50% win ratio, 1:1 sl tp, and random entries | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Coin tossing and random entries 44 replies

Random Entry Win Ratio! 11 replies

Pure Risk to reward trade System - 1 to 3- Totally random entries 178 replies

Expert Advisor with random entries 3 replies

Random entries discussion... 7 replies

Attachments: 50% win ratio, 1:1 sl tp, and random entries
Exit Attachments

50% win ratio, 1:1 sl tp, and random entries

  • #41
  • Jun 27, 2019 11:27pm Jun 27, 2019 11:27pm
Quoting Copernicus
{quote} Here is a 21 year test undertaken by Jez Liberty on a portfolio of currencies and futures using a random entry condition of flipping a coin for short or long with a trailing stop condition and occasional re-balancing. Pretty good overall results. The system first “tosses a coin” to decide whether to go long or short the market. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. That entry-stop distance is used to calculate the position size, so that the risk per trade (amount...
Dear Copernicus,

Thank you very much for the interesting linked article. I wonder how different the results may be if a coin entry was used every day. So, essentially, the gentleman tested the effectiveness of a trailing ATR exit? Monkey entry, but calculated exit

Wonder how a calculated entry with a monkey exit would perform..

Best Regards,
Tzamo
"Work hard, put in the effort, and follow your dreams..." -Me
  • #42
  • Jun 27, 2019 11:28pm Jun 27, 2019 11:28pm
Quoting Copernicus
{quote} Here is a 21 year test undertaken by Jez Liberty on a portfolio of currencies and futures using a random entry condition of flipping a coin for short or long with a trailing stop condition and occasional re-balancing. Pretty good overall results. The system first “tosses a coin” to decide whether to go long or short the market. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. That entry-stop distance is used to calculate the position size, so that the risk per trade (amount...
One thing that bugs me about this experiment is the use of SL.

  1. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure.

So, they didn't want to use something like 10 or 15 pips or whatever most ppl use. Instead they used a fixed multiple of the volatility measure. To me, that means they were trying to use the market itself to decide a figure for the SL and in doing so, they opted for an organic stop level instead of an arbitrary round figure so their personal levels wouldn't influence the outcome. Then they used that stop level to come up with position size. Very diligent work but who says that is the right way to go about things? Right there they have introduced biases to the experiment that they hope you approve of and therefore you accept the results.

My experiment has been based on no SL and my positions have been fixed at .01 (the smooth line is a result of that and not necessarily loss hoarding and martingaling) so, while their experiment involves random entries and so does mine, these are 2 different systems which is even more interesting because obviously, random entries are some force to reckon with. Now I wish the guys who made the big NO's would come up with experiments to back their claims.

  • #43
  • Edited 1:42am Jun 28, 2019 12:00am | Edited 1:42am
Quoting alphaomega
{quote} It's all good in theory and in principle I'm also big fan of diversified trading systems. The problem however is that today the Forex market is too erratic and slow and these necessary "mega events" simply never happen on a large enough scale. With such low interest rates and the global reduction of leverage FX is losing it's attractiveness and the big sharks are moving back to the stock markets or to some alternative exotic instruments like crypto and options. Historically the stock market was and is THE PLACE to apply diversified trend...
Hi AM. Great to hear from you. :-)

I have had a bit of luck finding diversified solutions in FX using a vast array of divergent systems that seldom trade individually and restrict opportunities to 'extreme' price moves.

Here is an EA example that is applied to the D1 timeframe that certainly nails it at a diversified level but I am not willing to to provide the dirty details of the code as it has been a labor of love and it is risk weighted. It comprises about 140 separate divergent systems that have a very low trade frequency but a high average hold. The reason for the high number of systems is to provide a suite of uncorrelated divergent design styles. This is where I am focusing my efforts towards. It is in the EA design that I can lock in permanent uncorrelated relationships. In fact the design considerations lock in co-integrated relationships which are more powerful and static in nature than correlation relationships. Diversification across markets is a moving feast and trying to find uncorrelated markets is, as you said, getting harder these days.

This EA does not use pyramiding at the strategy level but across many systems there can be many trades activated at once.....so this in a way applies principles of 'pyramiding' to a degree.

This exe file has an expiry but can be tested by anyone interested. I am bundling together many examples of this which I test in the live environment...but it is a long winded exercise where success is measured in years and not months.

If using on demo or live account, then make sure it is loaded on a XAUUSD chart. Once loaded it will create the universe of charts and also provide a dashboard to track status of trades. You will need to load a new XAUUSD chart once the portfoluio has been created. I use $10000 in settings as a starting balance for testing purposes and I don't apply compounding.

