Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Discussion (https://www.forexfactory.com/forumdisplay.php?f=11)
-   -   Scaling out of losing trades... (https://www.forexfactory.com/showthread.php?t=30111)

PeterFM May 17, 2007 9:40am | Post# 1

Scaling out of losing trades...
 
I've started this thread after reading some posts in an old thread here.

It has raised some interesting points, and as the thread is around 2 years old, I felt some noobs (and me) could benefit from some alternative strategies when dealing with Money Management.

I once raised a question regarding scaling out of winning trades, and am comfortable with this side of my trading, but would like to get input from some of the newer, experienced traders on this particular aspect of trading.

OK, court is in session

piccolo May 17, 2007 10:18am | Post# 2

it's all about statistics and probabilities.

so let's say you consider moving stop loss to break even when price reaches x pips in profit. then you take your trades from history and check how often would such an approach benefit or harm you.

same with scaling out of loosing trades. if you look back and see that in 9 out of 10 loosing trades such an approach would benefit.. then now you know the answer.

mrgreen May 17, 2007 10:59am | Post# 3

Scaling out of a loss IMHO is a means of postponing taking the loss, and thus one is 'hopeful' that the trade will come back. Scaling Out of a win or loss reminds me of the following from ROASO:
The speculator's chief enemies are always boring from
within. It is inseparable from human nature to hope and to fear.
In speculation when the market goes against you -- you hope that
every day will be the last day and you lose more than you should
had you not listened to hope -- to the same ally that is so
potent a success-bringer to empire builders and pioneers, big
and little. And when the market goes your way you become fearful
that the next day will take away your profit, and you get out
too soon. Fear keeps you from making as much money as you ought
to. The successful trader has to fight these two deep-seated
instincts. He has to reverse what you might call his natural
impulses. Instead of hoping he must fear; instead of fearing he
must hope. He must fear that his loss may develop into a much
bigger loss, and hope that his profit may become a big profit.

superdezign May 17, 2007 11:17am | Post# 4

Scaling out of a loss IMHO is a means of postponing taking the loss, and thus one is 'hopeful' that the trade will come back. Scaling In and Out reminds me of the following from ROASO:
I disagree, I think scaling out can be a good way to keep your risk small..

lets look at an example:
Lets say you buy the EU at 1.3500, and you buy 3 lots. You decide to place your stop at 1.3450. You get stopped out with a loss of 1,500

or

lets say you buy EU at 1.3500, and you buy 3 lots. You decide to close out your trade if it reaches 1.3450, but you decide to scale out as the trade goes against you. It moves against you 20 pips, so you close one lot for a loss of 200. You still have two lots and the price moves against you 20 more pips, so you close another lot for a loss of 400. You have one more lot which you will close in ten pips. If it goes against you once again you close out the final lot for a loss of 500 and a total loss of 1,100. A difference of 400 in losses.

Of course this limits your profit potential if it swings against you and then in your favor, but this is always true if you limit your risk

lower your risk = lower reward..

mrgreen May 17, 2007 11:39am | Post# 5

It is presupposed that the 1500 represents say a 2% risk to your acct. Depending on the situation, say the buy is at support, support is then broken by 20 pips and you scale out one lot, again hoping that this break is temporary and the other 2 lots, when they come back will make up for the 1 lot loss. The trade goes to 40 PIP net loss, you close another lot, hoping that when the one lot will make up for the smaller losses. In this situation I would close the whole trade at 20 pips (a 600 loss) and not hope for a turn around. Its how I would approach this scenario. Others may not do so.
I disagree, I think scaling out can be a good way to keep your risk small..

lets look at an example:
Lets say you buy the EU at 1.3500, and you buy 3 lots. You decide to place your stop at 1.3450. You get stopped out with a loss of 1,500

or

lets say you buy EU at 1.3500, and you buy 3 lots. You decide to close out your trade if it reaches 1.3450, but you decide to scale out as the trade goes against you. It moves against you 20 pips, so you close one lot for a loss of 200. You still have two lots and the price moves against you 20 more pips, so you close another lot for a loss of 400. You have one more lot which you will close in ten pips. If it goes against you once again you close out the final lot for a loss of 500 and a total loss of 1,100. A difference of 400 in losses.

Of course this limits your profit potential if it swings against you and then in your favor, but this is always true if you limit your risk

lower your risk = lower reward..

Dopey May 17, 2007 12:21pm | Post# 6

Scaling out works fine on longer-term trades. On shorter-term trades, scaling in and out doesn't work so well. So, one has to define their trading time frame first, then run the numbers and see how it works.

cesarnc May 17, 2007 12:26pm | Post# 7

Scaling out works fine on longer-term trades. On shorter-term trades, scaling in and out doesn't work so well. So, one has to define their trading time frame first, then run the numbers and see how it works.
Agree...

And no point in adding at every 20 pips. If you're in the wrong direction, better to add heavy at a bounce off support and resistance, as to bring the B/E level much closer. Otherwise you will be just adding the the loss insteady of helping the trade.

