Flipping a system around | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Can you make money by simply flipping a coin? 15 replies

Flipping directions but adding more 2 replies

Sentiment news trade system (A.K.A coin flipping) 3 replies

Another flipping Newbie 2 replies

Late Night Channel Flipping ... and then some. 1 reply

Attachments: Flipping a system around
Exit Attachments

Flipping a system around

  • #1
  • First Post: Sep 24, 2019 2:31pm Sep 24, 2019 2:31pm
Hello everyone!

I have been gone from this forum for quite some years. Very busy with work and at the same time I needed to take a break from this quest or else risk getting burned out and abandoning it completely.

As I get back into the swing of things, I have realized that the only immediate way that I can use in order to take my personal emotion out of trading is thru automated trading. This may sound crazy to some but I am just going to give it a go.

The subject of this thread is not to discuss a system in and of itself, but rather the aspects of creating one. Particularly, one related to an old article I read way back in the day. In it, some random contributor commented on the options available to anyone looking for a successful system. He wrote: "If you have a system and it turns out that when backtesting it, it is a complete and steady looser, then simply turn it on its back. When it gives a sell signal, buy and when it gives you a buy signal, sell"

Sounds simple at first but I have tried that over the last couple of weeks and I don't seem to come up with that same result. Provided again that we are talking about a fully automated system.

I know that without the logic behind it, not much can be said about such a system. I am here proposing we discussed, not the system itself, but the ins and out of the main proposition, turning a system around with the same signals, but in opposite direction. I'll start with an example

Spreads: When you test a system and it happens to be a steady looser, it is easy to imagine flipping the signals. However, for every trade you make, there is a loss incurred on the spread, that adds to the apparent ability of the system to lose money. When you flip the signals around, that spread still runs against you.

What other things play a role, in your opinion?

Thanks
Please pull my head out of my ass, I am running out of oxygen...
  • #2
  • Edited 2:53pm Sep 25, 2019 1:23pm | Edited 2:53pm
Post Removed Cheers !
  • #3
  • Sep 25, 2019 2:21pm Sep 25, 2019 2:21pm
Jeez! what a rant!. I am going to take a leap, a chance here. Not sure how long you have been doing this. I am just going to base my assessment from your FF profile that says you joined just about a year ago. I am not sure you understand what I meant and based on the philosophical approach you describe, I am going to assume you are going thru some platonic trading phase here.

I hope you take no offense. When you present the idea that I may be qualifying a system as "good" or "bad" on 10 trades, then you are kind of implying the above but about me. So let's call ourselves even and start over.

Most of what you are describing comes once you have a "system". I am specifically referring to the process of creating one. At the end, whether you use the same old indicators, trend lines, flipping a coin, follow a weather pattern or your bowel movements... none of that is good unless it translates into something that ultimately triggers you to place a buy or a sell on your trading platform. There is no other order type after execution.

So, if after placing thousands of trades over a number of years, you can't see anything but a dropping balance/equity curve on your testing account, could it not easily be imagined that if you revert your buys and sells, then your equity curve would reverse as well?

That is what I am trying to start this discussion on.
Please pull my head out of my ass, I am running out of oxygen...
  • #4
  • Sep 25, 2019 2:44pm Sep 25, 2019 2:44pm
1-my profile not reflect how many years i am trading ... but i am trading since long long time ago ..

2-somehow i expected that reply exactly from you.. i expected what you said about my post not related to what you talking about .... so that's my point of view if you read what i said carefully you will know i talking about the same topic ..
Quote
the process of creating one
.

3- are you thinking you only one in the world thought in the same? everyone thought in the same thing ...

4-reverse buys and sells mean reverse a lot of other aspects you can't ignore ..

5- you are professional trader ,and trade since years then you know very well what other aspects that i mean if you want reverse buy and sell signals that your system give.

