DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Stop Loss is not always your friend...
- Thread starter princeplanet
- Start date
- Joined
- 28 January 2009
- Posts
- 235
- Reactions
- 1
Re: Stop Loss is not always your friend....
Yeah, i often read about protecting the downside and often focus so much on it that i forget to keep an eye on the upside. I have practically given up on trailing the stop since so often the move has another 2/3 to go but i got stopped out on the first 1/3. Likewise with B/E stops, great mental relief but get stopped too often to what was a great move. It's a killer to find a balance but these days just stick to original stop and even that can be pain in the a*s sometimes. So many times if the stop was just a tick or two wider i would not be stopped out but that's a whole new lot of mine field.
:grenade:
Yeah, i often read about protecting the downside and often focus so much on it that i forget to keep an eye on the upside. I have practically given up on trailing the stop since so often the move has another 2/3 to go but i got stopped out on the first 1/3. Likewise with B/E stops, great mental relief but get stopped too often to what was a great move. It's a killer to find a balance but these days just stick to original stop and even that can be pain in the a*s sometimes. So many times if the stop was just a tick or two wider i would not be stopped out but that's a whole new lot of mine field.
:grenade:
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
R/Y
Just on reading your post it appears your looking at systematic trading
OR
are you encompassing discretionary trading in the same
bundle with your interpretations of signal and risk management?
Which is correct?
Will have time to devote to this in discussion form tonight.
R/Y
Just on reading your post it appears your looking at systematic trading
OR
are you encompassing discretionary trading in the same
bundle with your interpretations of signal and risk management?
Which is correct?
Will have time to devote to this in discussion form tonight.
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
Hi T/A
There is no correct about this. It's just the way I think about things like this if asked to get a bit serious. I find it is helpful to consider the issue by dividing an investment process into money making and risk management components. I think it is useful in a conversation about stops. I acknowledge that there are different ways to consider the issues and define what represents a contribution to value-add. We could readily be looking at the same thing, have an exactly matching understanding of what we are seeing and yet name different parts of the process different things. Hence, we can agree to agree and yet disagree on why we agree. Rumsfeld had nothing.
You have mentioned that you are a discretionary trader nowadays, albeit with a room full of knowledge and ability from various fields.
A signal is anything which is a directional prediction. Hence discretionary predictions are signals in their own right. You do not even have to have a good reason. "Oh, I reckon it's Monday and that's an up-day" is a buy signal on a discretionary basis, for example. That says absolutely nothing about how good a signal it is.
One thing about time stops. A signal almost always has a life-span. You may find, for example, that the discretionary call tends to be strongest for a day, has less power over a week and is essentially defunct over a two week period. That's still a signal - the call that a stock will move in a certain direction. It just has a time frame for operation. All signals are like this. They don't forecast over an infinite time frame. No stop is required to implement this. The call just becomes weaker with time and can shut itself down at a particular time after making the call. I have just introduced a deeper concept here. There is the directional prediction and the confidence in that prediction. A signal is actually a combination of both. Not necessarily just Buy/Sell but possibly also a spectrum between the High-Conviction Buy and Sell extremes. Hence, as your predictive ability declines with time, we can say that I expect a stock to go up, but that my confidence wanes with the passage of time with new information changing the landscape. Alternatively, you can say that my signal is only good till close. Same deal, different clothes. In other words, a time stop is actually part of a signal and we should set it apart from risk management if this framework is to be utilised.
Cheers
RY
Hi T/A
There is no correct about this. It's just the way I think about things like this if asked to get a bit serious. I find it is helpful to consider the issue by dividing an investment process into money making and risk management components. I think it is useful in a conversation about stops. I acknowledge that there are different ways to consider the issues and define what represents a contribution to value-add. We could readily be looking at the same thing, have an exactly matching understanding of what we are seeing and yet name different parts of the process different things. Hence, we can agree to agree and yet disagree on why we agree. Rumsfeld had nothing.
You have mentioned that you are a discretionary trader nowadays, albeit with a room full of knowledge and ability from various fields.
