Random entries - Let's take it to the extreme | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is random really random? 231 replies

Coin tossing and random entries 44 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: Random entries - Let's take it to the extreme
Exit Attachments

Random entries - Let's take it to the extreme

  • #21
  • Dec 11, 2019 7:52am Dec 11, 2019 7:52am
38 days in, results seem promising thus far.
Random entries and deliberate exits. How far will this go?
Attached Image (click to enlarge)
Click to Enlarge Name: 1.JPG Size: 226 KB
  • #22
  • Dec 13, 2019 6:40pm Dec 13, 2019 6:40pm
Quoting carddard
38 days in, results seem promising thus far. Random entries and deliberate exits. How far will this go?
Interesting... over 1500 trades in 38 days! Do your "deliberate" exits involve PA/TA with a hard TP/SL, or are you opening up trades and letting them run with a trailing stop?
  • #23
  • Dec 18, 2019 4:33am Dec 18, 2019 4:33am
Quoting davelansing
{quote} Rennaissance, That's the conclusion that I've come to as well by way of extensive backtesting (see my previous post on backtesting random entries with specific exit strategies).... but the arguments - and copious amounts of DATA - presented by 60minuteman (Thread: "Systematic Trend Trading", and Copernicus (Threads: "Diversified Trend Trading Approach", and "Systematic Trend Trading"), and many others, also supports the idea that asymmetrical risk:reward trading (i.e. trend trading) has long-term profitability- especially when compounded....
I have dabbled with alot with trend following approaches, my advice to anyone is to stay away from it. It loses money 80% of the times. I have sort of gone towards pattern recognition. I look for situations that have a bias of atleast 60%. That's what I will recommend, trend following is not a good approach in today's market. Why? Because everyone is trying to do it.cheers
When you see it, BET big.
  • #24
  • Feb 17, 2020 6:55am Feb 17, 2020 6:55am
Quoting Rennaissance
{quote} I have dabbled with alot with trend following approaches, my advice to anyone is to stay away from it. It loses money 80% of the times. I have sort of gone towards pattern recognition. I look for situations that have a bias of atleast 60%. That's what I will recommend, trend following is not a good approach in today's market. Why? Because everyone is trying to do it.cheers
My experience with GJ has been a happy one. Following is the process I followed before sticking to GJ trading a few years ago.
The process is as follows:

1. Choose the right candidate by taking a look at volatility statistics or ADRs Monthly or Quarterly dashboard .

2. Then use simple indicators (like provided on the mql link) or Pips Bands to measure the volatility and identify outliers. Free Indicators coded by https://www.mql5.com/en/users/mt-coder are quite helpful. Bulls and bears and Big times particularly.

3. My favoured symbol is GBPJPY. I have not tried averaging trades with any other currency pair because it provides momentum as well as volatility. Important is your judgement of momentum birth .

4, If I were to select any new candidate, I would look back at 2 years of data and plot the maximum days or hours a currency or stock index has taken on a daily TF between four consecutive Outliers. Flash crash is not considered an outlier.

For GBPJPY I have defined the size of the outlier as 1200 pips . I stuff my account with 1200 +5% money worth the pips. Lots can be anything from 0.10 to 5 lots depending upon the capital capacity. Even micro lots will do as long as you find a honest broker. Nanno or Micro lot traders may not easily find such a credible broker.

5. I can build Balance easily up to 50% by trading 60 pips sized trade within a period of 30-45 days and a TP of 30 pips when I feel momentum will be there. The Pain points start from the moment the drawdown starts hitting 30% and more and can last until next 90 days. Accumulation and Distribution go on uninterrupted.

6. Presently I feel the symbol lack momentum so I have just posted pending Orders at the last Outlier mark. In a year you can get maximum two outliers generating 300 to 400 pips each effortlessly. You have to have pending order in place for this kind of trading.

7. How you handle the pain until reversal happens is something that can be worked out depending upon your willingness to pick up new skills of managing drawdowns by Hedging method. Hedging needs further injection of 600 pips capital assuming you are not able to handle locking and unlocking loss positions. Hedging is not risky if you just want to lock your losses to regain margins and baby-sit drawdown.