I just do that when inline with the major long-term trend, too.

witchazel May 17, 2007 1:51pm | Post# 8

The generalized formula for a system's profit factor is ( winrate*payrate)^frequency.

Scaling out while not changing stops
Winrate will remain constant: TP and SL doesnt change.
Payrate will remain constant: each lot taken off will make the wining and losing amount less.
Think of a system with 60% winrate 3:4 payrate. it has a base pay per trade of (60x3-40*4)/100 = 0.2.
Now reduce the lot size by 50%, so you now have a 60%wr and a 1.5:2payrate. 60x1.5-40x2) = 0.2.
Its constant.


Scaling out while keeping the risk constant
Definition: you have a TP of 20 and a SL of 20. At -10, drop 50% of the lots. Extend the SL down another 6pips so the loss can be the same size.
Winrate will increase: Lowering the SL will increase the winrate
Payrate will decrease: lowering the amount that will be gained while keeping the amount lost constant.
Think of a system with 60% winrate 3:4 payrate. it has a base pay per trade of (60x3-40*4)/100 = 0.2.
The plan is to Risk 1% of an account. At -10 pips drop 50% of the lots and extend the SL so it will end up at -1% of account on loss. Say this increases winrate by 5%.
Now you have a 65% winrate 1.5:4 payrate. base ppt = (65x1.5-35x4)/100 = -0.425.
You now have a losing system.

tesla May 17, 2007 2:17pm | Post# 9

This is where the value of a detailed trade history really shows itself. Do a check of the last 50 or so trades for a specific strategy and see if scaling out of losing positions would have provided more benefit than your current stoploss rules. Witchazel gives you a starting point with a way to crunch the numbers, but the numbers have to be specific to your system.

I don't think that a vague discussion of the benefits of scaling out will provide nearly as much value.

Just one person's opinion.

Bemac May 17, 2007 6:00pm | Post# 10

I look at it this way.

Before I enter the Trade I have resolved some beliefs in my head {at least}.

If I get in Now, at Price 'n'
I have the potential of 'x' pips/lot.
I have the Risk of 'y' pips/lot.

Are there any significant Price Levels between 'n' & 'y'?
If so, how many? {let's say 2 [ note that that will give me 3 stop levels]}
I can use that info to subdivide my risk into three categories.
a) ah s#it {Loss Level 1}
b) oh fu@ {Loss Level 2} ... & and...
c) doG darn{Loss Level 3}


So I resolve to determine that at...
LL1 1 get out of 1/3 of my position.
LL1 2 get out of 1/3 of my position.
LL1 3 get out of my position.

I also resolve that IF the Last 1/3rd survives, I will reinstate 1/3rd on a subsequent break {in my direction this time} of LL2 and the same for LL1.

CAUTION.
Make sure you know what you are dealing with here because a strategy like this can absolutely whip[saw]your butt. But, if handled correctly, your acct. {not your butt } may benifit

ItalianSharp May 17, 2007 6:16pm | Post# 11

I look at it this way.

Before I enter the Trade I have resolved some beliefs in my head {at least}.

If I get in Now, at Price 'n'
I have the potential of 'x' pips/lot.
I have the Risk of 'y' pips/lot.

Are there any significant Price Levels between 'n' & 'y'?
If so, how many? {let's say 2 [ note that that will give me 3 stop levels]}
I can use that info to subdivide my risk into three categories.
a) ah s#it {Loss Level 1}
b) oh fu@ {Loss Level 2} ... & and...
c) doG darn{Loss Level 3}


So I resolve to determine that at...
LL1 1 get out of 1/3 of my position.
LL1 2 get out of 1/3 of my position.
LL1 3 get out of my position.

I also resolve that IF the Last 1/3rd survives, I will reinstate 1/3rd on a subsequent break {in my direction this time} of LL2 and the same for LL1.

CAUTION.
Make sure you know what you are dealing with here because a strategy like this can absolutely whip[saw]your butt. But, if handled correctly, your acct. {not your butt } may benifit
What about entering the position with 1 contract and adding to the position at S1, S2 and S3?

Bemac May 17, 2007 6:22pm | Post# 12

What about entering the position with 1 contract and adding to the position at S1, S2 and S3?
"Averaging In" requires Even Greater Skill. But... Yes... Good Point.
Problem being that I prefer to Enter with Multiples in the event that I am Correct.

ItalianSharp May 17, 2007 6:27pm | Post# 13

"Averaging In" requires Even Greater Skill
Agree. The most difficult part is to understand when it's time to add, how much longer to keep adding, when to take the loss and when to reverse the position.

Professional traders with large funds operate this way. I believe market makers do that too.

SunTrader May 17, 2007 6:27pm | Post# 14

I keep it simple. At major support/resistance I have my "line in the sand" beyond which if price goes I bail out and reverse the trade.

If I think it should only go so far and it goes beyond, then I am wrong and that means time to get out or reverse. Not hope.

Bemac May 17, 2007 6:29pm | Post# 15

I keep it simple. At major support/resistance I have my "line in the sand" beyond which if price goes I bail out and reverse the trade.

