IS this Karen Supertrader story legit?

[rb107]

Yes, and I do those trades also. It works just fine until the underlying gaps beyond the one or two deviations and wipes out what you've gained in the past 20 trades. A 5% chance is after all, a 1 in 20 probability. It has happened to me in the past year. Gotta watch for those outliers on the bell curve..

I note most pros always have protection in the form of a buy further out. That does no good after the fact, of course...

If you keep your position size small, you will not get wiped out. I trade the SPX and while I've had positions challenged, in most cases (for the Weekly SPX) it expires OTM. The one occasion where it did not (over 5 yrs), I lost a small amount.

Besides, using the Expected Move (and not the Option Chain skew), you can get Prob of Profit of over 95%. The premium is good when using a wide spread (I use $25). With that kind of spread, the probability of a max loss is less than 1% (typically less than 0.1%).

 

[Shakone]

Love the % figures being thrown around here without any specifics. Don't believe them of course.

 

[rb107]

Love the % figures being thrown around here without any specifics. Don't believe them of course.

The specifics, Shakone, is the Probability model which provides the std dev (and the percentages) based on IV. You are, of course, free to believe what ever you want.

 

[Shakone]

The specifics, Shakone, is the Probability model which provides the std dev (and the percentages) based on IV. You are, of course, free to believe what ever you want.

Are you talking about standard deviations from a mean under a normal distribution, and the probability of falling within those deviations?

 

[rb107]

Are you talking about standard deviations from a mean under a normal distribution, and the probability of falling within those deviations?

Yes I am. And based on the percentage daily change in the SPX (specifically), the price action is predominantly random. This is always another point of controversy amongst traders.

 

[Shakone]

Yes I am. And based on the percentage daily change in the SPX (specifically), the price action is predominantly random. This is always another point of controversy amongst traders.

So you believe the SPX is normally distributed do you?

 

[rb107]

So you believe the SPX is normally distributed do you?

Yes. But it's not just opinion; it's based on research that I and many others have performed.

I suspect that you do not. If so, I would ask what you base your opinion on. If your response is simply from experience, then it's still opinion. Besides, I've been trading since the '70s, and my experience has led me to test my opinion.

 

[Shakone]

Yes. But it's not just opinion; it's based on research that I and many others have performed.

I suspect that you do not. If so, I would ask what you base your opinion on. If your response is simply from experience, then it's still opinion. Besides, I've been trading since the '70s, and my experience has led me to test my opinion.

That's interesting, since academic research suggests that the normal distribution is a very poor fit for the S&P over the past 50 or so years.

You should publish your research. Academics would be very keen on this, since the majority seem to love the random walk normal distribution idea.

My research indicates that it's not normal, because it moves multiple standard deviations away from the 'mean' significantly more often than one would expect from a normal distribution. And therefore since it is not normal by my research, all these probabilities you mentioned earlier are wrong.

 

[rb107]

That's interesting, since academic research suggests that the normal distribution is a very poor fit for the S&P over the past 50 or so years.

You should publish your research. Academics would be very keen on this, since the majority seem to love the random walk normal distribution idea.

My research indicates that it's not normal, because it moves multiple standard deviations away from the 'mean' significantly more often than one would expect from a normal distribution. And therefore since it is not normal by my research, all these probabilities you mentioned earlier are wrong.

Actually, independent research over the decades have proven that the price action of the SPX is random on a daily basis.

In fact, Eugene Fama, Llars Peter Hansen, and Robert Shiller were awarded the Nobel Price in Economics recently. Their work showed that asset prices move randomly in the short term.

Now, if you want my research I will be happy to share it. It shows that the Skew of the SPX (tested over a 5-year period) is well below 1.0, which is an indicator of randomness (with the 5-year average of -0.053). Just send me an email and I will attach the results.

Now recall, I said it's predominantly random. This does not exclude the occasional outliers, due to unexpected events, that lead to large tails relative to a purely random action. Do these outliers mean the price action is not normally distributed? No.

Now, since I am willing to share the research, perhaps you would like to share yours.

 

[Shakone]

Actually, independent research over the decades have proven that the price action of the SPX is random on a daily basis.

In fact, Eugene Fama, Llars Peter Hansen, and Robert Shiller were awarded the Nobel Price in Economics recently. Their work showed that asset prices move randomly in the short term.

Now, if you want my research I will be happy to share it. It shows that the Skew of the SPX (tested over a 5-year period) is well below 1.0, which is an indicator of randomness (with the 5-year average of -0.053). Just send me an email and I will attach the results.

