Plutonite's Dynamic Trailing Stops | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Dynamic TakeProfit & Dynamic StopLoss EA/AI/Algorithm 0 replies

Trailing Entries & Trailing Stops? 26 replies

Plutonite's Trading Manual 154 replies

To Trailing Stop or Not To Trailing Stop.... 36 replies

Plutonite's Live Journal 103 replies

Attachments: Plutonite's Dynamic Trailing Stops
Exit Attachments

Plutonite's Dynamic Trailing Stops

  • #1
  • First Post: Edited 10:19pm Apr 28, 2007 9:23pm | Edited 10:19pm
Hello everybody,

We all have a dream: to be able to hold the profits we make on a trade, take every run to its finish, and protect ourselves from whips.

Everyone knows that it is easy to enter a trade, but usually very difficult to make a reasonable exit, so to protect their profits, many people use trailing stops (i.e stop losses that move upward with a profitable long, downwards with a short) but are still unhappy due to the simplicity of such methods, which results in most of their profits being lost.

As a possible solution, I have devised this system which I would like to share with you. I will not discuss how to enter the market, only how to exit. There is no T/P, only the S/L value which we will define as:

"The number of pips, below the last candle's closing value, at which we will terminate the trade "

Now we will actually vary the size of this stop loss via a simple calculation made at the close of every following candle once we enter a trade. I will stop here in this post because at this point several options (ATR..etc) are available, all suggestions are welcome. My own calculations are in the following post.
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #2
  • Edited 10:29pm Apr 28, 2007 9:59pm | Edited 10:29pm
Ok, we want to put a value on our stop loss, which is taken from the close of the previous candle. We will deal with only LONG positions for simplicity. Short positions are exactly the same, just do the reverse. We will deal with H4 bars in all examples.

Now you have entered a long trade. Start with your intitial S/L value X. The value of X depends on the chart, something like 60-70 pips for GBPJPY or 25-30 for EURUSD. At the close of this candle, you will either have:

  1. A candlestick greater in size than the previous, upwards in direction(i.e close > open)
  2. A candlestick smaller in size than the previous, upwards
  3. A candlestick in the opposite direction (i.e close < open)

We will assign the value based on the case. For th first case:

SL = (previous SL) + Y*R
where Y is say 20 pips, and R is the ratio of this candlestick's height (close-open) to the previous candlestick's height (close - open).

For the second and third case:

SL = (previous SL) - Y*R,

Where R is the same as above, and Y = 5 for the second case, Y=20 for the third. These numbers are all for GBPJPY.

Finally, you have a maximum value and min value for your SL that cannot be exceeded. This should be around 120 Max and 30 MIN, again for GBPJPY.

Examples to follow soon.

Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #3
  • Edited 11:25pm Apr 28, 2007 10:12pm | Edited 11:25pm
As a comment on the method, you will notice that it tightens the stop loss when the trade is not going favorably, and continues to shrink the SL value if the market persists against you. In consolidation periods it also tightens the SL bit by bit, helping you get a grip on the profits in case of a reversal.

In strong upwards movements however, it is designed to widen the SL, which is logical - think volatility and profit taking.

Try it out on a trending chart. You will be surprised. The numbers are generally dependent on the volatility of the chart at hand. Few charts will use bigger numbers than the GBPJPY "dragon"
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #4
  • Apr 28, 2007 10:39pm Apr 28, 2007 10:39pm
Here is a recent example from this month.
Attached Image (click to enlarge)
Click to Enlarge Name: shot.jpg Size: 83 KB
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #5
  • Apr 28, 2007 11:13pm Apr 28, 2007 11:13pm
And in the following week, another 300-350 pips. That's 700-800 pips in 2 weeks, entry based on following simple price action and exit based on this method. Life is not bad, eh?

Demonstration with ranging market coming soon!

Let me know what you think so far.
Attached Image (click to enlarge)
Click to Enlarge Name: shot2.jpg Size: 88 KB
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #6
  • Apr 29, 2007 6:31am Apr 29, 2007 6:31am
This looks good.

Do you have a formula for arriving at Y?

Thank you.
  • #7
  • Apr 29, 2007 7:11am Apr 29, 2007 7:11am
Hi Plutonite:
I have a question on determining Y. You mention " where Y is say 20 pips". How are you determining this number?

BTW - I have been thinking along the same lines as you. My thoughts that the standard "Exit trend on a break of an MA" does give back to much.

Quoting Plutonite
Ok, we want to put a value on our stop loss, which is taken from the close of the previous candle. We will deal with only LONG positions for simplicity. Short positions are exactly the same, just do the reverse. We will deal with H4 bars in all examples.