I have tried it with 100 pips lock and TP of 20 pips. When the pair is in momentum and gunning for consecutive runs, it is in no mood to pull back or retrace by more than 20% . Locking losses by way of hedging helps control margin. Unlocking the hedges opens up the possibility of increasing the drawdown. Most important is you have to monitor your trades and use any kind of bands BB, MA, Fibo Pivot, etc. to manage unlocking and re-locking.

8. I prefer to wait and bear the swap costs which might take 20% of the profits when profits are delivered. Psychologically, I am prepared for it because I have gone through this state of mind several times with no harm being done. It is a matter of managing your psychology. It takes at least 90 days to reverse but I am mentally prepared for 120 days and with an Equity Stop Loss. Luckily, it has not been hit in last 10 years. Unluckily, when flash crashes have happened, I was not trading. If I had, I would have made good money because I use TP and I only SELL. GBP Flash crashes have wiped out the buyers margins.

9. Your estimate of TIME for the whole cycle has to be based on lookback analysis of at least 2 outliers (for me it's 4 outliers and not 2 years). It should not be based on hopes or expectations. You also have to take a call on timing. Momentum is the key.

10. My experience says it is extremely difficult to be consistent with GBPJPY trading. So you have to be selective. Two cycles of 40% gains in a year. total 80% can be achieved with Outlier trading.
Practice makes a person perfect
  • #25
  • Feb 18, 2020 1:42am Feb 18, 2020 1:42am
It's been 3 months since this experiment was commissioned.
Some results for the experiment so far (see images below).

What I didn't like was the massive drawdown, which was expected and non-ideal, given that the entry criteria is not the most refined. But the crux of this experiment is to create a robust way to manage positions despite random entries at every interval. The natural next step was to refine the entry criteria (which I did), and that brings phase 2 of the experiment, which I started in Jan this year.

Let's see how this plays out..
Attached Image (click to enlarge)
Click to Enlarge Name: 1.JPG Size: 32 KB
Attached Images
  • #26
  • Feb 18, 2020 1:50am Feb 18, 2020 1:50am
Why is your all time pips gained negative?
When you see it, BET big.
  • #27
  • Jan 28, 2023 8:15am Jan 28, 2023 8:15am
It's been 2 years since I last posted on this thread. A recent FF member reminded me of this thread. You do see some similar threads around in the forum but nothing too serious. Quite the thrill, entering randomly in the markets and coming out on top. Using creative ways to combat the perennial problem of having an unlimited downside with an almost fixed upside is possible. A quadratic upside and downside, anyone?
  • #28
  • Edited 4:35am Jan 30, 2023 4:18am | Edited 4:35am
Quoting carddard
....The natural next step was to refine the entry criteria (which I did), and that brings phase 2 of the experiment, which I started in Jan this year. Let's see how this plays out.. {
How did phase2 work out? Were you able to overcome the DD?

