Kelly sizing...

[u21c3f6]

Quote from Gabfly1:

Frankly, I know nothing about Kelly's size and I have no interest in finding out.

As for you guys with all of your numeric specificity, you might do well to consider Sebastian Mallaby's words on page 231 of his book, More Money Than God: "The real lesson of LTCM's failure was not that its approach to risk was too simple. It was that all attempts to be precise about risk are unavoidably brittle."

I disagree, here's why:

The fact of the matter is that LTCM's approach to risk was too simple because it did not account for all possibilities. Kelly can be very dangerous if you are not aware of or take into consideration all your risks.

Despite that, Kelly can still be used but you have to make sure that your set-up warrants using Kelly. Several posters have mentioned that you can't use Kelly because you don't always know your risk. I think they write that because they are thinking about the way that they trade which usually involves "stop losses" which may not be limited or other methods of controlling risk that may not be set in stone. Those may be good reasons to not use Kelly or at least not use full-Kelly.

But think of an option spread set-up where you can define your risk to the maximum of the spread, this is an example where Kelly can be used as long as your set-up has an edge or in other words, your set-up is profitable after a valid sample of similar trades in bull, static and bear markets.

Joe.

 

[Gabfly1]

Quote from Daal:

Optimal F will get you broke in the long-run

Stated differently, one will get Optimally F'ed.

 

[Gabfly1]

Quote from u21c3f6:

I disagree, here's why:

The fact of the matter is that LTCM's approach to risk was too simple because it did not account for all possibilities. Kelly can be very dangerous if you are not aware of or take into consideration all your risks.

Despite that, Kelly can still be used but you have to make sure that your set-up warrants using Kelly. Several posters have mentioned that you can't use Kelly because you don't always know your risk. I think they write that because they are thinking about the way that they trade which usually involves "stop losses" which may not be limited or other methods of controlling risk that may not be set in stone. Those may be good reasons to not use Kelly or at least not use full-Kelly.

But think of an option spread set-up where you can define your risk to the maximum of the spread, this is an example where Kelly can be used as long as your set-up has an edge or in other words, your set-up is profitable after a valid sample of similar trades in bull, static and bear markets.

Joe.

LTCM's people were not stupid. They just weren't as smart as they thought they were. Unfortunately for them, they traded as though they were as smart as they thought they were.

I confess that I don't know either the Kelly or Vince formulas, nor do I care to find out. Each one invariably assumes an immediately stable probability distribution when, in fact, no legitimate probability distribution actually exists for the financial markets in real time. An ever changing frequency distribution of past behavior is a poor substitute if your intention is to get numerically specific. For my own limited purposes, I only rely on "the balance of probability," which invalidates overly precise fine-tuning and requires a sizable margin for error. That is the difference between true probability and uncertainty.

 

[u21c3f6]

Quote from Gabfly1:

LTCM's people were not stupid. They just weren't as smart as they thought they were. Unfortunately for them, they traded as though they were as smart as they thought they were.

I confess that I don't know either the Kelly or Vince formulas, nor do I care to. Each one invariably assumes an immediately stable probability distribution when, in fact, no legitimate probability distribution actually exists for the financial markets in real time. An ever changing frequency distribution of past behavior is a poor substitute if your intention is to get numerically specific. For my own limited purposes, I only rely on "the balance of probability," which invalidates overly precise fine-tuning and requires a sizable margin for error. That is the difference between true probability and uncertainty.

First, please note that I did not mean to intimate that LTCM's people were stupid; they were actually geniuses that got caught in a set of circumstances that they did not anticipate. Which goes to my point that Kelly is better used where your risk input is actually the floor and not a calculation where you ignore black swans and/or other problems that may make your position riskier than it appears on the surface.

Second, you can set-up trades that have exact maximum risks which goes to one of my other points. Posters tend to respond based on how they do things but of course there are many different ways to trade. Just because your distribution is not one where you feel that you can use Kelly does not mean that there aren't others that actually do have set-ups that lend themselves to Kelly calculations. I personally use half-Kelly very successfully. By using half-Kelly, I don't have to worry about being numerically specific, I just make sure that my trading has an edge and a defined risk.

Joe.

 

[kut2k2]

Quote from intradaybill:

Just because people do not answer your stupidity that does not mean you can say Q.E.D.

All of your formulas are wrong.

You gave me a challenge, I answered it. You weren't man enough to acknowledge that.

Now I give you a challenge. Using example 2 in my thread, prove my Kelly results are wrong, with real numbers, not a load of verbiage.

"If you dissappear [sic] it would mean that you are a hoax and a coward."