A signal is anything which is a directional prediction. Hence discretionary predictions are signals in their own right. You do not even have to have a good reason. "Oh, I reckon it's Monday and that's an up-day" is a buy signal on a discretionary basis, for example. That says absolutely nothing about how good a signal it is.
One thing about time stops. A signal almost always has a life-span. You may find, for example, that the discretionary call tends to be strongest for a day, has less power over a week and is essentially defunct over a two week period. That's still a signal - the call that a stock will move in a certain direction. It just has a time frame for operation. All signals are like this. They don't forecast over an infinite time frame. No stop is required to implement this. The call just becomes weaker with time and can shut itself down at a particular time after making the call. I have just introduced a deeper concept here. There is the directional prediction and the confidence in that prediction. A signal is actually a combination of both. Not necessarily just Buy/Sell but possibly also a spectrum between the High-Conviction Buy and Sell extremes. Hence, as your predictive ability declines with time, we can say that I expect a stock to go up, but that my confidence wanes with the passage of time with new information changing the landscape. Alternatively, you can say that my signal is only good till close. Same deal, different clothes. In other words, a time stop is actually part of a signal and we should set it apart from risk management if this framework is to be utilised.
Cheers
RY
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
So far so good. I prefer to use the word anticipation as I don't actually know where I'm going to exit (At the time of entering) but for ease of description---
Broadly I think its more than that --- various risks(Capital/Time/Portfolio/Opportunity/Trade/Market Liquidity and believe it or not Macro Fundamental risk)---to which effect what I do.
I think it a poor example but I presume you mean a which has no directly logical reason for an entry in a stock to be taken.
Again a broad brush.
If completely random I agree---I think---I haven't tested your theory to give me an outcome---but I have seen evidence of a simple filter and then a random entry having a positive expectancy.
I remember seeing a test once which returned a positive expectancy where all members of the data set were showing a positive up trend. The buy signal was Thursday. The results were improved with your explanation of Money Management. I didn't keep it but would have been perfect for this argument.There are other tests I have seen based on nothing more than a time of day. While not brilliant they returned a positive expectancy.
I'm sure they are out there and I'm sure they could be replicated---I don't have the time or inclination to do so---but if I find one ill be sure to post it up.
Yes agree with regard to expected outcome. A combination of risk minimizing factors can have quite an effect on trading results.
Broadly in agreement I guess.
So far so good. I prefer to use the word anticipation as I don't actually know where I'm going to exit (At the time of entering) but for ease of description---
Broadly I think its more than that --- various risks(Capital/Time/Portfolio/Opportunity/Trade/Market Liquidity and believe it or not Macro Fundamental risk)---to which effect what I do.
. In itself no but as a variable yes in the contexts of Risk/s I look at I think it does help It may skew a method from loss to profit---in the context of Risk Factors i consider.
I think it a poor example but I presume you mean a which has no directly logical reason for an entry in a stock to be taken.
Again a broad brush.
If completely random I agree---I think---I haven't tested your theory to give me an outcome---but I have seen evidence of a simple filter and then a random entry having a positive expectancy.
I remember seeing a test once which returned a positive expectancy where all members of the data set were showing a positive up trend. The buy signal was Thursday. The results were improved with your explanation of Money Management. I didn't keep it but would have been perfect for this argument.There are other tests I have seen based on nothing more than a time of day. While not brilliant they returned a positive expectancy.
I'm sure they are out there and I'm sure they could be replicated---I don't have the time or inclination to do so---but if I find one ill be sure to post it up.
Yes agree with regard to expected outcome. A combination of risk minimizing factors can have quite an effect on trading results.
Broadly in agreement I guess.
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
Heck Tech (A)...I think we agree to broadly agree. That's a relief.
Anyhow, to your recollection of Thursdays as a signal...
There are probably as many rules that make money as there are grains of sand...in the past. Vanstone with his neural net could probably generate a fair chunk of them by changing inputs and the internal structure of the net as well as the training methods. What happened in the past is interesting like history of the American Revolution is interesting. However, saying that it is predictive is an entirely different thing. There is a vast gap.
