The Three Keys

In my opinion to trade successfully intra-day we must adopt a policy which takes advantage of the theory. This simply means closing bad trades on a regression whilst only ever taking profits on prices of local extremity. By trading this way we will only ever be exchanging ‘value’ at points which are beneficial to ourselves.
An excellent post Steve that captures a number of key fundamentals to price action and speculative psychology very succinctly. Nice work.

Before anybody considers now placing their targets where their stops are and their stops where their targets are, they need to be able to have a reasonable shot at establishing the mean and extremity you refer to. The longer TFs which effectively drive this process have their own idea of future value as I’m sure you’re aware.

One of the most intelligent and informed posts for a while (apart from my own of course….)
 
An excellent post Steve that captures a number of key fundamentals to price action and speculative psychology very succinctly. Nice work.

Before anybody considers now placing their targets where their stops are and their stops where their targets are, they need to be able to have a reasonable shot at establishing the mean and extremity you refer to. The longer TFs which effectively drive this process have their own idea of future value as I’m sure you’re aware.

One of the most intelligent and informed posts for a while (apart from my own of course….)

Unfortunately it simply isn’t as simple as swapping your targets and stops around. It’s actually quite difficult to try and explain the point that I’m trying to get across. Can you see that with a stop the moment that it is triggered you are out of the market? Ask yourself what function a stop like that is actually performing? When we understand that a stop is performing a ‘service’ for us we can then move on to understand that this ‘service’ comes at a cost. It is this ‘cost’ that makes us consistently lose money. In simple terms, if we react to price is such a manner as to make us close our trade then we will more than likely be losing money.

Let me try the following as a semi practical demonstration....

Let us imagine we are trading GBP/USD. We will ignore trading costs to start with but will consider that a round trip cost is 2 pips.
Let us also assume that we are looking to scalp around 17 – 20 pips from a trade.
So, we see that GBP/USD has rallied to a round number at 1.9700. At this point we sell short at 1.9700. A trader prone to loss would now place a stop close by as he considers this ‘good practice’. Lets imagine that our generally losing trader places a 20 pip stop order to get him out of his short at 1.9720. At the same time he feels that his target for profit would be around 1.9680. He may or may not place a limit order to buy to close at this level. His feeling is that his ‘edge over the market’ is the fact that the market will normally reach points of resistance and support at round numbers and therefore by consistently making this kind of trade at these points in time he will, over the longer run, make money. At the moment this all sounds perfectly logical. (I will in due course try to prove otherwise.)
Imagine now that the market spikes higher still and reaches the stop at 1.9720. In an instant the trade is ‘stopped out’ and our friend is ‘flat’. Then what happens? Well obviously the market can still run higher or perhaps turn back lower. The chances are however that the market will run back lower and at some point will return to a level below the stop level set by our losing trader.
Try imagining now what would have happened if the trade would have gone into profit. Let’s suggest that over the course of 20 minutes the market fell those 20 pips. What would our trader do? The most likely thing is that he would have thought “Oh, this looks like it’s going lower” and held onto the position. Then, before you know it, the 20 point winner has diminished backwards into a 12 point winner... whoops... now it’s only an 8 point winner.... oh bugger lets close it for 8 while there’s still some profit left!

Do you see how both methods of ‘trade management’ are detrimental to profits? This is why our friend is a ‘generally losing trader’.

Psychologically our trader cannot win in the longer term because he is not capable of taking control of the trade management aspect of his trading. Instead he uses a ‘fixed stop loss’ because that ‘comforts’ him and causes him to believe that by doing this his potential losses from each trade are ‘under control’. Whilst this may be true to an extent our trader has failed to realise that, by allowing stops to get hit time and time again, he is consistently paying a premium to trade out of his positions at points of price extreme.

Of course the correct way to manage this type of trade is to use a mental stop which does not trigger an automatic exit from the market. Suppose in our example we set the following loose rules.