Do you understand the difference between random, and normally distributed? I have no doubt that it is random, but it is not normally distributed.

 

[rb107]

Do you understand the difference between random, and normally distributed? I have no doubt that it is random, but it is not normally distributed.

I believe I do. Randomness is based on Brownian motion which exhibits a normal distribution since the universality between Brownian motion and a normal distribution are closely linked. See http://www.trade2win.com/boards/newreply.php?do=newreply&p=2207118

In plain English, random motions are normally distributed. This is why the Probability Model applies.

In my research, I have plotted the random distribution of the SPX (measurement based on skew, which measures the difference between a normal distribution and the actual distribution of the price action) and it certainly closely resembles a normal distribution.

Again, I have my research which I can share with anyone interested. Is your research available?

 

[Shakone]

I believe I do. Randomness is based on Brownian motion which exhibits a normal distribution since the universality between Brownian motion and a normal distribution are closely linked. See http://www.trade2win.com/boards/newreply.php?do=newreply&p=2207118

In plain English, random motions are normally distributed. This is why the Probability Model applies.

No then. You don't understand anything at all.

Things can be random. We can take results and the results can be distributed a certain way. A common one is the bell-curve or normal distribution. But this is not the only distribution.

As an example there is log-normal distribution, Poisson distribution, Exponential distribution, Cauchy distribution etc. To be honest I think there are hundreds of distributions with different names and I don't know them all. So random does not mean normal distribution.

The fact that there are other distributions should be enough to indicate to you that something random does not need to be normally distributed. Do you get this or not? Why would anyone refer to these distributions if everything random were normal? To believe this indicates a serious flaw in your understanding of randomness and probability.

Stop quoting probabilities that you haven't a clue about, and don't give me BS research that you claim to have done from the 70's. Randomness is not based on Brownian motion. Brownian motion is a stochastic process which is normally distributed with mean 0, variance t and has some other properties. The S&P is not a Brownian motion.

Stop pretending. If you don't understand, ask for help. But don't give advice about probabilities when you don't have a clue about what you're talking about because this is downright dangerous to anyone who will listen to you.

 

[rb107]

No then. You don't understand anything at all.

Things can be random. We can take results and the results can be distributed a certain way. A common one is the bell-curve or normal distribution. But this is not the only distribution.

As an example there is log-normal distribution, Poisson distribution, Exponential distribution, Cauchy distribution etc. To be honest I think there are hundreds of distributions with different names and I don't know them all. So random does not mean normal distribution.

The fact that there are other distributions should be enough to indicate to you that something random does not need to be normally distributed. Do you get this or not? Why would anyone refer to these distributions if everything random were normal? To believe this indicates a serious flaw in your understanding of randomness and probability.

Stop quoting probabilities that you haven't a clue about, and don't give me BS research that you claim to have done from the 70's. Randomness is not based on Brownian motion. Brownian motion is a stochastic process which is normally distributed with mean 0, variance t and has some other properties. The S&P is not a Brownian motion.

Stop being a fraud. If you don't understand, ask for help. But don't give advice about probabilities when you don't have a clue about what you're talking about because this is downright dangerous to anyone who will listen to you.

Ok. Now we get into irrational responses. Randomness is based on Brownian motion which exhibits a normal distribution. The fact that there are numerous other distributions is not proof that randomness is uniquely Brownian (or that the SPX does not, in your opinion, exhibit Brownian motion).

Now my research was conducted recently; and as I said, I measured the price action of the SPX against a normal distribution (skew). It doesn't get more straight-forward than that.

My research is not unique; it has been performed by many other researchers as proof that price action is random and closely resembles a normal bell curve.

Since you are not willing to share your research (which you claimed you performed), then I suspect you really haven't performed anything, and this is simply an opinion on your part.

Perhaps you can share your background in statistics (your training) which would lead anyone to believe your opinion is correct.

 

[Shakone]

Ok. Now we get into irrational responses. Randomness is based on Brownian motion which exhibits a normal distribution. The fact that there are numerous other distributions is not proof that randomness is uniquely Brownian (or that the SPX does not, in your opinion, exhibit Brownian motion).

Now my research was conducted recently; and as I said, I measured the price action of the SPX against a normal distribution (skew). It doesn't get more straight-forward than that.

My research is not unique; it has been performed by many other researchers as proof that price action is random and closely resembles a normal bell curve.

Since you are not willing to share your research (which you claimed you performed), then I suspect you really haven't performed anything, and this is simply an opinion on your part.