There actually is a daily seasonal for the ASX 200, for example, that is statistically significant. Fridays are significantly different. However, the gap is so small you cannot trade it in terms of arbitrage.
I, or KTP amongst others, could generate a MA type signal where the two variables are short and long term averaging periods. By varying these over the infinite space, a meaningful parts of the possible combinations will appear profitable. Run it through with those same fixed variables and things will not stay the same. The underlying can be anything that moves. We can generalize to any type of rule making process. Some combination of input variables will produce profit for anything approximating an attempt at a decision rule. That is a given and there is nothing special about the decision rule just because some profitable combination of variables exist. What matters is whether that decision structure and set of parameters has any reason at all for being useful on a forward looking basis.
On other bits:
Static, to people like me, means random noise. Flipping coins, dice, pulling numbers out of a barrel...as a basis for forecasting/prediction. There is absolutely nothing in it that is predictive, but it is a signal anyway.
Heck Tech (A)...I think we agree to broadly agree. That's a relief.
Anyhow, to your recollection of Thursdays as a signal...
There are probably as many rules that make money as there are grains of sand...in the past. Vanstone with his neural net could probably generate a fair chunk of them by changing inputs and the internal structure of the net as well as the training methods. What happened in the past is interesting like history of the American Revolution is interesting. However, saying that it is predictive is an entirely different thing. There is a vast gap.
There actually is a daily seasonal for the ASX 200, for example, that is statistically significant. Fridays are significantly different. However, the gap is so small you cannot trade it in terms of arbitrage.
I, or KTP amongst others, could generate a MA type signal where the two variables are short and long term averaging periods. By varying these over the infinite space, a meaningful parts of the possible combinations will appear profitable. Run it through with those same fixed variables and things will not stay the same. The underlying can be anything that moves. We can generalize to any type of rule making process. Some combination of input variables will produce profit for anything approximating an attempt at a decision rule. That is a given and there is nothing special about the decision rule just because some profitable combination of variables exist. What matters is whether that decision structure and set of parameters has any reason at all for being useful on a forward looking basis.
On other bits:
Static, to people like me, means random noise. Flipping coins, dice, pulling numbers out of a barrel...as a basis for forecasting/prediction. There is absolutely nothing in it that is predictive, but it is a signal anyway.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
Yes---and your point being?
Yes---and your point being?
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
Stop loss is not expected to be your friend.....on average, through time.
Stop loss is not expected to be your friend.....on average, through time.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
Hmm
Don't think it's as clear cut as that
Your implying that unless you have an
Entry that is better than random then
Any sort of stop is simply pointless
The entry is the focus
Am I reading that correctly ?
Hmm
Don't think it's as clear cut as that
Your implying that unless you have an
Entry that is better than random then
Any sort of stop is simply pointless
The entry is the focus
Am I reading that correctly ?
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
On average, there is zero alpha. Trading is generally regarded as extracting alpha given the time frames are very short and risk from trading completely dominates anything else. When you have zero alpha, which is the population average alpha (ie no idea what the entry points should be) then no amount of risk management is going to make you money.
If you have got an idea of what superior entry points should be, the addition of stops will actually decrease your expected return.
On the other hand, if you are a wanton value destroyer (essentially taking the opposite positions of the people with an idea of what is going on), then stop losses will add value by helping you to stop losing value. The stop should be so tight that you never trade.
The relationship between having stops at the entry point when you are a value destroyer and having no stops when you are a value creator should be evident.
If your entry points are selected by chook raffle or listening to static electricity on your AM band radio, having super-tight stops or no stops produces the same expected outcome....zero, before expenses.
The above is only talking about money making and says nothing about risk adjusted returns. Stops do not add value on a population average in these terms.
It is reasonably argued that they can add value in psychological ways by helping you sleep at night. Further, if you do not value a loss and gain as equals, it can help on that. However, these are issues of psychology and not outright money making.