Yes, a quadratic upside and quadratic downside... let profits run if you're going to let the losses run... e.g., open random trades of a Buy basket and random trades of a Sell basket (opposing trades relative to the buy basket) of (+) and (-) correlated pairs, and for each basket, keep the trades open until the profit target for the basket is achieved and then close the profitable basket down (keep the non-profitable basket open). For the Buy basket, and for each pair in the basket, open trades at specific pip intervals (50-100 pips) only if a pair is moving up, and for the Sell basket, open trades only if a pair is moving down - this should minimize the DD...but, probably will need to keep the trade sizes small relative to your equity to account for the potentially high number of trades that would be open at any given time...
  • #29
  • Jan 30, 2023 4:53am Jan 30, 2023 4:53am
Quoting carddard
Hi all, Before you go on, let's keep an open mind. Here's a thought experiment. We like to talk about how edges decay over time (some swear by an average of 3 years before the efficacy of a strategy falls). What if all trades were entered at random and all you could control is the exit strategy? A system based in random entries has to no edge to speak of, so by that logic, there is no edge to decay. But how would you trade it? I've been dabbling with this wild thought for a while now and wanted to see if there is a like-minded individual who wants...
Great thread, been trading with a flip of a coin for a couple of years now and I love it...
  • #30
  • Jan 30, 2023 4:59am Jan 30, 2023 4:59am
Quoting davelansing
{quote} How did phase2 work out? Were you able to overcome the DD? Yes, a quadratic upside and quadratic downside... let profits run if you're going to let the losses run... e.g., open random trades of a Buy basket and random trades of a Sell basket (opposing trades relative to the buy basket) of (+) and (-) correlated pairs, and for each basket, keep the trades open until the profit target for the basket is achieved and then close the profitable basket down (keep the non-profitable basket open). For the Buy basket, and for each pair in the basket,...
In phase 2, the DD levels were the same. The only way to reduce it is to introduce a more accurate entry method other than random. But the position mgt bit worked as it should.
  • #31
  • Edited 9:02pm Jan 30, 2023 5:03am | Edited 9:02pm
Quoting carddard
{quote} In phase 2, the DD levels were the same. The only way to reduce it is to introduce a more accurate entry method other than random. But the position mgt bit worked as it should.
How about open trades randomly, but only add to the trades when they are moving in the direction of the original trade? This should minimize the DD... I discussed this above...
  • #32
  • Jan 30, 2023 5:09am Jan 30, 2023 5:09am
Quoting davelansing
{quote} How about open trades randomly, but only add to the trades when they are moving in the direction of the original trade? This should minimize the DD... I discussed this above...
No significant difference over time it seems. U tried and had any luck?
  • #33
  • Jan 30, 2023 5:20am Jan 30, 2023 5:20am
Quoting carddard
{quote} No significant difference over time it seems. U tried and had any luck?
Early stages of live testing...it does take patience, but maybe the lack of patience is why so many trader's lose? We shall see...

Carddard,

Not quite sure what you mean by "no significant difference over time?"
  • #34
  • Jan 30, 2023 6:19am Jan 30, 2023 6:19am
Quoting davelansing
{quote} Early stages of live testing...it does take patience, but maybe the lack of patience is why so many trader's lose? We shall see... Carddard, Not quite sure what you mean by "no significant difference over time?"
Drawdown level - No significant difference between the 2 DCA methods. But maybe I misunderstood. What criteria do you use to assess if price is "moving in the same direction as the original trade"? I'm assuming it's adding onto losing positions only?
  • #35
  • Edited Jan 31, 2023 4:31am Jan 30, 2023 7:59am | Edited Jan 31, 2023 4:31am
Quoting carddard
{quote} What criteria do you use to assess if price is "moving in the same direction as the original trade"? I'm assuming it's adding onto losing positions only?
Not necessarily… I only add trades if price is moving in the direction of the original trade. So for instance, if I buy pair xxxyyy @ 1.09567, my next buy trade will be opened at 1.10567 (100 pip differential) and every 100 pips upwards thereafter. If price drops below my entry < 100 pips, I do nothing. However, if price drops to 1.08567 (-100 pips) and then moves up to 1.09567, I open another buy trade and will continue to open buy trades every 100 pips of upward movement. If the market continues to move down, I’m buying the rallies as long as the 100 pip up move requirement is met. So, the market does have to make an effort to move in the direction of my original trade. Vice versa for sell trades. Is this perfect? No… can I potentially be buying 100 pip rallies as the market moves 1000 pips down… indeed. But in real time we will never know if the current downward move becomes a 1000 pip adverse move or if the 100-pip rally turns into a 1000 pip uptrend, hence the basket of neg. and pos. correlated pairs to offset risk - i.e. basket1: Buy: AUDUSD/EURUSD/GBPUSD/USDCAD/USDJPY/USDCHF; basket2: Sell: AUDUSD/EURUSD/GBPUSD/USDCAD/USDJPY/USDCHF).
  • #36
  • Jan 30, 2023 9:09pm Jan 30, 2023 9:09pm
Quoting davelansing
{quote} Not necessarily… I only add trades if price is moving in the direction of the original trade. So for instance, if I buy pair xxxyyy @ 1.09567, my next buy trade will be opened at 1.10567 (100 pip differential) and every 100 pips upwards thereafter. If price drops below my entry < 100 pips, I do nothing. However, if price drops to 1.08567 (-100 pips) and then moves up to 1.09567, I open another buy trade and will continue to open buy trades every 100 pips of upward movement. If the market continues to move down, I’m buying the rallies as...
I see. That's quite different from what I've been working on. Once a trade reaches profit, I close it. Different approach to yours.
  • #37
  • May 28, 2023 1:12pm May 28, 2023 1:12pm
Interesting. I have been trying something similar