Perhaps you can share your background in statistics (your training) which would lead anyone to believe your opinion is correct.

Nothing irrational.

Randomness is not based on Brownian motion. Get this first. The study of probability and random behaviour existed hundreds of years before Brownian motion was mathematically modelled. So stop spouting nonsense.

I never said Randomness is uniquely Brownian motion, in fact that's what you said and I said the exact opposite.

I also never discussed whether S&P exhibits some Brownian motion type behaviour (whatever that means), simply that it wasn't normally distributed based on my research and some academic research. And suggested that if your research differs you should publish it, as it would be interesting.

You have demonstrated to me that you don't understand randomness or distributions, and that puts any statistical research you may or may not have done, on shaky footing. You said that anything random is normally distributed. I already suspected you knew next to nothing when you were quoting probabilities associated with standard deviations from the mean, but that really does take the biscuit.

You can be assured that my qualifications on this matter are sufficient, and demonstrably more than yours. I don't want to get into a pissing contest, but you would lose. Lets just leave it at the level of stats and what is correct.

I like how you're now saying 'closely' resembles a Bell curve. Now you're on the right path of honesty, well done. But since it is NOT a bell curve, since it is NOT normally distributed, the probabilities that you quoted are not accurate (which was what my first post here was about).

 

[rb107]

Nothing irrational.

Randomness is not based on Brownian motion. Get this first. The study of probability existed hundreds of years before Brownian motion was mathematically modelled. So stop spouting nonsense.

I never said Randomness is uniquely Brownian motion, in fact that's what you said and I said the exact opposite.

I also never discussed whether S&P exhibits some Brownian motion type behaviour (whatever that means), simply that it wasn't normally distributed based on my research and some academic research. And suggested that if your research differs you should publish it, as it would be interesting.

You have demonstrated to me that you don't understand randomness or distributions, and that puts any statistical research you may or may not have done, on shaky footing. You said that anything random is normally distributed. I already suspected you knew next to nothing when you were quoting probabilities associated with standard deviations from the mean, but that really does take the biscuit.

You can be assured that my qualifications on this matter are sufficient, and demonstrably more than yours. I don't want to get into a pissing contest, but you would lose. Lets just leave it at the level of stats and what is correct.

I like how you're now saying 'closely' resembles a Bell curve. Now you're on the right path of honesty, well done. But since it is NOT a bell curve, since it is NOT normally distributed, the probabilities that you quoted are not accurate (which was what my first post here was about).

Ok. Let's focus on what was said. I said the SPX price action is predominantly random (not totally), and that there are outliers due to unexpected events. These outliers occur infrequently (you claim frequently, based on your research which you cannot provide).

I also said that the SPX price action closely resembles a normal distribution. Based on the very low level of skewness, it is very close. You don't have to have a perfect normally distributed curve to apply the Probability Model; a reasonably close approximation does quite well (after all, when dealing with statistics, we are dealing with probabilities; not certainties).

Now I can understand your disinterest in reviewing the research since you are convinced you know more than I. And I also can understand your reluctance to share neither your research or background as some indication that you know anything at all.

If you can't provide any support for your opinion, then this just devolves into, as you call it, a pissing match.

Btw, my background is a BA in Economics, and an MBA in Accounting and Finance. Now, does this mean I am an expert in statistics? No. But it does show a strong understanding of statistics, probabilities, and the ability to do research.

Again, what do you bring to the table that should lead me (or anyone else) to listen any further to your opinion? You do not provide any research to backup your claims, nor any background to show you have any understanding of what you're talking about. The fact that you admit to no knowledge of Brownian motion does lead me to believe your understanding of statistics and probability are limited.

If you can't provide any support for your opinion, then there really is no point to further this discussion.

 

[Shakone]

Ok. Let's focus on what was said. I said the SPX price action is predominantly random (not totally), and that there are outliers due to unexpected events. These outliers occur infrequently (you claim frequently, based on your research which you cannot provide).

I also said that the SPX price action closely resembles a normal distribution. Based on the very low level of skewness, it is very close. You don't have to have a perfect normally distributed curve to apply the Probability Model; a reasonably close approximation does quite well (after all, when dealing with statistics, we are dealing with probabilities; not certainties).

Now I can understand your disinterest in reviewing the research since you are convinced you know more than I. And I also can understand your reluctance to share neither your research or background as some indication that you know anything at all.

If you can't provide any support for your opinion, then this just devolves into, as you call it, a pissing match.