1 Mental stop of 20 pips.

2 Hard stop of 40 pips.

What we could do in terms of trade management is this. If our trade goes into loss then we will monitor it closely. If our mental stop is triggered we will move into ‘damage limitation mode’. This means that we have now accepted that our trade is a poor one and we are looking to get out. What we are waiting for is a better price than is currently available. It is possible that the market will move either way. If our mental stop of 20 pips gets hit and then the market retraces 12 pips then we can get out for a loss of 8 – see how remaining in the trade beyond our mental stop has saved us money? In most cases it will! Suppose, instead of retracing 12, the market moves on another 10 so our loss is now 30 – this can and does happen – it is still more than possible that the market will retrace 10 – 15 pips in the short term which still means a loss of less than 20 pips.
I hope that you can see that by taking responsibility for the trade management the losses can be far better controlled. This is because your exit from a trade is likely to be nearer to ‘true value’ than if you have a fixed stop set which is guaranteed to exit you from a trade at a price extreme.
Likewise by placing a limit close at a price extreme you are fair more likely to take advantage of poor losing traders when your trade closes at a point well away from ‘true value’


I hope this explains a bit more.

Steve.
 
Of course the correct way to manage this type of trade is to use a mental stop which does not trigger an automatic exit from the market. Suppose in our example we set the following loose rules.

1 Mental stop of 20 pips.

2 Hard stop of 40 pips.

What we could do in terms of trade management is this. If our trade goes into loss then we will monitor it closely. If our mental stop is triggered we will move into ‘damage limitation mode’. This means that we have now accepted that our trade is a poor one and we are looking to get out. What we are waiting for is a better price than is currently available. It is possible that the market will move either way. If our mental stop of 20 pips gets hit and then the market retraces 12 pips then we can get out for a loss of 8 – see how remaining in the trade beyond our mental stop has saved us money? In most cases it will! Suppose, instead of retracing 12, the market moves on another 10 so our loss is now 30 – this can and does happen – it is still more than possible that the market will retrace 10 – 15 pips in the short term which still means a loss of less than 20 pips.
I hope that you can see that by taking responsibility for the trade management the losses can be far better controlled. This is because your exit from a trade is likely to be nearer to ‘true value’ than if you have a fixed stop set which is guaranteed to exit you from a trade at a price extreme.
Likewise by placing a limit close at a price extreme you are fair more likely to take advantage of poor losing traders when your trade closes at a point well away from ‘true value’

To be honest, I think this encourages bad trading habits. If you have your mental stop at 20 pips and you don't take it out but wait for a retracement, when will you decide to stop yourself out? Suppose it is currently retracing 10 pips already. You are saying to yourself: "oh nice, maybe I can get out for breakeven". You hold on a little longer, price comes back to breakeven. Now you say to yourself: "I can't exit now, I'm just in touch of a profit". But two seconds later it plunges and you're down 30 pips, about the get taken out by your hard stop.

What I'm trying to say is, that if you have a hard stop at 40 pips and a mental stop at 20 pips, one of them has to be wrong on the account of a stop being the instrument to determine when the market proofs you wrong on your current position. If that 'trigger level' is at 20 pips you should get out at 20 and not wait to get taken out even further along the road. If that 'trigger level' is at 40 pips than you should not get out at 20 pips because it might just drop to 25 before reversing and finishing nicely in profit.

If you are going to determine in real-time when to stop yourself out instead of beforehand - that's basically what you're saying - then there's überhaupt no point in having predetermined stops. Except for those disaster events or should any external factor influence your ability to close out manually - e.g. power shortages etc. And that doesn't sound like taking much responsibility to me.

I can see your point and understand where you're going with this. But here I am, being a contrarian pain in the ass :p
 
Last edited:
In my opinion to trade successfully intra-day we must adopt a policy which takes advantage of the theory. This simply means closing bad trades on a regression whilst only ever taking profits on prices of local extremity. By trading this way we will only ever be exchanging ‘value’ at points which are beneficial to ourselves.