Good Luck ... i am not here to argue about something i already made enough experiments on it ..
and if you thinking what i said is useless and not related ... ok ignore my contribution ... i deleted it
  • #5
  • Sep 25, 2019 4:36pm Sep 25, 2019 4:36pm
Good Luck ... i am not here to argue about something i already made enough experiments on it ..
and if you thinking what i said is useless and not related ... ok ignore my contribution ... i deleted it

Thanks! Glad to hear!
Please pull my head out of my ass, I am running out of oxygen...
  • #6
  • Sep 25, 2019 7:24pm Sep 25, 2019 7:24pm
This simply cannot be done. And the reason is also simple - 99.99% of all "losing" systems are not really losing systems from the markets perspective. In other words they don't have inverted negative expectation. Their actual long term expectation is exactly zero! And the loss comes from the trading cost. (spreads+commissions+swaps+ slippage). All of this is because the systems are in fact random. And the market is mostly random as well. So random+random equals what??? RANDOM again. You can't get order out of chaos when you are not the one in control.

A real losing system is a system with real negative expectancy which must be at least 3 times the size of the cost! Therefore a real losing system and a real winning system is pretty much the same thing. You just have to create exact mirror image of the rules and you have a winning system. Unfortunately these real systems are extremely difficult to find. So difficult that most likely they don't even exist in practical terms. Or it's close to impossible to find them.

What you see in the public space is random results. The winners think they are smart and the losers think they have bad luck or not enough knowledge/skill.
While in fact both (winners and losers) are for the most part simply fooled by randomness.

Here with my curve simulator you can see exactly what is happening:
1500 trades per per curve, 100 curves per simulation.

1. Simulation of 55% win probability RR 0.8 and fixed risk 10:
Attached Image (click to enlarge)
Click to Enlarge Name: W55 RR 0.8 fix 10.png Size: 350 KB


2. Simulation of 55% win probability RR 0.8 with risk 1%.
Attached Image (click to enlarge)
Click to Enlarge Name: W55 RR 0.8 R%1.png Size: 291 KB


3. Simulation of 55% win probability RR 0.8 with risk 1% + random additional multiplier ( In order to simulate bad and good luck)
Attached Image (click to enlarge)
Click to Enlarge Name: W55 RR 0.8 R%1 + random luck.png Size: 163 KB


4. Same as 3. but with risk 2%
Attached Image (click to enlarge)
Click to Enlarge Name: W55 RR 0.8 R%2 + random luck.png Size: 87 KB


Why 55% with RR 0.8?? Because this is close to the typical result of the average retail trader. For example looking for a 100 pips profit on 1 pip spread.... this is what you get on average.

So what do we learn?
We learn that:

All of these simulations have negative expectation.
All of these simulations have positive curves! (Winners)
And as you increase the risk by adding additional random components ("Luck") you get larger and larger winners, sometimes extremely large winners, but they represent smaller and smaller % of the whole sample.

Now, when you understand all of this, you can change the input parameters of the simulator in any way possible, and you can create any equity curve.
And then you realize that everything related to trading the financial markets as a way to make a living is one big bullshit, and humans in general are not very smart.
  • #7
  • Sep 25, 2019 7:42pm Sep 25, 2019 7:42pm
if your expecting one system to win 24 hours a day every day... thats hard. Conditions have to be right for most systems to work. Thats where watching the market for years comes into play. Don't try and master everything.. Just find the one thing your good at. I do a lot of my trades at certain times of the day, that I've had the best luck with.. I avoid other times of day that tend to be too slow or too wild. I have to do a lot of waiting for conditions to be right.
The only system that will work is one designed by and for yourself.
  • #8
  • Sep 26, 2019 3:09am Sep 26, 2019 3:09am
Quoting bloodpoodle
if your expecting one system to win 24 hours a day every day... thats hard. Conditions have to be right for most systems to work. Thats where watching the market for years comes into play. Don't try and master everything.. Just find the one thing your good at. I do a lot of my trades at certain times of the day, that I've had the best luck with.. I avoid other times of day that tend to be too slow or too wild. I have to do a lot of waiting for conditions to be right.
Very true. If my systems trade more than 40-50% of the days I lose money. The reason behind this is that the random nature of the markets is far bigger than most people think.
  • #9
  • Sep 26, 2019 4:05am Sep 26, 2019 4:05am
Quoting alphaomega
This simply cannot be done. And the reason is also simple - 99.99% of all "losing" systems are not really losing systems from the markets perspective. In other words they don't have inverted negative expectation. Their actual long term expectation is exactly zero! And the loss comes from the trading cost. (spreads+commissions+swaps+ slippage). All of this is because the systems are in fact random. And the market is mostly random as well. So random+random equals what??? RANDOM again. You can't get order out of chaos when you are not the one in control....