If you want to apply compounding based on equity then set to 0 in settings and it will compound based on equity of the entire portfolio .....but in StrategyTester it simply compounds the single instrument as you can only test an instrument at a time.

Apply it to the following universe on the D1 timeframe only:
EURUSD D1
GBPJPY D1
GBPUSD D1
EURJPY D1
AUDUSD D1
USDJPY D1
NZDUSD D1
CADJPY D1
XAUUSD D1
USDCAD D1
EURCAD D1
GBPNZD D1

Do not apply to other markets.

Each year I re-balance and review these portfolio compilations.

Here are the results from Strategy Tester using Pepperstone in $Aus with spread, commissions, SWAP and slippage assumptions applied. Note that this style of portfolio attracts drawdowns early as the average hold per trade is long and loss trades are incurred initially on starting the portfolio.

Performance Results
Attached Image (click to enlarge)
Click to Enlarge Name: Capture1.PNG Size: 44 KB


Portfolio Equity Curve
Attached Image (click to enlarge)
Click to Enlarge Name: Capture2.PNG Size: 91 KB


Individual Return Stream Contributions of Portfolio
Attached Image (click to enlarge)
Click to Enlarge Name: Capture3.PNG Size: 350 KB



Guys....this is to play with only to demonstrate that some divergent models work well in the Forex context if you are patient and prepared to look at things from a long term perspective and as a wealth building as opposed to cash flow generation exercise.

Cheers

C
Attached File(s)
File Type: zip 1000 Portfolio.2019.06.26.zip   571 KB | 578 downloads
  • #44
  • Edited 12:22am Jun 28, 2019 12:10am | Edited 12:22am
Quoting OutThere
{quote} Okay, so random entries work. Why isn't everyone using random entries then? I mean there are these established do's and don'ts in trading such as martingale is bad. SL is good. Don't revenge trade, read news, trend is your friend Etc. Why don't we have random entries as being good? People spend untold amounts of energy and time to get the entry right and I don't blame them. A good entry is obviously good but hey, if random entry works, maybe everyone should focus on exits. That was the whole point of my random entry EA screenshot. Work on...
It is a tough nut to crack. If you are applying predictive methods such as mean reversion then entries matter. If you apply price following methods under diversification then exits rather than entries matter. I personally think that very few retail traders apply price following successfully. They might think they are price followers...but probably underneath it all have prediction in the back of their mind :-)
  • #45
  • Jun 28, 2019 12:16am Jun 28, 2019 12:16am
Quoting tzamo
{quote} Dear Copernicus, Thank you very much for the interesting linked article. I wonder how different the results may be if a coin entry was used every day. So, essentially, the gentleman tested the effectiveness of a trailing ATR exit? Monkey entry, but calculated exit Wonder how a calculated entry with a monkey exit would perform.. Best Regards, Tzamo
Hi tzamo. I hope all is well mate :-)

There is another Blog post of his which looks deeper into this mystery.

In this post he compared random entries with trailing stops, random entries with random exits, random entries with profit targets, MA entries with random exits and MA entries with profit targets. This is in the context of a diversified 'divergent' portfolio. This principle does not apply to convergent systems such as mean reversion.

Cheers C
  • #46
  • Jun 28, 2019 12:18am Jun 28, 2019 12:18am
Quoting OutThere
{quote} One thing that bugs me about this experiment is the use of SL. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. So, they didn't want to use something like 10 or 15 pips or whatever most ppl use. Instead they used a fixed multiple of the volatility measure. To me, that means they were trying to use the market itself to decide a figure for the SL and in doing so, they opted for an organic stop level instead of an arbitrary round figure so their personal levels wouldn't influence...
I must admit I am a big fan of volatility based trailing and position sizing methods. With divergent market conditions when the ***** hits the fan, things move very fast and furious. Volatility based methods really assist risk management in conditions such as these and prevent you from trading above your pay grade. :-)
  • #47
  • Jun 28, 2019 12:43am Jun 28, 2019 12:43am
Your math is flawed. You are using the same logic people use at a craps table. "A seven has to come up every 5 to 10 rolls. so I bet when there have been 5 rolls with no seven. You know the odds of a seven not coming up after 5 rolls? ". The correct answer is still the probability that the next roll is a 7. 1 in 6. Just like I can't lose 25 times in a row. You flip a coin. You always have a 1 in 2 chance of the next trade losing....
  • #48
  • Jun 28, 2019 2:16am Jun 28, 2019 2:16am
Quoting OutThere
I wasn't going to mention anything about this topic but big words and twisted explanations and the plethora of colorful NO's got me. It is amazing how many clueless trades are around doling out advice. I don't even know what ' diversified divergent trading systems' means. Oh is it like 'trend following/momentum breakout styles'? I thought random entries were random entries which means you enter pairs anytime, any trend any pair and for no reason. Similar to the monkey throwing darts at the wall with a random list of pairs written all over it. News...
"No SL. SL is for losers." This is very popular misconception among the people that blow accounts consistently . They are usually fooled by some short term success that ends catastrophically (as showed below )