A question for you.

(1) Suppose we are in a long position which already generated nice profits. We are looking to close out the position. Suddenly price seems retrace a bit. We don't want to exit now because we are convinced we could've exited at a better price. So we sit this one out and see that the retracement was actually a reversal. We now need to decide what we do, hold on for a better price higher, or take what we have (still a decent profit) and close out completely.

(2) Suppose we are in a long position and price looks way overextended. We close out completely, only to see price rally a whole lot further. Frustrated that we closed out so early, we think of getting back on, but we manage to resist.

So, what causes the most emotional pain? Having missed some profits, or having to give back some to the markets? For most people, it will be much easier to close out a winning position when it's going in the favourable direction. Closing a winning position when it's going against you, is much harder, although in most cases, that's exactly when you should be doing it.
 
Firewalker – Well made points.

My theories are based around the experiences which I set out in Post #88. In that post I mentioned how I lost, on average, about 104 pips per week on GBPUSD (Before trading costs were further deducted). This was based on an average of around 8 trades per week.

104 / 8 = 13 pips per trade average loss before spreads. That to me spells out the real cost of setting stops. This was amazingly consistent over a fairly long period and therefore I cannot determine that this is a coincidence.

I’m not sure whether you are wholly grasping the concept. I’m not trying to suggest a method where people should be ignorant of knowing when to get out of a bad trade. What I’m trying to point out is the relative costs of employing a system where a rigid stop loss is set on entry. I guess that what I am trying to get across is that a rigid stop loss is a form of panic. In the market panic costs you. You are in effect exchanging your position, at an inferior price, for peace of mind / relaxation.

Your questions are difficult to answer directly given that your overall trading style could be very different to mine.
In Q1 you ask about exiting a trade. If I was long then I would be looking to exit on sharp spike upwards. Sure the market might go on upwards after I exited but you cannot know that this is going to happen. What is important is that your exit is a reaction to a positive movement to price.
Q2 isn’t really a question. I agree with you. As humans we can always have an emotional involvement with a trade even after it is closed. I’m sure that we’ve all looked at prices of instruments after we have closed positions ‘just to see if we were right or wrong’ or to see if we can fathom a scheme which will allow us to trade slightly better next time out. I guess that this is part of trading. What we must understand is that we are never going to get ‘the perfect price’ every time we enter or exit a position so there is little point beating ourselves up over that.

In your Post #103 you say that you think this is promoting bad trading habits – I would have to disagree. What I would acknowledge is that a rigid stop loss system will stop bad trading habits occurring – all I would say however is that a) rigid stops cost you, and b) there are other ways of developing good trading habits. In my opinion rigid stops are not the only way to establish these ‘good habits’.
In your reply you actually disobey the loose rules I mentioned. Once the mental stop is triggered you are looking to exit the trade but slightly more cheaply than the current price. As a rule of thumb on GBP/USD you can normally manage around half the value of the maximum that the scalp has been against you. So if it went 20 against you, triggered the mental stop, then I’d be looking for around a 10 pip retrace. I would certainly not be looking to try and turn my losing trade into a winning one once the mental stop was triggered – if you did that you’d be running the risk of simply ‘running a loser’ which is what we don’t do.
 
correct way to manage this type of trade is to use a mental stop

good post steve

my solution is dynamic stops(just an other option):
1. place initial stops at 2 (or 3) different levels
2. geurilla the stops wisely, one by one, only in my favor to better levels


Of course the correct way to manage this type of trade is to use a mental stop which does not trigger an automatic exit from the market. Suppose in our example we set the following loose rules.

1 Mental stop of 20 pips.
2 Hard stop of 40 pips.
Steve.
 
I have to agree with firewalker. Neither your "poor trader" nor your "good trader" are exercising what I would call trade management, anymore than operating a vehicle while blindfolded and bound is "driving".

