I Am What I Am

Filed Under (Journal) by Casey on 29-04-2011

Dear Erik

I thought that I would dedicate a post to your comment on my previous post. Since you are my most avid commenter with regards to my recent event and also your comments did spark a much deeper thought into my Martingale escapade. And probably talking about it might give me a clear mind of who I am and what I am doing.

Anyway, this is what you commented

Casey,

Instead of just “guessing” how to run strategies that work you should study statistics man, learn how to perform Monte Carlo simulations and learn how to study systems for real. I would hate to see you get all wiped but unless you get a grip and truly learn the stuff you are on the way for that. Don’t be stubborn, take advice, study statistics.

Erik

——————————-

Casey,

Backtests without an out of sample period are useless (you’re just curve fitting) not to say about the simulation capabilities of MT4 which are very poor for a system on a 5 minute time frame. Once you do truly reliable backtests accounting for broker dependency, spread widening and slippage and you also do Monte carlo simulations you will notice that this is a sure wipeout in the long term. Get your statistics together Chris, don’t ignore the shortcomings of your simulations. I hope you are humble enough to take this advice!

Erik

So… the first thing that came to my mind was “What is Monte Carlo simulation?” In your previous comment you did mention about that simulation, but I did not understand so I has assume that you were mentioning about backtesting. I mean… this was the first time I here this kind of testing on forex.

Ok… granted my forex knowledge is not as vast as the internet… So I went to do some research and this is the first site that came out when I google “Monte Carlo Simulation” was of course Wikipedia…

Wikipedia says that this type of simulation involves “….are a class of computational algorithms that rely on repeated random sampling to compute their results. Monte Carlo methods are often used in simulating physical and mathematical systems. These methods are most suited to calculation by a computer and tend to be used when it is infeasible to compute an exact result with a deterministic algorithm.[1] This method is also used to complement the theoretical derivations.”

To be honest Erik, I like fell asleep at …random sampling…

Here is another definition by Investopedia here…. http://www.investopedia.com/terms/m/montecarlosimulation.asp. “A problem solving technique used to approximate the probability of certain outcomes by running multiple trial runs, called simulations, using random variables.”

And another definition… http://www.palisade.com/risk/monte_carlo_simulation.asp

I did more google-ing to see what Monte Carlo Simulation Trading System would yield and viola…

http://www.forexfactory.com/showthread.php?t=17771

http://www.stocktradersbulletin.com/documents/articles.php?entry_id=1244494832

http://www.elitetrader.com/vb/showthread.php?threadid=209751

These are but a few information I got about Monte Carlo Simulation and I think so far the best is this site.

http://www.tradingsystemlab.com/files/Trading%20System%20Evaluation.pdf

Which leads me to this understanding. To use Monte Carlo simulation you need to backtest a system. Then pluck the statistics from the backtest and simulate it forward.

And boom… you will know if the system is profitable are not. Next… I try to google-ing Monte Carlo Simulation Software….

I gave up then… So… unless you can tell me what Monte Carlo is all about… I shall stick to my understanding as above. Which means to say… your simulation result depends on backtesting result. If backtesting result is good… your forward testing result is good.

Another words it is another simulation forward… and if by gawd… this is the best true model to simulate the future… then this is the holy mother of gawd strategy for forex.

Unfortunately it is not…because if backtesting is flawed in the first place like in your second comment then the future simulation is flawed.

So sorry Erik but I do not subscribe to your advice.

I have already thought about your second comment about backtesting. I myself have also wondered about the true aspect of backtesting. I have already thought about it since two years back… and thought that backtesting means $hit and what we want is forward testing…

Thus, now you see all forex ea sellers are putting their EA on live demo account.

Given an inch we would want a yard… So now we are saying live demo account means $hit because demo does not simulate real trade like spread widening or even slippage.

Then you have these ea sellers go for live account with real money.

We then say… this live account with real money is bull $hit because they are using a fu(ked up broker…

So now… we ask them to put their money in Alpari… or GoMarket… or FXCM… or one of the more reputable broker… and there is no complain from the market. Even then I know of some who are not happy with these sort of testing results.

I will admit this… My recent losses were due to my lack of backtesting… and again I mention in my last post… I was lazy and greedy.

So far *fingers crossed* the only really work I did with Megan was with Cyborg EA and it has yielded me 5% every month because of the backtesting exercise I did on all currency pair base on one single idea.

When I put in PipStrider which is another mistake it failed and pull my Megan account slightly under because of the same reason that I did not backtest PipStrider.

In any system there will be shortcomings. It’s how we deal with it when it comes or how we manage it when it does arrive. Monte Carlo is a risk analysis and tells you the possibility that this situation might happen. If it does… what are we going to do about it… If it doesn’t then it is happy times.

I work for a construction company and I do know about risk analysis and risk management. This is also one of the reason why I love forex.

Stubborn… I am. That is allows me to continue with forex.

Take advice… I do. Only if it makes sense to me.

Study statistics… I won’t. It just a bunch of numbers and graph which means nothing because there are always assumptions when we do statistics… unless I subscribe to the assumptions.

Backtesting… yes I need to re-evaluate my position with Cyborg. That is so far my success story…

Chris… I am not… I am Casey.

Humble… I maybe… I need a few more knock on the head to probably be more humble.

With Martingale, no one has beaten the forex using this strategy… probably because they maybe right. Martingale will be a loser in the long run. Or probably no one has tried it hard enough. Maybe I have not tried hard enough.Yeah… So I guess I am stubborn.

As I say… I can only try until I can not try any more.

With that… I vow again not to touch a martingale system unless I did backtesting on it like how I did with Cyborg.

I probably won’t look at Monte Carlo Simulation… but I will continue to look at backtesting.

Starting for square one? Oh well… at least I am starting again…

Oh and Erik… thanks a lot for you input. Appreciate it.

Regards
Casey

PS… Except for my GoMarket I am testing another Martingale idea live… without any backtesting. Because this is going to be a semi automatic exercise. If this idea is correct… I will then move to convert this into an EA… :)

Be Sociable, Share!

Comments:

There are (5) Comments for the I Am What I Am

Post a comment

Anti-Spam Protection by WP-SpamFree