Btw, my background is a BA in Economics, and an MBA in Accounting and Finance. Now, does this mean I am an expert in statistics? No. But it does show a strong understanding of statistics, probabilities, and the ability to do research.

Again, what do you bring to the table that should lead me (or anyone else) to listen any further to your opinion? You do not provide any research to backup your claims, nor any background to show you have any understanding of what you're talking about. The fact that you admit to no knowledge of Brownian motion does lead me to believe your understanding of statistics and probability are limited.

If you can't provide any support for your opinion, then there really is no point to further this discussion.

You are putting statements as if from me that I haven't written, for the second post running. You made some statements about the probability of events. I said I didn't believe those probabilities (because I knew how you obtained them).

You have applied the concept of standard deviations from the mean of a normal distribution and the probability associated with that. Fine. But it is not normally distributed. You now accept that the distribution is not normal. Good. You say it's 'close'. How is that measured? Lots of things are close. What is a measure of closeness? So now take the probability that you had for the normal distribution, and tell me what the mathematical probability is for something that's distributed 'close' to the normal, using your theory of 'closeness'. If your answer is 'well it was close to normal, so lets just use that probability' then you must accept the probability is not accurate, which is what I initially posted. Be nice if you could be very specific and give me a precise probability.

I have already provided support of my opinion. It's called being correct, and having an understanding of probability and statistics.

A BA in economics and an MBA in Accounting and Finance does not show a strong understanding in statistics or probability, but congratulations on your degrees. You've already demonstrated you have a poor understanding, since you think the world of probability and randomness revolves around Brownian motion.

In terms of academic research, you could google empirical data of stocks, S&P 500, open up a copy of Hull and look at the references etc. I've just googled and found several papers that indicate normal is not so good as a distribution for the S&P 500. Can you google?

 

[rb107]

You are putting statements as if from me that I haven't written, for the second post running. You made some statements about the probability of events. I said I didn't believe those probabilities (because I knew how you obtained them).

You have applied the concept of standard deviations from the mean of a normal distribution and the probability associated with that. Fine. But it is not normally distributed. You now accept that the distribution is not normal. Good. You say it's 'close'. How is that measured? Lots of things are close. What is a measure of closeness? So now take the probability that you had for the normal distribution, and tell me what the mathematical probability is for something that's distributed 'close' to the normal, using your theory of 'closeness'. If your answer is 'well it was close to normal, so lets just use that probability' then you must accept the probability is not accurate, which is what I initially posted. Be nice if you could be very specific and give me a precise probability.

I have already provided support of my opinion. It's called being correct, and having an understanding of probability and statistics.

A BA in economics and an MBA in Accounting and Finance does not show a strong understanding in statistics or probability, but congratulations on your degrees. You've already demonstrated you have a poor understanding, since you think the world of probability and randomness revolves around Brownian motion.

In terms of academic research, you could gogle empirical data of stocks, S&P 500 etc. I've just googled and found several papers that indicate normal is not so good as a distribution for the S&P 500. Can you google?

Ok. I understand you need to have things spelled out several times.

How do we measure how close the SPX price action is to a normal distribution? It's called SKEW. Skew measures this difference (I suggest you Google the definition and its application). Therefore, we have a quantitative measure; and this measure shows that the average SKEW over a 5-year period is -0.053 (anything under absolute 1.0 is considered random).

Now, since you really don't have any understanding of statistics (this is fairly clear based on your responses), you don't understand the concept of inexactness, which is what probabilities are about. There is no certainty when dealing with probabilities; we could express the accuracy of a distribution over many occurrences using the standard error (which drops as the number of occurrences increases), but rather than lecture you, I suggest you do the research and start learning about randomness and its relationship to Brownian motion, probabilities, and Google independent white-papers on the randomness of the S&P.

Btw, stating that you are "correct" and that you understand statistics is not a proof of anything. It is simply your opinion; an opinion that is based on, imo, wishful thinking and nothing else.

On the other hand, if you do have a background in statistics, then I suggest you run a simple test on the distribution of the daily price action of the SPX and compare it to a normal distribution over the last 5 years (or more if so inclined). Then come back with the results and we can talk some more.

If you chose to look at the test results of others in charting the SPX distribution on a daily basis (not annual), then you will see that it resembles closely a bell curve (with high tops at the center, and long tails). Does the high tops and long tails exclude the use of the probability model? Imo, and from my experience, NO. One other point: the longer the time-frame (daily vs weekly vs monthly vs annual) the less random is the distribution which diminishes the applicability of the probability model.

 

[Shakone]

Ok. I understand you need to have things spelled out several times.

