Why martingale EAs look profitable… until they aren’t (especially on gold) - page 12

New comment
 
Thomas Olof Gardling #:

Any trade taken merely with Moving Average, RSI o any other standard forex indicator isn't thinking outside the box. Those indicators are more or less only the average price of N bars.


Any indicator measures what price did in the past and it doesn't predict anything. Traders should know that no matter what they use the price direction cannot be predicted with much over the probability of a coin toss. So when martingale is involved the only thing you can control is thrown off the window: a constant risk per trade. The probability of a trade to succeed does not change but the potential loss is higher.

 
Daniel-gheorghe Muresan #:

Any indicator measures what price did in the past and it doesn't predict anything. Traders should know that no matter what they use the price direction cannot be predicted with much over the probability of a coin toss. So when martingale is involved the only thing you can control is thrown off the window: a constant risk per trade. The probability of a trade to succeed does not change but the potential loss is higher.

You are talking about lagging forex indicators, they will never be better than 50/50 in the long run.


But then there are leading indicators from outside the forex realm.

You just need to figure out what moves the currency/commodity you are trading and where to get that information from.

 
Andrea Michelon's 72-variant test is basically my last two years in one table. I ran three different grids on gold, all died between February and April. The one that hurt was a $500 account that survived 11 weeks and then lost 85% in a single NY session when spreads blew from 22 to 65 cents around CPI. I disagree with Daniel-gheorghe's line that indicators can never beat 50/50, but he's right that martingale throws the one controllable thing, constant risk, out the window. That's the real killer.
 
Bruce Fx #:
I disagree with Daniel-gheorghe's line that indicators can never beat 50/50

Then show me a lagging forex indicator that gives better results (over time) than 50/50, like a crossover of 2 MA lines or RSI high/low values.

The thing is, its easy in hindsight to see what should have been the right settings in the indicators for just that historical chart.

The challenging is to make it work in the future with the historical settings.


Bruce Fx #:
martingale throws the one controllable thing, constant risk, out the window. That's the real killer.


Yes, if you are referring to the original martingale which doubles the next trade after a loss until it gains the previous loss/losses.

Its the strategy behind it that matters. If you don't have a fix or something that gives you better odds, it will not matter what you throw at it. 


You can have the best controllable risk settings in a bad strategy and it will only give you a slow death instead.

 
There's NO "leading indicator".

And if people thinks indicator signal is at the cross only etc, they don't know what they think they now.
One of the main purpose of the indicator is, it'll simplyfy the chart's data so we can analyze and see patterns on it. And that's actually how to trade indicators, and cross is just one of them.
Some people thinks indicators are lagging, price action is faster, no, they're more or less the same. Indicators are based on chart's data too.

All indicators and price actions are lagging in terms of everything are based on past data.
Nobody know the future.
All we do is finding proftable patterns that happened many time in the past, and this lead us to believe if next time these patterns appear again, it'll likely to profit.
There's no indicator that will lead the market. It's the fundamental economics and mass psychology that will lead the market.
 
Erman Ardianto #:
There's NO "leading indicator".
That's debatable. There are indicators that attempt to predict future price movements and, inherently, throw many "false signals." The most common one is the Stochastic Oscillator (5, 3, 3).
 
Erman Ardianto #:
There's NO "leading indicator".


Of course there are, every time Trump opens his mouth there are people who know what he is going to say before hand. 

Banks sell information for large amounts to hedge funds and people with deep pockets too. 


Here is the problem with information, the large players cant just put a couple of hundred million into the market without getting noticed.

If you can some how follow these big players and trade behind them, that's a leading indicator.


But you are right in the sense that there are no leading indicators in forex or anywhere in the realm of retail trading.

 
Thomas, fair challenge in #114 and honestly I can't show you one either. I trade gold manually and I've never found an indicator that predicts direction better than a coin flip. What changed my results was admitting that. I stopped hunting entries and started obsessing over exits. These days I risk 0.5% per trade on gold, max two trades in a London session, and my win rate is barely above 50% but the account actually grows. The edge was never the crossover.
 

One thing that hides the risk in gold grid/martingale backtests: commission modelling.


I reverse-engineered a public gold grid account and rebuilt it as an EA. In my first tick-data tests the commission was fixed per fill, so every added grid level quietly cost more than the backtest showed. The equity curve looked smooth until the floating loss of the stacked positions met a long one-directional move.


If you test these, check (1) commission per lot vs per fill, (2) swap on positions held for days, and (3) the worst adverse run in the data, not the average.

1...56789101112
New comment