great stuff..
Have one request:
a. Simulation of 20% win probability RR 4,5 and risk of 0,25% per trade. Number of trades 1500

I appreciate
  • #10
  • Sep 26, 2019 5:42am Sep 26, 2019 5:42am
Quoting 21vs7
{quote} great stuff.. Have one request: a. Simulation of 20% win probability RR 4,5 and risk of 0,25% per trade. Number of trades 1500 I appreciate
Positive expectancy but not very good. The win% is very small, so even with such minimal risk per trade you can still expect to see over 40% drawdown in the long run. And this is without adding the possibility of good or bad luck.
Attached Image (click to enlarge)
Click to Enlarge Name: win 20% rr 4.5 risk 0.25%.png Size: 359 KB

Here you can make experiments with my simulator.
Attached File(s)
File Type: ex4 Curve Simulator.ex4   21 KB | 645 downloads

Just keep in mind that if you try some aggressive settings the simulation will spit out some crazy large numbers and MT4 will crash because the charts cannot display such large numbers or you simply run out of memory.
  • #11
  • Sep 26, 2019 6:52am Sep 26, 2019 6:52am
Quoting alphaomega
{quote} Positive expectancy but not very good. The win% is very small, so even with such minimal risk per trade you can still expect to see over 40% drawdown in the long run. And this is without adding the possibility of good or bad luck. {image} Here you can make experiments with my simulator. {file} Just keep in mind that if you try some aggressive settings the simulation will spit out some crazy large numbers and MT4 will crash because the charts cannot display such large numbers or you simply run out of memory.
Thanks alphaomega!
  • #12
  • Sep 30, 2019 4:10am Sep 30, 2019 4:10am
Quoting alphaomega
{quote} Positive expectancy but not very good. The win% is very small, so even with such minimal risk per trade you can still expect to see over 40% drawdown in the long run. And this is without adding the possibility of good or bad luck. {image} Here you can make experiments with my simulator. {file} Just keep in mind that if you try some aggressive settings the simulation will spit out some crazy large numbers and MT4 will crash because the charts cannot display such large numbers or you simply run out of memory.

Hey a,

cant make it work..doing only little simulation, but still I get a freeze.

Have a bit larger request..in your own time..if you have a chance please..

1. 4,9 RR; 36% win rate; 1500 trades..risk 0,25%; 100 curves.
2. 12,5RR; 26% win rate: 1500 trades; risk 0,25%; 100 curves
3. 5,4RR; 38,5% win rate; 1500 trades; risk 0,25%; 100 curves.
4. 14,9RR; 28% win rate; 1500 trades; risk 0,25%; 100 curves.