Clearly you have put a great deal of thought into this, but there is also a lot of "feeling". And, for both traders, a great deal of hope. Hope and feeling and could and might are not friends of responsible trade management. I understand the point you're trying to make about stops, but placing so much responsibility on stops when the entries and trade management are so poor (on both sides) is a bit like bailing out a sinking dighy.

Trading all in/all out creates all sorts of problems, particularly with regard to relying on hope. Focusing on true value can help, but since true value is in a continuous state of flux, this focus must be likewise flexible. One can also avoid the issue of stops by being in the market at all times, and if one is going to work with channels/envelopes/bands, this course may be just the thing, even though it carries its own set of challenges.

In any case, placing more than one level of stops and hoping that one's losing trade will eventually revert closer to an ever-changing mean is probably not the best choice for he who wants to improve his trading performance. When your poor trader does not exit at his target, his issue is not stops. When your good trader finds himself in a losing trade and hangs onto it in the hope that he will be able to exit at a better price, his issue is likewise not stops.
 
I propose that a simple test is conducted.

1) Spin a coin - Heads we go long, tails we go short.

2) We enter the GBPUSD market in accordance with our coin toss.

3) Our target will be 20 pips.

4) Our 'mental stop loss' will get triggered once the market moves 20 pips against us. Once the mental stop is triggered we will monitor how large the position is ever against us. We will close if the position goes 50 against us or upon the market retracing 50% of the maximum which it was against us.

5) We will enter the next trade at the start of the next 60 minute period after the close of the prior trade.

Any takers?

Steve.
 
I propose that a simple test is conducted.

1) Spin a coin - Heads we go long, tails we go short.

2) We enter the GBPUSD market in accordance with our coin toss.

3) Our target will be 20 pips.

4) Our 'mental stop loss' will get triggered once the market moves 20 pips against us. Once the mental stop is triggered we will monitor how large the position is ever against us. We will close if the position goes 50 against us or upon the market retracing 50% of the maximum which it was against us.

5) We will enter the next trade at the start of the next 60 minute period after the close of the prior trade.

Any takers?

Steve.

So you're making a random entry at a random price rather than at any sort of price extreme or mean?
 
So you're making a random entry at a random price rather than at any sort of price extreme or mean?

Correct - What I want to try and demonstrate is the 'cost of a fixed stop loss' vs 'a managed exit strategy' as I am convinced that it is a fixed stop loss that is of most 'expense' to an intra-day trader. In order to do this I want everything else to be totally random.

Steve.
 
Correct - What I want to try and demonstrate is the 'cost of a fixed stop loss' vs 'a managed exit strategy' as I am convinced that it is a fixed stop loss that is of most 'expense' to an intra-day trader. In order to do this I want everything else to be totally random.

Steve.

Assuming no control of any other variables, how many trials are necessary in order to provide you with a statistically significant result? (I assume that anyone interested in a mechanical system would be interested in this.)
 
Steve, before you are goaded into delivering an empirical test on this, can I suggest you continue with what you have started, in your own way?

You’re bound to get feedback and agreement/disagreement, but regardless of the validity (how would we ever measure that?) of your argument, I sensed in your initial foray, a glint of something very fundamental which you were on the verge of expressing - something with which I have also found to be a rather wriggly beastie.

I’ve gone all the way from tight stops to no stops to stops based on the next higher timeframe’s previous swing high/low (and correspondingly small position size). Each have their own pros and cons. But we’ve all heard that before. I really sense you have a way of expressing this that will be very interesting. ‘Right’ or ‘Wrong’.

Those that need to test this theory before you’ve fully explicated it should do so, themselves, I suggest.

If you were to get to involved in the experimentation, tests and results on a post by post basis of the needs of those who are likely well-meaning, but for whom you have no responsibility (including me), you may well find your original thinking and the development of an idea of potentially great significance, has been snuffed out with the fusty old fire-blanket of tired dogma. And that would be a shame.

Please plough on and do it your way.
 