Now you have entered a long trade. Start with your intitial S/L value X. The value of X depends on the chart, something like 60-70 pips for GBPJPY or 25-30 for EURUSD. At the close of this candle, you will either have:

  1. A candlestick greater in size than the previous, upwards in direction(i.e close > open)
  2. A candlestick smaller in size than the previous, upwards
  3. A candlestick in the opposite direction (i.e close < open)

We will assign the value based on the case. For th first case:

SL = (previous SL) + Y*R
where Y is say 20 pips, and R is the ratio of this candlestick's height (close-open) to the previous candlestick's height (close - open).

For the second and third case:

SL = (previous SL) - Y*R,

Where R is the same as above, and Y = 5 for the second case, Y=20 for the third. These numbers are all for GBPJPY.

Finally, you have a maximum value and min value for your SL that cannot be exceeded. This should be around 120 Max and 30 MIN, again for GBPJPY.

Examples to follow soon.

In trading, there is no bullshit. You either make money or you don't.
  • #8
  • Apr 29, 2007 8:46am Apr 29, 2007 8:46am
Give the full set of rules un-ambiguously and I will program an EA that can be attached to any order and will manage the stop with this method.

Like the others, I don't know where Y comes from.
  • #9
  • Edited 10:48am Apr 29, 2007 10:17am | Edited 10:48am
Hi guys

Thanks for the response!

Y is not arrived at by "rules", it is just the number of pips by which we widen or shrink the S/L depending on whether the movement is speeding up, slowing down, or reversing against us.

I chose 20, -5 and -20 (for each case respectively) because that works out nicely through trial and error for this pair(GBPJPY). For EURUSD something like 6, -2 and -6 could work.

I'm sorry for any confusion, I'll try to post a lengthy numerical example. And thanks for the offer of EA, I would've programmed it myself but I want to arrive at the best method first!

Please play around with it and see if you can come up with a better equation/method. The only requirement is that we shift the SL to the open of the current candle(close of the previous), for every new candle.
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #10
  • Apr 29, 2007 12:52pm Apr 29, 2007 12:52pm
For those who want to want to experiment with how conservative/loose the SL can be made, I suggest having a Y1, Y2, and Y3 (one for each situation) and playing with each on it's own. The same can be done for R, which I am trying to improve now because a ration of the lengths of the previous 2 candles is by no means an ideal measure of volatility.

By the way all of this is for the sake of automation/aiding the trader's decision. Many good traders will be doing this subconsciously although they may not move the SL every candle as we do here
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #11
  • Apr 29, 2007 7:55pm Apr 29, 2007 7:55pm
Quoting Plutonite
For those who want to want to experiment with how conservative/loose the SL can be made, I suggest having a Y1, Y2, and Y3 (one for each situation) and playing with each on it's own. The same can be done for R, which I am trying to improve now because a ration of the lengths of the previous 2 candles is by no means an ideal measure of volatility.

By the way all of this is for the sake of automation/aiding the trader's decision. Many good traders will be doing this subconsciously although they may not move the SL every candle as we do here
In addition, today I realized that stochastic methods from statistical analysis produce similar results(because our method borrows from probability theory). To see for yourself, try a full stochastic indicator with settings (20, 8, 4) on a H4 chart and trace after a good entry in a similar manner to the examples above.

Regards,
-P
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #12
  • May 17, 2007 8:52am May 17, 2007 8:52am
Very interesting concept here. Rather than providing an aribtrary figure for y, had you considered using ATR or multiples of ATR, then you have something fully dynamic that takes into account ATR and well as the ratio's that you speak of which take into account the rate and direction of movement.

For example: the modified formula might look like this (where ATR is a 9 period ATR):

SL = (previous SL) + (ATR*3)*R

This might mean that rather than a fixed y-figure for a trending market and a congested market, you have a variable ratio that bears some significance to the R-Ratio.

Without testing it, I am not sure whether or not this would increase whip, but in my opinion bringing ATR helps look at market breadth, which is uesful when you are measuring volatility.
  • #13
  • May 17, 2007 9:04am May 17, 2007 9:04am
Quoting rogerha
Very interesting concept here. Rather than providing an aribtrary figure for y, had you considered using ATR or multiples of ATR, then you have something fully dynamic that takes into account ATR and well as the ratio's that you speak of which take into account the rate and direction of movement.

For example: the modified formula might look like this (where ATR is a 9 period ATR):

SL = (previous SL) + (ATR*3)*R

This might mean that rather than a fixed y-figure for a trending market and a congested market, you have a variable ratio that bears some significance to the R-Ratio.

Without testing it, I am not sure whether or not this would increase whip, but in my opinion bringing ATR helps look at market breadth, which is uesful when you are measuring volatility.
Thank you for that rogerha! I definitely have, but I couldn't for the life of understand this nonsensical ATR stuff before, and decided to ditch it as a celebration of my ignorance. Now its a good thing you brought this topic up, USDJPY is making a trend run at the moment and I was going to return to this thread to find some SL enlightenment from...myself haha.

