Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
I fully agree that 16 trading day span is insufficient amount of data to make a judgement. I've just started thread and I can't show you tests and trades for years back or forward trades in just two posts. My plan was to record trades on chart for a short time back in time and most importantly, to show you how the system trades in forward tests.
Backtests for certain instruments and period can be done, of course.
I'm not sure why do you mean its not real money system. There is no expectation to do well, you can clearly see exact trades that would've been taken.
This is exact point of putting your trading models and ideas into indicator, you simply attach it to chart and you see if it works or don't.
If you believe that someone who has profitable system would be very successfully rich, it is completely wrong.
I'm sure you now trading is not just buying and selling market, but much more, especially discipline in risk management, keeping emotional control and sufficient capitalization. Anyone who believes that with 1000$ you can make millions is just delusional.
We have a bit of misunderstanding here. Case study means that I am trying to do is to share something that works, to see how people will react to it, will they believe it or not.
I'm not sure where you see convincing narrative to buy my system.
No one is being pushed into anything.
Only way I would commercialize my tools is by making every person that wishes to buy it, test it for themselves, for free.
I would never scam a single person.
Of course, according to FIO/Vendor relationship rules.
Thanks.
Can you help answer these questions from other members on NexusFi?
So, explain me, what is the point of your post? You post screenshot of profits.
If I came here to scam people and brag how much money I made, luxury cars, etc. I would do same as you. Put bunch of profits and voila.
It is actually a sad world, where people believe in ridiculous nonsense, but are so ignorant to most obvious stuff that is usually lying in front of their eyes.
Well, I really don't believe your system works or that it makes you money. But that is not the point of my post. If you think you have a great system and you really don't, that is not actually a problem. I only care whether you are promoting it to our members and whether you are attempting to generate an interest in their buying something, even if in the far future. It can be a wonderful system or a terrible system, but if you're trying to sell it you will need to do that somewhere else.
This is the entire point. If you are posting so that people will think you system would profit them and you are preparing them to buy it, you are in violation of our rules. If not, you are fine. Promotional intent or promotional language are assessed in the usual meanings of the terms: intent to get someone interested in buying. I am telling you that you cannot do that here, and you cannot do the preparation for it here either.
You are explaining nothing. You present charts to say, "See, it works, see how great it is," but there is nothing to discuss because it's a complete black box, the workings unavailable. The only point seems to be to push the system. This is close enough to being promotional to have prompted these warnings. If you want to continue to present the system's results and if you want to discuss them then you can certainly do so, but this is fairly close to a sales job, and if you stray into frank promotions, we will need to step in.
Bob.
When one door closes, another opens.
-- Cervantes, Don Quixote
1. Curve fitting - you have many automated strategies that show perfect equity curve. Then you start to trade it in real time and it falls apart like house of cards. What does this mean? They use genetic algorithm to test millions even billions of combination of parameters to find the perfect one to fit the data. This never works, and it will never work. Most of system backtests are done this way.
Model in the beginning is completely wrong and it can never be profitable in real trading.
2. Commissions - most people make backtests without any commissions included.
3. Martingale systems, gambling.
Not to mention hundreds of other software tricks, to manipulate H/L prices in order to make perfect equity line.
Thanks, you mention all true things, but the same issues also hold true for manual backtesting that you say you do instead.
One really bad thing about manual backtesting is the ability to cheat - to say "I would not have taken that trade" or "it was a tick from my target, so I would have exited."
It is true that a backtested system can be susceptible to various problems including data snooping biases. However, that doesn't mean that you skip the backtest. Backtesting is important because it is the only way to empirically test the efficacy of a trading system. Empirical means verifiable by observation or experience. I can give you a strategy and a data set to test it against, and anyone can verify the results of that system. Without such evidence how can a person come to an evidence based decision on whether or not the idea is worth pursuing further? Moving forward without that information would be like putting the cart before the horse.
A good backtest is just the first step in the testing and verification process. If you can show the system has some historical validity then you can go about trying to show that the results continue to hold into the future. Before that point though there's not really much point in discussing anything else.
Four years worth of backtesting against ES futures is a good start.
I agree with you, that is why I show my charts with indicators that I use to trade. If you apply the same rules, you should get similar or same results as in backtest. But trust me, there is hundreds more ways to cheat in automatic backtests.
I never said I didn't do any backtests.
What I've done is mining years of data from most popular instruments to find statistical frequencies that occur in the market.
This is years and years of data used. Later, using those data, indicators are constructed and fed with the statistically verified periods of the market. And even today, indicators are adapted with every new bar that market prints.
This is empirical evidence you are looking for, backtests are just tip of the iceberg.
Preprocessing the market and mining the data is what gives robustness to an indicator, if model is valid, it will work always. It will always have statistical value and through adaptiveness, it will always follow market movements.
To spit some numbers, my effective edge over the market is around ~21.5 percent.
It can go few points up or down, but using law of large numbers, it will always return to above percentage.
Time for backtest really depends on timeframe you are trading. For Daily, yes, 4 years should be good, but with intraday timeframes, few months is perfectly okay.
Here is the backtest for last 4 months using 15m TF. Same trades as charts I posted in previous posts.
True.
You can add commissions depending on your preferred broker, multiply it by number of trades, and get how much money will go into commissions.
Since this is not scalping system, commissions are not so important and will not affect system performance considerably.