How do we measure how close the SPX price action is to a normal distribution? It's called SKEW. Skew measures this difference (I suggest you Google the definition and its application). Therefore, we have a quantitative measure; and this measure shows that the average SKEW over a 5-year period is -0.053 (anything under absolute 1.0 is considered random).

Now, since you really don't have any understanding of statistics (this is fairly clear based on your responses), you don't understand the concept of inexactness, which is what probabilities are about. There is no certainty when dealing with probabilities; we could express the accuracy of a distribution over many occurrences using the standard error (which drops as the number of occurrences increases), but rather than lecture you, I suggest you do the research and start learning about randomness and its relationship to Brownian motion, probabilities, and Google independent white-papers on the randomness of the S&P.

Btw, stating that you are "correct" and that you understand statistics is not a proof of anything. It is simply your opinion; an opinion that is based on, imo, wishful thinking and nothing else.

On the other hand, if you do have a background in statistics, then I suggest you run a simple test on the distribution of the daily price action of the SPX and compare it to a normal distribution over the last 5 years (or more if so inclined). Then come back with the results and we can talk some more.

If you chose to look at the test results of others in charting the SPX distribution on a daily basis (not annual), then you will see that it resembles closely a bell curve (with high tops at the center, and long tails). Does the high tops and long tails exclude the use of the probability model? Imo, and from my experience, NO. One other point: the longer the time-frame (daily vs weekly vs monthly vs annual) the less random is the distribution which diminishes the applicability of the probability model.

You're just trying to wind me up now.


I said in the beginning I don't believe the probability you quoted is accurate. You argued.

You stated a probability that was based on it being normally distributed. You now accept that it is not normally distributed, since you're mentioning skewness. That means it differs from normal distribution. It has skew. It also has fatter tails. As a consequence this means that the probability does not need to be 84%. Concentrate on this. You're using an assumption which doesn't hold.

You then admitted that it might be close to normally distributed. I said good, you're getting there in terms of 'close' and asked what the accurate probability is. You haven't answered what this probability is. You seem to have just given up and decided that the incorrect assumption (normal dist.) will do for you as an approximate. Not the worst estimate in the world, but not accurate and not something you'd want to risk money on.

If as you say a probability can never be certain, then where do you get the 84% probability from anyway? Why say something as if it's true when you're not certain? Why argue with someone who doesn't believe it, when you're not even certain yourself? I think what you meant to say is that the outcome cannot be certain (because it is random), or perhaps you meant that due to limited data the probability is not precise. Either way, the probability of being within standard deviations of the mean for a normal dist. is certain, but again, we haven't got a normal dist.

 

[jerry2dt]

Fascinating to watch a couple of academics chat about probabilities, skew, etc. I'm just a daily trader who relies on the odds and so am not that surprised when my trade which has a 95% statistical chance of success goes against me 5% of the time, more or less. But it's only when it gaps that I seem to get hurt. When the trade goes slowly against, I can easily make adjustments and be just fine. I am talking about short strangles primarily with 90% inside window.

Can either or both of you give me an idea of using stats more beneficially? I've been doing this for the past 4 years and have had a few losing months but am up nicely over the time span. Is disaster just waiting for me?

Thanks,

Jerry

 

[Shakone]

Fascinating to watch a couple of academics chat about probabilities, skew, etc. I'm just a daily trader who relies on the odds and so am not that surprised when my trade which has a 95% statistical chance of success goes against me 5% of the time, more or less. But it's only when it gaps that I seem to get hurt. When the trade goes slowly against, I can easily make adjustments and be just fine. I am talking about short strangles primarily with 90% inside window.

Can either or both of you give me an idea of using stats more beneficially? I've been doing this for the past 4 years and have had a few losing months but am up nicely over the time span. Is disaster just waiting for me?

Thanks,

Jerry

I don't know what you're doing, so can't really comment on whether you'll have continued success or not.

You don't need sophisticated mathematics to trade, although some do use it. This talk isn't very sophisticated though. Skewness indicates a more likely direction and fat tails indicate extreme moves are more likely than normal would expect. If you can successfuly identify either then there may be some potential for profit, just as if you can identify a highly volatile instrument or a mean-reverting instrument or a trending instrument, then you may be able to capitalise on that property. The market changes over time though, so it's questionable about how useful all this is. What I do know is that it's hard to put accurate and rigorous probabilities on many things without making some assumptions that may not apply. And unfortunately the saying 'a little knowledge is a dangerous thing' applies quite well here.