If I think it should only go so far and it goes beyond, then I am wrong and that means time to get out or reverse. Not hope.
So how do you determine where to Draw The Line.
ie. How do you determine Major S/R

bluebuddha May 17, 2007 6:40pm | Post# 16

Scaling out while not changing stops
Winrate will remain constant: TP and SL doesnt change.
Payrate will remain constant: each lot taken off will make the wining and losing amount less.
Think of a system with 60% winrate 3:4 payrate. it has a base pay per trade of (60x3-40*4)/100 = 0.2.
Now reduce the lot size by 50%, so you now have a 60%wr and a 1.5:2payrate. 60x1.5-40x2) = 0.2.
Its constant.
Actually, it's more complicated than that. Let's say that of the times you win, 75% of it went straight to your TP, and 25% hit the first SL or scale out point, then retraced back to your TP.

So what you would have is 0.40*1+0.4*2= 1.2 for your losses (one part is half position at 2 points, the other is half position at 4 points). For your wins, you have 0.6*0.75*3+0.6*0.25*(1.5-1.0) = 1.425 (half position lost 2 points, and other half gained 3 points). In this case, the expectancy is slightly better, at 0.225. Also, what if you re-entered your original entry on the retrace? (See jacko's "anti-hedging" strategy). In this case, your win expectancy is 0.6*0.75*3+0.6*0.25*(3.0-1.0) = 1.65, which would boost overall expectancy to 0.45.

But what if we had the price hit your first scale out 75% of the time and went straight to your TP only 25% of the time. The expectancies in this case would be:

Not re-enter: [0.6*0.25*3+0.6*0.75*(1.5-1.0)] - 1.2 = -0.525
Re-enter on retrace: [0.6*0.25*3+0.6*0.75*(3.0-1.0)] - 1.2 = 0.15

This time the expectancy is worse off, and if you don't re-enter, you wind up losing. So whether your expectancy improves or gets worse when you scale out depends on how often the market goes straight to your TP on your wins, and if you decide to re-enter when the market goes back into your direction. Not only that... but the situation gets even more complicated if you elect to re-enter, and the market goes against you again.

Scaling out while keeping the risk constant...
Personally, I think reverse trailing stops is a bad strategy no matter how you trade... sort of like Martingale or adding positions to a losing trade. So I'm not going to get into this.

SunTrader May 17, 2007 7:22pm | Post# 17

By using fib time and price ratio retracements/projections/expansions and trade pattern.

But don't get me wrong. It is what I use. I don't claim it to be "the line" and the one and only one - each and every trade. It works for me more times than not.

I just believe it is better to bail out then add to a loser or scale out. If I am wrong then there is no other decision but to cut and run.

BTW I don't just blindly reverse if my stop is hit.

WTB May 17, 2007 7:47pm | Post# 18

I personally scale out of winners in two steps: I always have one pre-defined level mid-way through the swing where I take half the lot size out and move stop to even, letting the second half run to its full potential target.

Scaling out of losers is bollocks as far as I am concerned: as soon as you detect that you're in the wrong side of the market, or you simply miss-timed your entry, get out of there ASAP. Dont scale out, just get out with a small loss and wait until a clearer signal is given.

As far as scaling in, I have never had much luck scaling in to be honest. I always end up adding up to my position at the wrong time and the profits made with the original position are eaten up by the loss on the add up. I think adding up works best for buy-and-hold long term trades off the daily/weekly charts. But on the shorter time frames I never managed to get any decent results out of it myself.

Bemac May 17, 2007 8:07pm | Post# 19

[quote=WTB;1444251]...bollocks as far as I am concerned: /quote]
Correct me if I'm wrong but doesen't Bollocks {ballaxe} relate to a Neutered Bull?

Everybody now...
Ouuucccchhh.... That's gonna leave a mark !

SedonaPipper May 17, 2007 8:37pm | Post# 20

Using the stop loss as a new entry
 
I actually spent some time putting together a spreadsheet that looked at using my first stop loss as an average down point. I calculated how many lots I would have to enter on the trade to BE on a 20 pip retracement.
I took the idea to 4 entry trades. When the market consistantly hits your stop loss then turns around and hits your first limit it seemed like a great idea.

The real problem arose when the market just went through my stop because I was in a real bad trade. With each succeeding entry, I became an emotional wreck because I was needing to get a retracement and it wasn't happening. And as the market continued to go against the position just added to my emotional stress - that's when the real bad decision happens, you cut your losses and an hour later, the market rallies 400 pips and your sitting on the sideline licking your wounds. Nope it created too much pain for me to live with.

Eventually I gave up on the idea because I was back with the same dilema, knowing that if I blew it, I would be wiped out forever. After the fourth trade my lot size was beyond my account size.

I still toy with the idea in my revenge thoughts and now I have learned to live with the first loss.

Better to take a small loss and live for another trade than to be working for the man to get enough of a bankroll to trade again. Only this time I will do it smarter - yeah right, that doesn't work either. To be a good trader requires the willingness to go through an intense behavior modification. Taking losses is part of that process.

Sedonapipper


© Forex Factory