ATC.Experience, knowledge and practice. - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
For testing take x2 x3 spreads from a typical intraday, tp at least x10 spreads. Changes within 20% of key input parameters should not affect profitability result by more than 5%. If sl>tp, make a grid in the size of the resulting sl.
I've read your thread about martingale and I don't understand what the problem is with using elements of martingale in a strategy. We are here to make money, not to defend our principles. If the martingale = big risks and loss of securities, there has to be some criteria for these risks. So I return to my question - what conditions must be fulfilled for the bot to be considered normal and acceptable for using it for real money? How "honestly" should the backtest be performed in order to meet these criteria?
In general, what is the framework to fit in, or there is no such framework, and the whole discussion boils down to smirks and sarcasm (this is not about you).
But since forex is a more flat market than gsb, martin works better on it.
If martingale = big risks and losing a deposit, then there must be some criteria for these risks.
So I return to my question - what conditions must be fulfilled for the bot to be considered normal and acceptable for use on real money? How should the backtest be performed "honestly" in order to meet these criteria?
on the gsb (random walk chart), martin does not offer any advantage.
but since forex is a more flat market than gsb, martin works better on it.
the same as in other systems ;)
i need it to show a normal result on the backtest and on the forward (on a different period)
My strategy trades on indicators on small timeframe, but if the price does not go in the right direction the bot continues to open positions against the trend (martin). It is effective until the trend is strong, when the trend starts, the bot, so as not to become a hostage of martingale, leaves losing positions and continues to trade as usual, ignoring the hanging positions, the small foref allows him to catch waves even in a trend. Then, either he manages to close the hovering positions earlier or later in the bubble, or he cuts them to the minus. The result is visible on the profit graph above, the peculiarity of the strategy is that the bot is always in the drawdown of 100-300 quid (green line lagging behind the blue) - this abandoned positions create such a background. The scheme is working, but not 100% universal yet, on some pairs there is no profit or small loss.
There are no stops and takeoffs in the order, the bot does everything "by hand" according to the situation.
It is not possible in manual trading either. Drawdowns are inevitable. And a guaranteed profit is a fantasy. Although some people are lucky for life. If you are lucky you will earn, if you are unlucky you will lose. You can come up with your own ideas. They haven't learned how to influence luck.
Apart from the strategy, the success of the broker's reliability, the bank account, the glitches of the terminal, the broker's failure. It all really affects it all.
Even a profitable strategy will not guarantee earnings.
You have an opportunity - trade. No, work and trade. And then you get old. You can only get rich by cheating others - publish signals, or take them for trust.
To begin with go away from Forex. On the stock market. I understand you are still at the beginning of the path. All have experience.
Two points are important to understand here:
1) By definition, any "homemade" robots developed and/or potentially developed by a single person at home, despite all personal obsession to find a robo-grale, in reality are hundreds of times weaker than automated or semi-automated trading systems of large trading funds and investment houses (GoldmanSachs, BlackRock etc.), where hundreds of creative mathematicians, strategists and programmers are working; Therefore, in a competition between robots "crowd" algodtraders will always lose (the results of their robo-strategies are approximated and calculated by neural networks, just like the behavior of gambling traders with autonomous decision-making),
2) Artisanal robots are many times weaker than a human brain trained to solve dynamic computational problems generated by news and trades of trading funds.
This leads to one simple conclusion: there is no sense for a single trader to spend time on algorithmic junk trading, but only to train their own "neuronet" to solve manual trading problems (+ to develop drivers/scanners and other support algorithms in MQL5 or another platform). This is the way of gladiators in the financial market, practicing and fighting for money.
About me: came to this conclusion last summer, after 5 years of algo research, and now I'm fully focused on training my brain + developing back-office systems (market scans, risk management, etc.). No WOW results so far, but I'm going this route of learning, coaching-> trading, figuring out my blunders, and trading again. The only downside so far is that I have had to give up alcohol completely.
- A bot that works at 1H or higher yields 2-3 trades a week, it's bullshit, it looks nice on the backtest, it's unrealistic to try it out on the real market))) you have to wait a year...
- I can't give up martingale completely, some of its elements are still needed,
The results of my TC league refute this statement.
I have a lot of TS that have been working for a long time (some have been working for more than a year) on the demo, and show normal results on the demo.
Here, ran mine on EURUSD from 2018.01.19 until today, start 1000, final 4200, entry 2 quid, leverage 500, open up to 50 positions, depo load within 5%, about 10 trades a day
You should put it on demo. And see what happens. Then tell me about the profits.
There are two points that are important to understand here:
1) By definition, any "homemade" robots developed and/or potentially developed by individuals in a cottage industry, despite all the personal obsession to find a robo-grial, in reality are hundreds of times weaker than automated or semi-automated trading systems of major trading funds and investment houses (GoldmanSachs, BlackRock etc.) employing hundreds of creative mathematicians, strategists and programmers; Therefore, in a competition between robots the "crowd" of algodtraders will always lose (the results of their robot strategies are just as approximated and calculated by neural networks, as the behavior of gambling traders with independent decision-making),
I disagree.
Again, the experience of TC League tells me that the simplest systems can earn as much money as the "super-duper systems". In fact, my introduction to trading started with writing a robot based on this super-system of hundreds of rules. It showed good results on history for 15 years. When I tried it on real account it seemed tostart earning and three months later I lost all my gains, like the simplest TS. The question arises - why use sophisticated systems if they work the same way as the easiest? Wouldn't it be better to concentrate on selecting simple systems that work in the same way rather than creating a super-duper system?
I disagree.
Again, the experience of the TC League tells me that the simplest systems may well earn as well as the "super-duper systems". In fact, my acquaintance with trading started with writing a robot using this super-system of hundreds of rules. It showed good results on history for 15 years. When I tried it on real account it seemed to start earning and three months later I lost all my gains, like the simplest TS. The question arises - why use sophisticated systems if they work the same way as the easiest? Wouldn't it be better to concentrate on selecting simple systems that work in the same way rather than creating a super-duper system?
What's the point of spending time on simple ones if, as you say, they work just as badly? It's better to spend time on the maths then.
What's the point of wasting time on simple ones if, as you say, they work just as badly? It's better to spend time on the math.
Yes, because less time is spent on simple ones, and there are a lot of simple ones to choose from. I've got over 600 TCs working at the same time, so there's plenty to choose from. And I can assert, not without proof, that simple systems are quite capable of earning profits.
If you spend time on the basics, and then you get a loss after a week of work, what's the point of studying the basics?