https://twitter.com/PJSmith/status/1662440710891532291

I've done some testing on this before, and the results always surprised me when compared to 'real' strategies. Certainly, be interested to follow this thread.
  • #38
  • Aug 17, 2023 1:18am Aug 17, 2023 1:18am
It's been 4 years since I started this thread, and many have reached out to find out more about this strategy.
I'd post my remarks here to address some of the questions:

1) Are you still using this strategy?
Ans: I started this as an experiment 4 years ago, and the concept has evolved overtime, but the core principles remain the same. This strategy is meant to focus on the exit leg only, hence a random entry was added to test the robustness of the concept. Yes, I am still using parts of this strategy, but with curated entries. This does not change the efficacy of the exit component. It would continue to work well with random entries (with caveats). With that said, this only forms a small part of my portfolio.

2) What starting capital should I have and what lot size do I trade?
Ans: I started this experiment with the smallest lot possible. 0.01. With a minimum capital of $5k with as high a leverage as possible so that you have ample margin to play with.

3) How much returns did you get with this strategy?
Ans: With random entries, you can refer to some of the earlier screenshots that I posted in this thread as a reference. The lower the lot size you trade with, the lower your returns (obviously).

4) Can this strategy be backtested?
Ans: Yes. For those who know me personally, I never trade manually. This strategy can be backtested. I've attached some screenshots below showing a backtest of 25 pairs from 2011 to mid 2022.

5) Can you share more about the strategy and its inner workings?
Ans: I started this experiment because before I got into fx, I dabbled with scaling in and out of trades with equities, and also tried various methods of dollar cost averaging with other instruments. That was the original inspiration for this strategy. I applied the same models into fx and made some tweaks. The core concept of this strategy is to optimize the use of your margin. Since entries are random, you need to constantly watch your margin to ensure that you have sufficient ammunition to scale in and out of trades. As time goes by, reduce your overall exposure as soon as you can. This allows you to free up your margin, which allows your to further reduce your exposure until a trade is made good. There are many caveats to this strategy, and I don't think it is for everybody. Some characteristics of this type of trading is:

  1. You are almost always in drawdown - It can be a psychological torture when trading this live. Can you handle it?
  2. You need margin for this to work consistently in the long term - It isn't going to work with $100 or $1,000 account
  3. It is a nightmare to trade this manually. I'd go as far as to say that you absolutely need this automated for it to work in the long term

I hope this helps you out if you ever stumble upon this thread.
Good luck guys.

Attached Image(s) (click to enlarge)
Click to Enlarge Name: 2.JPG Size: 211 KB
Click to Enlarge Name: 3.JPG Size: 67 KB
Click to Enlarge Name: equitychart.png Size: 38 KB
  • #39
  • Jan 10, 2024 12:47pm Jan 10, 2024 12:47pm
Quoting carddard
I hope this helps you out if you ever stumble upon this thread.
Well, I indeed stumbled on your interesting thread :-)

In my thread "The Myth of Averaging" I experimented with random entries and controlled exits. You can indeed achieve a positive expectancy using Averaging Down (adding to losing positions):
Attached Image (click to enlarge)
Click to Enlarge Name: screenshot.png Size: 10 KB

And also by Averaging Up (adding to winning positions):
Attached Image (click to enlarge)
Click to Enlarge Name: screenshot2.png Size: 10 KB

However, your description suggests that you do something else. I take it you diversify over many different symbols and use the winning trades to pay off the losing trades if they have too much drawdown. For the rest you just wait out the market fluctations until the losing positions swing back into profit. Interesting idea! I will code some experiments myself to see if I can get that to work.

Thanks for the inspiration!
  • #40
  • Jan 17, 2024 5:16am Jan 17, 2024 5:16am
Hi carddard. Is "spread" included in your backtest from August, and if so, how much?
0 traders viewing now