The original statement I made was that some tech practitioners, including those who publish and are widely followed in discussion and courses, seem to think adding stops makes you money just by doing it (on average, through time). It does not. Individual outcomes may yield that result as tossing coins will yield profit sometimes, but there is no systematic profit to it. That's all I'm claiming on this thread.
On average, there is zero alpha. Trading is generally regarded as extracting alpha given the time frames are very short and risk from trading completely dominates anything else. When you have zero alpha, which is the population average alpha (ie no idea what the entry points should be) then no amount of risk management is going to make you money.
If you have got an idea of what superior entry points should be, the addition of stops will actually decrease your expected return.
On the other hand, if you are a wanton value destroyer (essentially taking the opposite positions of the people with an idea of what is going on), then stop losses will add value by helping you to stop losing value. The stop should be so tight that you never trade.
The relationship between having stops at the entry point when you are a value destroyer and having no stops when you are a value creator should be evident.
If your entry points are selected by chook raffle or listening to static electricity on your AM band radio, having super-tight stops or no stops produces the same expected outcome....zero, before expenses.
The above is only talking about money making and says nothing about risk adjusted returns. Stops do not add value on a population average in these terms.
It is reasonably argued that they can add value in psychological ways by helping you sleep at night. Further, if you do not value a loss and gain as equals, it can help on that. However, these are issues of psychology and not outright money making.
The original statement I made was that some tech practitioners, including those who publish and are widely followed in discussion and courses, seem to think adding stops makes you money just by doing it (on average, through time). It does not. Individual outcomes may yield that result as tossing coins will yield profit sometimes, but there is no systematic profit to it. That's all I'm claiming on this thread.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
You are benchmarking this against what?
Any profit at all?
You know this how?
Sorry I don't see this having meaning---is it a sarcastic comment?
Well they're not--what is a value destroyer--how do I destroy value?
What's is a value creator--how do I create value?
Hardly an argument of value when looking at whether stops are important in helping skew your numbers your way---as I don't know of anyone on the planet who uses such entries??
Again a broad brush---how do you know this?
Agree
Sorry don't agree.
If we are to let profit's run and cut losses short one way is to introduce trailing stops.
I don't know of anyone who says this---can you give an example?
So back to random.
Questions.
(1) Can the inclusion of stops(any of Initial stop/Breakeven stop/Trailing stop---blah blah---skew a method which has an entry which is designed to identify a potential movement in the direction of a desired trade---more toward a profit or in fact increase profit from a method where an exit is the only way to close out a trade---in your view??
(2) Are stops a complete waste of time---any stop in any situation?
You are benchmarking this against what?
Any profit at all?
You know this how?
.
Sorry I don't see this having meaning---is it a sarcastic comment?
.
Well they're not--what is a value destroyer--how do I destroy value?
What's is a value creator--how do I create value?
Hardly an argument of value when looking at whether stops are important in helping skew your numbers your way---as I don't know of anyone on the planet who uses such entries??
Again a broad brush---how do you know this?
.
Agree
Sorry don't agree.
If we are to let profit's run and cut losses short one way is to introduce trailing stops.
I don't know of anyone who says this---can you give an example?
So back to random.
Questions.
(1) Can the inclusion of stops(any of Initial stop/Breakeven stop/Trailing stop---blah blah---skew a method which has an entry which is designed to identify a potential movement in the direction of a desired trade---more toward a profit or in fact increase profit from a method where an exit is the only way to close out a trade---in your view??
(2) Are stops a complete waste of time---any stop in any situation?
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
Cash
Options pricing.
Not attempting to be sarcastic. If you (this is a general reference, as in plural. Not targeted at you) are actually hindered and do the opposite of making money (behavioural biases can cause this) then you are better off not trading at all. Stops add value to these kinds of people by stopping them from trading at all.
I should add, all of this relates to the kinds of stops that reduce position size.
Value is defined as making money over cash. You create value by having predictive ability and deploying it into markets. This can come from certain types of technical analysis, fundamental analysis, combinations, information arbitrage...anything which has an effective edge in the market.