I really appriceate the help!
  • #13
  • Sep 30, 2019 5:58am Sep 30, 2019 5:58am
Quoting 21vs7
{quote} Hey a, cant make it work..doing only little simulation, but still I get a freeze. Have a bit larger request..in your own time..if you have a chance please.. 1. 4,9 RR; 36% win rate; 1500 trades..risk 0,25%; 100 curves. 2. 12,5RR; 26% win rate: 1500 trades; risk 0,25%; 100 curves 3. 5,4RR; 38,5% win rate; 1500 trades; risk 0,25%; 100 curves. 4. 14,9RR; 28% win rate; 1500 trades; risk 0,25%; 100 curves. I really appriceate the help!
Check the settings. If you set the risk as % of balance then you should also adjust the risk per trade. Because by default it is set at 10. With 1500 trades per curve you will get huge numbers due to exponential growth. As I said before, the charts in MT4 cannot handle such large numbers.
Also, if your computer is older or not very powerful then simply try to do simulations with smaller number of curves and trades.
  • #14
  • Sep 30, 2019 7:54am Sep 30, 2019 7:54am
Quoting alphaomega
{quote} Check the settings. If you set the risk as % of balance then you should also adjust the risk per trade. Because by default it is set at 10. With 1500 trades per curve you will get huge numbers due to exponential growth. As I said before, the charts in MT4 cannot handle such large numbers. Also, if your computer is older or not very powerful then simply try to do simulations with smaller number of curves and trades.
trying different settings. dont get any lines. could be MT4
thanks anyway
  • #15
  • Sep 30, 2019 10:03am Sep 30, 2019 10:03am
Quoting Drolph
{quote} The moment when you realize, that financial markets cannot be sufficiently reproduced by basic Monte Carlo simulations.. Quotes are not random in the way you are simulating them in your simple calculations, but driven by expectations. There is a difference between the random development of a price based on a single timestamp or of the one beeing in relation to other markets/recent developments. It cannot be summarized to simple math like that. Sounds like a small influencing parameter, but it makes your whole approach showing an artificial...
Almost everything in the physical world can be artificially simulated. Of course some things are extremely difficult and require massive computing resources but it's not impossible.

But the price action of the markets - this in fact is very very simple to simulate. It's quite easy to create random charts, (or tick history) to be completely indistinguishable from real price history.

You can simulate every single aspect of the price action. The cycles of the volatility, the gaps, the news events, and of course the occasional black swans.
For each of the different aspects you just have add another layer of random variables, some of which have to have limited ranges, upper and lower boundaries....and so on. You can take the exact scope of the parameters from the real historical data to achieve perfect realism. And all of this can can be done even automatically with simple machine learning.

If you think about it there is nothing complicated in the price action. It' just a random string of numbers. +1, -1. tick up, tick down....there is no mystery.
  • #16
  • Oct 1, 2019 1:03am Oct 1, 2019 1:03am
Quoting alphaomega
{quote} Almost everything in the physical world can be artificially simulated. Of course some things are extremely difficult and require massive computing resources but it's not impossible. But the price action of the markets - this in fact is very very simple to simulate. It's quite easy to create random charts, (or tick history) to be completely indistinguishable from real price history. You can simulate every single aspect of the price action. The cycles of the volatility, the gaps, the news events, and of course the occasional black swans. For...
You did not get my point. I have got a diploma in natural science and I am pretty used to numerical calculations. There is no physical background you can adress with equations like a weather forecast for example.

Price action is not random in the way you are taking it, it is driven by expectations. So having a Monte Carlo simulation with a simple range of risk/reward and TP and SL variations thrown on a random Tick development is not reproducing financial markets sufficiently, as I said before.
  • #17
  • Oct 5, 2019 11:34pm Oct 5, 2019 11:34pm
Quoting alphaomega
So what do we learn? We learn that: All of these simulations have negative expectation. All of these simulations have positive curves! (Winners) And as you increase the risk by adding additional random components ("Luck") you get larger and larger winners, sometimes extremely large winners, but they represent smaller and smaller % of the whole sample. Now, when you understand all of this, you can change the input parameters of the simulator in any way possible, and you can create any equity curve. And then you realize that everything related to...
Thanks for pointing this out! RR imho is the biggest BS. People tell you to use a good RR, e.g. risking 1 bucks for a 5 buck return. But they don't realize that changing the RR will change your winning %, too. In the case of RR = 1:5 and under the assumption that market is random you'll only win 20 % of your trades.
  • #18
  • Last Post: Oct 16, 2019 5:50am Oct 16, 2019 5:50am
Can anyone here prove (empirically or theoretically) that the market is not random on the long run ?
Because if it is random, if you reverse random you'll get ..."another random"

If you flip the shit, you won't get gold. You'll get just another shit !
Member
0 traders viewing now