If you have your equity curve dragging below your balance curve, your system is RISK ACCUMULATING. All this system look great until the catastrophic failure that always comes sooner or later(never to late, max 1-2 years)

here are some examples
Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 157 KB
  • #49
  • Edited 4:09am Jun 28, 2019 3:30am | Edited 4:09am
Quoting OutThere
{quote} One thing that bugs me about this experiment is the use of SL. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. So, they didn't want to use something like 10 or 15 pips or whatever most ppl use. Instead they used a fixed multiple of the volatility measure. To me, that means they were trying to use the market itself to decide a figure for the SL and in doing so, they opted for an organic stop level instead of an arbitrary round figure so their personal levels wouldn't influence...
The big NO still in place until you prove you can do it.

I have done research in this area before; so I am not talking just for the sake of it.

Let's explain why it won't work;
- Say you have entered a random trade; and it was Long for the sake of it.
- Market went against you, and you added legs to your position, martingale style
- Market kept going against you, forever; i.e. EURUSD used to be 1.4, and now it is in 1.14; it never went back to that area again.

this scenario can happen; and it will happen again and again.

Random entries does not have an edge, period. You may enjoy a nice equity curve; but one day that death cliff is going to happen.

In order to improve a martingale system; and lower the risk; you need to introduce an entry filter that has a slight edge; and the same filter will help in exiting the martingale; without waiting for a net profit to occur.

that net profit of positions; will NEVER occur all the time if your entry is random.

You may get away with it; if you add a time exit. for example; exit at EOD no matter what; or exit when the basket age is 4 hours, etc. This will introduce some volatility to your equity curve; but it is better than waiting and hoping.

If anybody remembers FXViper momey manager who works with Nick Mcdonald; may remember that he did very well in the PNL; until that **** hit the fan; he waited days, then weeks, then months, then eventually lost more than 50% of the fund; and closed down. He did it with the above style; entered a basket of trades against the market; hoping it will revert; and it never did; he kept adding and adding; and during the months of DD; the man was down, morale destroyed, and clients kept hassling. that morale ride is tough; and killing.
  • #50
  • Jun 28, 2019 4:06am Jun 28, 2019 4:06am
I did a quick backtest of a random martingale system; with 0.3/0.3/0.3 chances for Long/Short/Neutral. it then adds legs as it go.

It does not exit at EOD; it carries on.

The only exit method is net profit of the basket of 5 pips.

2019 seen a fantastic curve; 45" smooth curve.

2015 seen the death cliff; which took few months in DD; and almost blow the account

I stress the importance of time exit in this type of strategies; even if the martingale is not random; use a time exit to get the hell out before that cliff comes visiting.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: random2019.jpg Size: 339 KB
Click to Enlarge Name: random2015.jpg Size: 277 KB
  • #51
  • Edited 10:38pm Jun 28, 2019 5:22am | Edited 10:38pm
A “50% win-rate @ 1R” brings us nowhere. In fact, with trading being a negative-sum game, we can only lose using that strategy ...
Market is not random but unpredictable
  • #52
  • Edited Jun 29, 2019 5:12am Jun 28, 2019 6:15am | Edited Jun 29, 2019 5:12am
all techicals are meant to predict, while FA is meant to anticipate.

I assume SRs work because there's liquidity there (aka stop hunting)..
  • #53
  • Jun 28, 2019 6:29am Jun 28, 2019 6:29am
Quoting alphadude
I did a quick backtest of a random martingale system; with 0.3/0.3/0.3 chances for Long/Short/Neutral. it then adds legs as it go. It does not exit at EOD; it carries on. The only exit method is net profit of the basket of 5 pips. 2019 seen a fantastic curve; 45" smooth curve. 2015 seen the death cliff; which took few months in DD; and almost blow the account I stress the importance of time exit in this type of strategies; even if the martingale is not random; use a time exit to get the hell out before that cliff comes visiting. {image} {image}
You might be right about all that you said and/or might not.

What I mean is yes and no.

Any monkey can place a trade. Right? After that, the trades start a life of their own. That applies to any entry method. All entries are more or less random. Maybe you can explain a win or a loss by the reasoning behind your move to enter but things could have easily gone the other way.