Static and coin tosses, monkeys with darts...are all analogies (albeit doable) for a prediction process that has no value. People do fundamental and technical analysis. Just because they do doesn't mean that this is a license to print money. The value of most of it is actually zero.
It is a tautology. No maths is required or any simulations.
If your position is expected to go up and you have predictive ability, stopping out via trailing stop destroys value on average through time. Options pricing.
Mr NR. "Successful Stock Trading - A Guide to Profitability". Chapter 1. An example is given relating to an Excel simulation which leads on to all sorts of stuff on win/loss ratio at the expense of winning percentage. You can make of it what you will. Clearly a lot of people believe this. It might be fruitful to ask Kris to check the maths behind the simulation and the claims. I will not comment further.
The average player is effectively random in alpha terms. It is a tautology.
Questions.
Adding stops etc. changes the distribution of expected outcomes. You can skew it anyway you like. If you have skill, the kind of stop which involves cutting positions will net destroy value in terms of money making expectations. For example, you can start with an initial stop loss and roll up to B/E. If you have skill, this will lose you money. It will, however, do this by producing a string of small losses and a smaller number of gains, some of which can be large.
No. They are valuable. I am only saying they don't make money on average.
Using them in the manner you described increases the chances that you will make money over the longer term if you have skill. It does this at a cost to your expected return at the outset.
If you have liquidity issues or leverage issues, their value is present just to keep you alive, if you have an edge.
Nothing can save you in the longer term if you have no predictive value or are a wanton value destroyer (generic).
-----
All I am saying is that stops do not make a person money in and of themselves on average, through time. That's it.
Cash
Options pricing.
Not attempting to be sarcastic. If you (this is a general reference, as in plural. Not targeted at you) are actually hindered and do the opposite of making money (behavioural biases can cause this) then you are better off not trading at all. Stops add value to these kinds of people by stopping them from trading at all.
I should add, all of this relates to the kinds of stops that reduce position size.
Value is defined as making money over cash. You create value by having predictive ability and deploying it into markets. This can come from certain types of technical analysis, fundamental analysis, combinations, information arbitrage...anything which has an effective edge in the market.
Static and coin tosses, monkeys with darts...are all analogies (albeit doable) for a prediction process that has no value. People do fundamental and technical analysis. Just because they do doesn't mean that this is a license to print money. The value of most of it is actually zero.
It is a tautology. No maths is required or any simulations.
If your position is expected to go up and you have predictive ability, stopping out via trailing stop destroys value on average through time. Options pricing.
Mr NR. "Successful Stock Trading - A Guide to Profitability". Chapter 1. An example is given relating to an Excel simulation which leads on to all sorts of stuff on win/loss ratio at the expense of winning percentage. You can make of it what you will. Clearly a lot of people believe this. It might be fruitful to ask Kris to check the maths behind the simulation and the claims. I will not comment further.
The average player is effectively random in alpha terms. It is a tautology.
Questions.
Adding stops etc. changes the distribution of expected outcomes. You can skew it anyway you like. If you have skill, the kind of stop which involves cutting positions will net destroy value in terms of money making expectations. For example, you can start with an initial stop loss and roll up to B/E. If you have skill, this will lose you money. It will, however, do this by producing a string of small losses and a smaller number of gains, some of which can be large.
No. They are valuable. I am only saying they don't make money on average.
Using them in the manner you described increases the chances that you will make money over the longer term if you have skill. It does this at a cost to your expected return at the outset.
If you have liquidity issues or leverage issues, their value is present just to keep you alive, if you have an edge.
Nothing can save you in the longer term if you have no predictive value or are a wanton value destroyer (generic).
-----
All I am saying is that stops do not make a person money in and of themselves on average, through time. That's it.
- Joined
- 30 September 2012
- Posts
- 741
- Reactions
- 373
Re: Stop Loss is not always your friend....
Agree - there's no systematic profit to stop losses. I'm not sure how someone could argue otherwise.