Anyway that is a great suggestion. I will try out some numbers and get back to you. By the way guys this is NOT a trailing stop executed by your broker! You do not need your broker to have a trailing stop. This is a manual method that requires you to change your SL on each bar completion. It can be programmed, and we might just see an EA soon.

Thanks again rogerha. PS: any thoughts on how we can put signs on the ATR factor so that the SL doesnt just grow in magnitude every time?
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #14
  • May 18, 2007 3:42am May 18, 2007 3:42am
Actually, thank you Plutonite !

Not sure what you mean with your last question. My idea would mean that the ATR would not be a static figure but actually derived from the live figure of ATR(9) (or a another more approriate length). I am not sure if that answers your question though (?)
  • #15
  • Edited 5:24pm May 21, 2007 2:38pm | Edited 5:24pm
This looks like a fabulous idea. If you guys think it is worth the effort, I would like to implement it into my order managing EA which you can find here:

http://www.forex-tsd.com/expert-advi...anagement.html

I can also make a stand-alone EA or script if that is wanted too. Does this type of trailing stop usually prove to be more effective than a standard one, or you just posting good examples?

Since you are using candle closing values, this method would also changed depending on the timeframe given, which would make it work a lot like an indicator. Does this dynamic TP work well on other time frames smaller or larger than 4h? AND, finally, examples so far examples are on the gbp-jpy; some examples on other pairs would be nice.
  • #16
  • May 23, 2007 9:22am May 23, 2007 9:22am
This is an interesting topic that needs to be explored more deeply. I thought of doing the same once with Parabolic SAR. When the market goes into a consolidation phase (measured by something like ATR, fractal indicator), the acceleration setting would increase to get you out the market faster. I never explored it more thoroughly, but this method would be easier to implement as the process involved in Parabolic SAR is kind of complex.

Quoting Plutonite
Thank you for that rogerha! I definitely have, but I couldn't for the life of understand this nonsensical ATR stuff before, and decided to ditch it as a celebration of my ignorance.
Don't let ATR intimidate you... it's just an average range that is adjusted for gaps. Since FOREX never gaps, except occasionally over the weekend, there really is no need to adjust for gaps in this market. Therefore, using a plain average range will give you results that are almost the same... and average range is just a moving average of High minus Low... simple as that.
  • #17
  • May 27, 2007 4:49pm May 27, 2007 4:49pm
Hey Guys,

any news here ? i would love to see an indicator/ea based on the dynamic trailing stop..i find that quite an good idea !!

Many thx to Plutonite for introducing this !

Kind Regards,
Aidan
  • #18
  • Jun 5, 2007 7:33am Jun 5, 2007 7:33am
charging............... !!! CLEAR !!!! POW!!!!!!!

Beep, Beep........Beep, Beep..........Beep, Beep............

Just doing what I can to keep this thread alive. I have a dream as well Plutonite. Hopefully we can see this thing come to life!
  • #19
  • Jun 5, 2007 8:24am Jun 5, 2007 8:24am
Shhhhhh...!!! I'm supposed to be on vacation for chrissakes! Don't you guys ever take a break from trying new stuff? You must be real fun at parties! *sigh*



Ok here's the deal, there are a few ATR based EAs for managing stop losses out there. They are generally pretty decent for trending periods. If someone can hack together an indicator script to change those into a dynamically moving stop loss that modifies the trade every bar with an accumulated ATR value taking into account increments/decrements according to direction, you're an angel.

I personally like to manage my exits myself, because ATR simply looks at candle shapes, with no understanding of whats happening whatsoever. So if there is an important news item coming up right after the point where this method is going to throw you out of the market, you might want to wait instead of re-entering and losing the spread. And heavy news spreads are ugly things, not to mention servers smoking up..etc.

I therefore suggest that you learn to use this method manually and apply intelligent discretion rather than a program. A small calculation every 4 hours can't hurt. Even better, if you're a little more experienced, you can simply look at the price action without the ATR thingie at all.

Ok I've written more than I've vowed not to, time to dissapear into the shadows again. BTW does anyone know what the hell is happening to the AUDUSD? Friggin Russians. We should have bombed them during the cold war, saved millions of traders endless headaches. Yes, I just advocated global nuclear war in favor of a hard trading day And if it doesn't dive soon I will leave this mad market.

Good luck all. Someone yell at me if you want to program this thing, I'd do it but the fam would kill me. No programming for another 2 months.

Regards,
-P
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #20
  • Apr 13, 2008 5:49am Apr 13, 2008 5:49am
It's a shame this one up and died without so much as a single indicatorized version of the formula(visualized stop level) posted (instead of an EA).
Trust price. Know yourself.
0 traders viewing now