This is not an attempt to come up with reasons why not use SL or a proof for anything. I leave that for people who benefit from providing such proofs. I think if there is anything to take away from this, it is that maybe a trader can suspend all attempts to enter the market at a "correct" place and still make a few bucks. That's all. Maybe you are saying that going that route is finding a dead end. Best not to go there. Right?

You might be right, you might be wrong but it's worth a try.
  • #54
  • Jun 28, 2019 9:35am Jun 28, 2019 9:35am
Quoting OutThere
{quote} You might be right about all that you said and/or might not. What I mean is yes and no. Any monkey can place a trade. Right? After that, the trades start a life of their own. That applies to any entry method. All entries are more or less random. Maybe you can explain a win or a loss by the reasoning behind your move to enter but things could have easily gone the other way. This is not an attempt to come up with reasons why not use SL or a proof for anything. I leave that for people who benefit from providing such proofs. I think...
If market basics - momentum and symbols with strong mean reversion tendencies like EU and GY - are in place, it is worth attempting entering random entries with equity profit and equity loss as exits.
Practice makes a person perfect
  • #55
  • Jun 28, 2019 2:14pm Jun 28, 2019 2:14pm
btw FX is NOT a negative sum game LOL

http://www.businessdictionary.com/de...-sum-game.html


not even zero sum if we want to be precise: central banks print money therefore it's positive sum.
  • #56
  • Jun 28, 2019 3:46pm Jun 28, 2019 3:46pm
Quoting andoseg2
Martingale work with two conditions : 1) You have a 100.000$ account 2) you risk 0.10% per entry (100$)
you mean the first entry lol
  • #57
  • Jun 29, 2019 3:56am Jun 29, 2019 3:56am
I guess my point is..

I've never seen any EA with a perfectly flat equity curve.

Now I get it.. no one is trying to make a flat equity curve. it's either up or we failed and it's down.

But a 50 50 and 1:1 system without factoring in spread.. would be flat.

Has anyone ever seen that??

Prove that first.. and you can prove the rest.

Random entries.. random 50/50 profit and loss with similar standard deviations and streaks..

basically.. create a system that is exactly like flipping a coin.. win 30 pips half the time.. lose 30 pips half the time.. with no patterns for entries..

And I'll show you how to make money with it like OP asked. without martingale.

Any takers?
  • #58
  • Jun 29, 2019 5:14am Jun 29, 2019 5:14am
Quoting Copernicus
{quote} Here is a 21 year test undertaken by Jez Liberty on a portfolio of currencies and futures using a random entry condition of flipping a coin for short or long with a trailing stop condition and occasional re-balancing. Pretty good overall results. The system first “tosses a coin” to decide whether to go long or short the market. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. That entry-stop distance is used to calculate the position size, so that the risk per trade (amount...

looks very interesting !
  • #59
  • Jun 29, 2019 9:37am Jun 29, 2019 9:37am
Quoting Copernicus
{quote} Here is a 21 year test undertaken by Jez Liberty on a portfolio of currencies and futures using a random entry condition of flipping a coin for short or long with a trailing stop condition and occasional re-balancing. Pretty good overall results. The system first “tosses a coin” to decide whether to go long or short the market. An initial stop is set below/above the entry price at a distance equal to a fixed multiple of the volatility measure. That entry-stop distance is used to calculate the position size, so that the risk per trade (amount...
pardon me for asking, can the principle of holding and trailing be used for other entry method and actually make it profitable?
for example a moving average cross of SMA 50 and SMA 5 (thats just some random setting)
enter long when SMA 5 cross upward SMA 50 and hold + trail stop to SMA 50 until SMA 5 cross downward SMA 50 and vice versa for short
In the end its just a stupid numbers game
  • #60
  • Jun 29, 2019 10:04am Jun 29, 2019 10:04am
Quoting JaxPacific
I guess my point is.. I've never seen any EA with a perfectly flat equity curve. Now I get it.. no one is trying to make a flat equity curve. it's either up or we failed and it's down. But a 50 50 and 1:1 system without factoring in spread.. would be flat. Has anyone ever seen that?? Prove that first.. and you can prove the rest. Random entries.. random 50/50 profit and loss with similar standard deviations and streaks.. basically.. create a system that is exactly like flipping a coin.. win 30 pips half the time.. lose 30 pips half the time.. with...
What do you mean by " a perfectly flat equity curve"?

Do you mean that there are several open trades in the terminal and no matter how much the prices of underlying pairs go up and down, the equity stays the same?

If that is what you are asking then it is relatively easy to do.

Or you want trades closed constantly in profit and loss while the equity remains the same? Brokers would just love such a system. I have had EA experiments that went nowhere despite of hundreds of trades opened and closed. I consider those garbage EA's.

Meaning something else? Explain better please.
0 traders viewing now