Potential for stop losses to reduce volatility & drawdown is possibly worth looking at. I've avoided them completely until recently, but only in relation to volatility (and really, drawdown) reduction. Some market states will see that costing you some profit though. For me it's a personal decision re: whether the risk of profit reduction is greater or less than the potential for draw down reduction.
Agree - there's no systematic profit to stop losses. I'm not sure how someone could argue otherwise.
Potential for stop losses to reduce volatility & drawdown is possibly worth looking at. I've avoided them completely until recently, but only in relation to volatility (and really, drawdown) reduction. Some market states will see that costing you some profit though. For me it's a personal decision re: whether the risk of profit reduction is greater or less than the potential for draw down reduction.
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
- Joined
- 1 May 2007
- Posts
- 2,904
- Reactions
- 52
Re: Stop Loss is not always your friend....
I would argue that if say you are an amazing trader with great expectancy, and your stops are based on signals rather than fixed x% away, you are actually adding value as the stop is in effect another 'trade'.
I would argue that if say you are an amazing trader with great expectancy, and your stops are based on signals rather than fixed x% away, you are actually adding value as the stop is in effect another 'trade'.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
Ok about 2%
I've lost the value of introducing a completely different instrument to the discussion of stops in Stock Trading.
Wow that's really encompassing everyone with zero knowledge.
While some gamble the stock market most wouldn't use any sizable funds. I really don't think this discussion was ever meant to encompass total novices---well I didn't.
Hang on an arbitrary Ill buy 10000 CBA is a stop in itself.
Another example.
I can have No stop and at the point of X place a stop at B/E
Well with the experience I have with Systems testing I have seen improvement in all with the addition of stops. (Tautology)
Yes I agree that most actually lose---but I wouldn't say it could be directly related to placing stops.
Ok
Again I cant see the benefit of comparing with options.
You have all sorts of stuff to consider with options--time decay in the money out of the Money At the money options. How does this help in the discussion.---are we talking options?
Ill have a look. But why is it CLEAR a lot of people believe anything?
Negative I would suggest---net losers.
Questions.
Disagree
My experience is that the END reward to Risk is actually increased as we have not only really small losses we also have No loss (other than Brokerage) on losses and accumulated winners (Profit) far exceeds Accumulated Losses.
In fact 1 win can demolish many B/E stop outs and many small losses. The number of trades required to claw back a string of losses is far less than if I had a stop only and way way less if I didn't have one at all.
----Tautology
Perhaps in some circumstances but the exact opposite in others.
Yes can be the case but need not be a sole reason
If I read this as complete gambling novice --yes.
-----
Ok but I don't think this discussion has helped those who think they have an edge ---don't use stops and have found that their portfolio is anything but impressive due to a few really destructive losses.
EG GALUMAY
Ok about 2%
I've lost the value of introducing a completely different instrument to the discussion of stops in Stock Trading.
Wow that's really encompassing everyone with zero knowledge.
While some gamble the stock market most wouldn't use any sizable funds. I really don't think this discussion was ever meant to encompass total novices---well I didn't.
Hang on an arbitrary Ill buy 10000 CBA is a stop in itself.
Another example.
I can have No stop and at the point of X place a stop at B/E
Well with the experience I have with Systems testing I have seen improvement in all with the addition of stops. (Tautology)
Yes I agree that most actually lose---but I wouldn't say it could be directly related to placing stops.
.
Ok
Again I cant see the benefit of comparing with options.
You have all sorts of stuff to consider with options--time decay in the money out of the Money At the money options. How does this help in the discussion.---are we talking options?
Ill have a look. But why is it CLEAR a lot of people believe anything?
Negative I would suggest---net losers.
Questions.
Disagree
My experience is that the END reward to Risk is actually increased as we have not only really small losses we also have No loss (other than Brokerage) on losses and accumulated winners (Profit) far exceeds Accumulated Losses.
In fact 1 win can demolish many B/E stop outs and many small losses. The number of trades required to claw back a string of losses is far less than if I had a stop only and way way less if I didn't have one at all.
----Tautology
Perhaps in some circumstances but the exact opposite in others.