 

[kut2k2]

Quote from u21c3f6:

I disagree, here's why:

The fact of the matter is that LTCM's approach to risk was too simple because it did not account for all possibilities. Kelly can be very dangerous if you are not aware of or take into consideration all your risks.

Despite that, Kelly can still be used but you have to make sure that your set-up warrants using Kelly. Several posters have mentioned that you can't use Kelly because you don't always know your risk. I think they write that because they are thinking about the way that they trade which usually involves "stop losses" which may not be limited or other methods of controlling risk that may not be set in stone. Those may be good reasons to not use Kelly or at least not use full-Kelly.

But think of an option spread set-up where you can define your risk to the maximum of the spread, this is an example where Kelly can be used as long as your set-up has an edge or in other words, your set-up is profitable after a valid sample of similar trades in bull, static and bear markets.

Joe.

LTCM failed because they ignored Kelly. They were overleveraged to the hilt.

 

[intradaybill]

Quote from kut2k2:

You gave me a challenge, I answered it. You weren't man enough to acknowledge that.

Now I give you a challenge. Using example 2 in my thread, prove my Kelly results are wrong, with real numbers, not a lot of verbiage.

"If you dissappear [sic] it would mean that you are a hoax and a coward."

I gave you a textbook problem. You are giving me a problem based on a wrong formula you have derived, which has nothing to do with Kelly, yet you are calling it "Kelly result".

The reason you underestimate Kelly is because you are limiting the number of trials to a number equal to the number of outcomes. That should not be the case. You are essentially maximizing for only 3 trials because this is how your set up the Gain function.

I explained already to you that you are maximizing the wrong Gain function. You are limiting the gain function by the number of outcomes. This is false. The gain function is for arbitrary n, the number of trials. You need to maximize that.

 

[kut2k2]

Quote from intradaybill:

I gave you a textbook problem. You are giving me a problem based on a wrong formula you have derived, which has nothing to do with Kelly, yet you are calling it "Kelly result".

The reason you underestimate Kelly is because you are limiting the number of trials to a number equal to the number of outcomes. That should not be the case. You are essentially maximizing for only 3 trials because this is how your set up the Gain function.

I explained already to you that you are maximizing the wrong Gain function. You are limiting the gain function by the number of outcomes. This is false. The gain function is for arbitrary n, the number of trials. You need to maximize that.

The problem isn't based on any formula, my results are based on formulas. If the formulas are wrong, then the results are wrong and a person who really understood the mathematics could prove that. With numbers, not verbiage.

You've been exposed as a fraud. As well as an idiot.

 

[logic_man]

Quote from u21c3f6:

I disagree, here's why:

The fact of the matter is that LTCM's approach to risk was too simple because it did not account for all possibilities. Kelly can be very dangerous if you are not aware of or take into consideration all your risks.

Despite that, Kelly can still be used but you have to make sure that your set-up warrants using Kelly. Several posters have mentioned that you can't use Kelly because you don't always know your risk. I think they write that because they are thinking about the way that they trade which usually involves "stop losses" which may not be limited or other methods of controlling risk that may not be set in stone. Those may be good reasons to not use Kelly or at least not use full-Kelly.

But think of an option spread set-up where you can define your risk to the maximum of the spread, this is an example where Kelly can be used as long as your set-up has an edge or in other words, your set-up is profitable after a valid sample of similar trades in bull, static and bear markets.

Joe.

Agreed. All of this discussion about black swans invalidating Kelly ignores the fact that in a worst-case scenario some trading vehicles simply go to zero if hit by a black swan, rather than continue to sustain losses beyond the Kelly limit.

 

[intradaybill]

Quote from kut2k2:

The problem isn't based on any formula, my results are based on formulas. If the formulas are wrong, then the results are wrong and a person who really understood the mathematics could prove that. With numbers, not verbiage.

You've been exposed as a fraud. As well as an idiot.

The idiot is he, who first defines the Gain in terms of percent returns over an arbitrary number of trials (trades), optimizes the function for the fraction k but then in the results uses in the place of percent return the payout over a number of trials equal to the number of outcomes.

If you do not understand what I wrote above, which is how you formulated the problem exactly, you not only an illiterate idiot but your persistent failure and resistance to comprehend your errors point to this

http://en.wikipedia.org/wiki/Asperger_syndrome

You underestimate Kelly because you compute the fraction over a few initial trials where Kelly computes it for the average of trials. The error is some percentage of the Kelly fraction because geometric growth mostly builds up in the first steps.

Try to understand what I wrote it will help you in many respects:

- help you to understand Kelly

- help you to understand geometric growth

- help you to understand the difference between trials and outcomes - random variables and their valus

- help you to understand that math variables cannot be changed to anything you want after you define them

- help you to become a better person by realizing your autism