Assuming no control of any other variables, how many trials are necessary in order to provide you with a statistically significant result? (I assume that anyone interested in a mechanical system would be interested in this.)

Not sure about number of results required. Any ideas on that one?
What I did think however was that 2 results can be recorded for each hour without the need for a coin toss. A result can be recorded for both outcomes. ie a long trade and a short trade.

Just had a quick check of 14:00 / 15:00 / 16:00 / 17:00 hours trades and the results are as follows....

WT = Winning Trade / LT = Losing Trade

14:00 WT +20 LT -12
15:00 WT +20 LT -14.5
16:00 WT +20 LT -12
17:00 WT +20 LT -11

Steve.
 
Not sure about number of results required. Any ideas on that one?
What I did think however was that 2 results can be recorded for each hour without the need for a coin toss. A result can be recorded for both outcomes. ie a long trade and a short trade.

Just had a quick check of 14:00 / 15:00 / 16:00 / 17:00 hours trades and the results are as follows....

WT = Winning Trade / LT = Losing Trade

14:00 WT +20 LT -12
15:00 WT +20 LT -14.5
16:00 WT +20 LT -12
17:00 WT +20 LT -11

Steve.

So in each of the above LTs, the trade went against you by 24, 29, 24, and 22?
 
Guys, you're off on a wrong track with this. The coin toss was a bad idea.

With the data and rules you're proposing, you need to be right 1:1.67332 = 83.66% of the time to come out ahead. A coin toss will give you around 50%.

What was the point?

You can't come down from the abstract to the practical without getting agreement on the abstract. That normally requires going to a higher level of abstraction - not coming down to tossing coins.
 
Guys, you're off on a wrong track with this. The coin toss was a bad idea.

With the data and rules you're proposing, you need to be right 1:1.67332 = 83.66% of the time to come out ahead. A coin toss will give you around 50%.

What was the point?

You can't come down from the abstract to the practical without getting agreement on the abstract. That normally requires going to a higher level of abstraction - not coming down to tossing coins.

I removed the 'coin toss' and replaced it with the production of a result for both outcomes for each hour.

Can you expand on your math with the 83.66%? I've not picked that up correctly.

Many thanks,
Steve.
 
My theories are based around the experiences which I set out in Post #88. In that post I mentioned how I lost, on average, about 104 pips per week on GBPUSD (Before trading costs were further deducted). This was based on an average of around 8 trades per week. 104 / 8 = 13 pips per trade average loss before spreads. That to me spells out the real cost of setting stops.

How you equate a losing trade with the cost of setting a stop is beyond me, but the logic behind it is flawed. For some a stop signal might be a reason to get out, for others it might be a reason to reverse one's position. I don't consider a stop to have an inherent cost, in fact I consider it to be the market's way of showing me I'm wrong.

This was amazingly consistent over a fairly long period and therefore I cannot determine that this is a coincidence.

Correlation does not mean causality.

I’m not sure whether you are wholly grasping the concept. I’m not trying to suggest a method where people should be ignorant of knowing when to get out of a bad trade. What I’m trying to point out is the relative costs of employing a system where a rigid stop loss is set on entry. I guess that what I am trying to get across is that a rigid stop loss is a form of panic. In the market panic costs you. You are in effect exchanging your position, at an inferior price, for peace of mind / relaxation.

Perhaps your view of a rigid stop loss as a form of panic might be because your stops are placed at price extremes, where they exactly should NOT be placed. Although I haven't seen your trades/charts to illustrate this, I suspect this is the case by what you are describing.
 
Interesting debate - can I use traffic lights :whistling

Your entry triggers and gives you a green light

The price goes against you to the extent (pre-determined) that you get an amber light warning that the trade looks like a wrong 'un but that's not certain yet. It's reasonable at this stage to decide to exit but to watch the action closely to take advantage of a better price if it arises.

The price continues against you to the extent (pre-determined) that you get a red light where you must exit, it's certainly a wrong 'un and no further questions to be asked.

good trading

jon
 
Top