Yes can be the case but need not be a sole reason
If I read this as complete gambling novice --yes.
-----
Ok but I don't think this discussion has helped those who think they have an edge ---don't use stops and have found that their portfolio is anything but impressive due to a few really destructive losses.
EG GALUMAY
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
With that caveat...'and your stops are based on signals'...I am on the same page. This is where it comes to a definition. Stops as you propose are actually predictions as they are based on informative signals. It's like the reverse of an initiation that was formed via informed signal. You are just initiating another informed trade which offsets the first one. In the context of the recent exchange where stops are expressed as risk management only (ie. no predictive ability in an of themselves), they don't add value.
If you have predictive ability, skyQ is the limit.
With that caveat...'and your stops are based on signals'...I am on the same page. This is where it comes to a definition. Stops as you propose are actually predictions as they are based on informative signals. It's like the reverse of an initiation that was formed via informed signal. You are just initiating another informed trade which offsets the first one. In the context of the recent exchange where stops are expressed as risk management only (ie. no predictive ability in an of themselves), they don't add value.
If you have predictive ability, skyQ is the limit.
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
A stop is a binary put option with the stock as an underlying. It is insurance.
You don't have to be a novice to be a value destroyer or have static as your prediction. Bogle is hardly a novice or lacking in knowledge, for example. He led Vaguard. They don't exactly have small AUM. Buffett will move all his estate into index funds and cash. Hardly the action of an inexperienced person, I would think.
When a person datamines, they will get these results. I can find profit with stops in backtests of static electricity and coin flips of any length. It's no skill whatsoever to find this. Stops are not expected to add value on average through time for the market as a whole. That is a tautology. It is also a tautology that if you torture data long enough, it will say whatever you want. Stops decision rules will be found that work on any dataset including white noise and stops will also be found that destroy value. It's no big deal at all, it exists in everything, and their existence says nothing about making money.
Most actually lose because alpha is zero sum less expenses. Stops do not add value on average through time. They don't subtract value on average through time either.
Questions.
That set of examples is exactly the belief that has been espoused in the prior referred book. This statement is exactly why I believe people believe in this argument. Please ask Kris to check the maths or Vanstone. Or get someone on this site who knows about option pricing using binomial lattices or some other risk-neutral stochastic model to jump in. When you buy an option, is it free? I will say no more.
Yes. However, any single draw will yield a bunch of different results. One bad trade does not make you an idiot. One good one will see people strut around, but doesn't actually make them good. We need to work in expectations. Talking about possibilities without attaching probabilities is worth very little short of saying "you could lose all you money" or "you could become seriously rich". It is a known behavioral bias that we focus on attention-grabbing scenarios to the detriment of attaching probability. Both are important to make an informed decision.
What matters is whether a person actually has an edge. What a person thinks only matters if they can actually back it with outcome. If you (generic) don't know that you don't know, you're cactus. If you know that you don't know, at least you know something and can build from there. The 'mortality rate' of posters since the inception of ASF might provide some evidence that belief in personal skill and actual possession of skill that matches that belief are somewhat weakly related.
The posts that Gulamay has made in terms of hard data is insufficient to make any strong assertions from my point of view.
----
Tech/A. Stops do not add value in expectations terms. They will not add value in and of themselves on average, through time. That's it. I might step off now as I think I've said all I'd like to about the main aspects of this point. Seems like we do have some common ground though and that is good to find. On the rest, I guess we can agree to disagree.
Systematic and SkyQuake made very specific and nuanced statements which tell me they are familiar with the arguments and issues.
Thanks a lot for the exchange.
A stop is a binary put option with the stock as an underlying. It is insurance.
You don't have to be a novice to be a value destroyer or have static as your prediction. Bogle is hardly a novice or lacking in knowledge, for example. He led Vaguard. They don't exactly have small AUM. Buffett will move all his estate into index funds and cash. Hardly the action of an inexperienced person, I would think.
Exactly. All of these are arbitrary random positions. No information, no expected return.
