normal winning/losing trading strategy, and pushes the losses off into the future. Everything is wonderful until those losses come back home again, all at once. And you can't be sure you'll have time to recover between those losses. Given the vagaries of chance, you might have several of those "car accidents" in a row, possibly even right after starting trading the approach. So in that respect you hugely increase your risk.

It's like living a lavish lifestyle and spending more than you can afford, and borrowing the extra money from Mafia loansharks. It works great while it works, and you just hope they "only" break your legs when they come to collect. And you never know if they're going to be in a bad mood and do much worse than break your legs. That's no way to live, and it's no way to trade.

But I agree a Martingale approach can work. **IF** you thoroughly understand the behavior of your system, and **IF** you thoroughly understand the risk/reward profile of the Martingale you're using, and **IF** you are certain to size your positions so you can survive even a series of several "car accidents," then a Martingale might be a workable way to trade.

That series of "**IF**"s probably disqualifies 99.9% of the people who want to use Martingales.

And even if you're the 1 in 1000 who genuinely understands the dangers of this method, I claim it's still a bad approach.

I think "living the easy life" borrowed from Mafia loansharks is a very bad way to live, and I think "preferring continuous small winnings" borrowed from a Martingale is a very bad way to trade. I think you should "man up" and learn to accept and live with the actual performance of your system, and/or you should find or develop a system that matches your psychology better.

[Iceworld79]

well. I appreciate your inputs and I fully agree with most of them.
For my own practice, my hard stop is 20% of my initial balance, so it takes 5 continuous hits to wipe me out. I have done tick-by-tick backtesting on my own strategy (4 years, all cross-pairs) and worst case is 2 continuous hits. I also has other strategies running in parallel with the martingale one. So the risk is spread out and I think it should put me in the safe zone.

let's think about the general case of a wining strategy. Assuming you are using a martingale style, during the period of [T1, T2), you are lucky and win profit (P). Eventually, you are hit at the moment of T2 and The total amount of your loss is SL.

1. If P>SL, it is fine and you may keep using this system and eventually you will see green. Now, if we convert this strategy into a non-martingale one, since it does not change the gained profit during [T1, T2], the equility curve may be very bumpy and it is very likely that you terminate a wining strategy early.

2. If P<SL, no matter which strategy you choose, since you are under water, you may choose to terminate it anyway.

All my point is that the Martingale is not a definitely "No-No-Zone". If you FULLY understand the risk and manage it in a reasonable way, it can be a useful tool.

Every strategy/idea in your trading life has its own advantage/disadvantage. We should not just simply say NO to something because 99.9% person can not understand or master it. Instead, we should put a serious warning to Martingale with something like "ONLY applicable if you FULLY understand it, otherwise a disaster will come at your own cost".