When a person datamines, they will get these results. I can find profit with stops in backtests of static electricity and coin flips of any length. It's no skill whatsoever to find this. Stops are not expected to add value on average through time for the market as a whole. That is a tautology. It is also a tautology that if you torture data long enough, it will say whatever you want. Stops decision rules will be found that work on any dataset including white noise and stops will also be found that destroy value. It's no big deal at all, it exists in everything, and their existence says nothing about making money.
Most actually lose because alpha is zero sum less expenses. Stops do not add value on average through time. They don't subtract value on average through time either.
Stops actually are options.
I read ASF.
After expenses, yes. Before expenses, no.
Questions.
That's a tautology if your world is devoid of probability and focuses on selected outcomes only. However, the very fabric of the universe is probabilistic. You will find these things occur, but the expectations inferred in terms of money making ability do not follow.
That set of examples is exactly the belief that has been espoused in the prior referred book. This statement is exactly why I believe people believe in this argument. Please ask Kris to check the maths or Vanstone. Or get someone on this site who knows about option pricing using binomial lattices or some other risk-neutral stochastic model to jump in. When you buy an option, is it free? I will say no more.
Yes. However, any single draw will yield a bunch of different results. One bad trade does not make you an idiot. One good one will see people strut around, but doesn't actually make them good. We need to work in expectations. Talking about possibilities without attaching probabilities is worth very little short of saying "you could lose all you money" or "you could become seriously rich". It is a known behavioral bias that we focus on attention-grabbing scenarios to the detriment of attaching probability. Both are important to make an informed decision.
What matters is whether a person actually has an edge. What a person thinks only matters if they can actually back it with outcome. If you (generic) don't know that you don't know, you're cactus. If you know that you don't know, at least you know something and can build from there. The 'mortality rate' of posters since the inception of ASF might provide some evidence that belief in personal skill and actual possession of skill that matches that belief are somewhat weakly related.
The posts that Gulamay has made in terms of hard data is insufficient to make any strong assertions from my point of view.
----
Tech/A. Stops do not add value in expectations terms. They will not add value in and of themselves on average, through time. That's it. I might step off now as I think I've said all I'd like to about the main aspects of this point. Seems like we do have some common ground though and that is good to find. On the rest, I guess we can agree to disagree.
Systematic and SkyQuake made very specific and nuanced statements which tell me they are familiar with the arguments and issues.
Thanks a lot for the exchange.
DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Re: Stop Loss is not always your friend....
My bad. It's just a regular option payoff, although it exercises on touch and is of unstated maturity. It's still a form of insurance.
My bad. It's just a regular option payoff, although it exercises on touch and is of unstated maturity. It's still a form of insurance.
tech/a
No Ordinary Duck
- Joined
- 14 October 2004
- Posts
- 20,400
- Reactions
- 6,317
Re: Stop Loss is not always your friend....
But that's where a stop is placed---where you know clearly your analysis is wrong????
This is something you have adopted and was never expressed by the Duck!
I think you've created an argument around a topic which you have designed.
Not enough time today or tonight to continue but-----But that's where a stop is placed---where you know clearly your analysis is wrong???? That's why its there!
You don't take a random trade then place a random get out of jail point.
Some do ---gamblers-----and end up with bottom draw stocks---to afraid to cop it sweet.
I certainly take issue with some of your statements
particularly when as someone who initially said you liked hard evidence you have defaulted to tautology.
While you may disagree I certainly don't agree.
But that's where a stop is placed---where you know clearly your analysis is wrong????
.
This is something you have adopted and was never expressed by the Duck!
I think you've created an argument around a topic which you have designed.
Not enough time today or tonight to continue but-----But that's where a stop is placed---where you know clearly your analysis is wrong???? That's why its there!
You don't take a random trade then place a random get out of jail point.
Some do ---gamblers-----and end up with bottom draw stocks---to afraid to cop it sweet.
I certainly take issue with some of your statements
particularly when as someone who initially said you liked hard evidence you have defaulted to tautology.
While you may disagree I certainly don't agree.
