Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Fooled by randomness 15 replies

Determining randomness in a trading system 35 replies

Best book on randomness in markets and market probabilities? 4 replies

Thoughts on randomness and systems 40 replies

Poll: Are trends rooted in randomness? 16 replies

MM - Using randomness against Asymmetric Leverage

  • Post #1
  • Quote
  • First Post: Aug 21, 2009 7:19am Aug 21, 2009 7:19am
Does Randomness eliminate Asymmetric Leveraging?
What factor should determine my X Variable (see below) and does it have to be random too?

I wondered if Asymmetric Leveraging could be somehow eliminated. The only logical way of "fighting" against it, is by having a fixed position size/percentage/whatever!

But say I use a perfect die so each time I roll a number, and THAT number IS the % of risk I will use for the trade I'm entering. After X (X = Variable such as time or profit) I use a higher numbered die (think D&D dice) starting with a 6 sided die.

Unquestionable FACTS for the sake of my questions:
1) Perfect die = pure randomness.
2) My trading gives me more then 55% expectancy.
Quoting ozwilson
Does Randomness eliminate Asymmetric Leveraging?
What factor should determine my X Variable (see below) and does it have to be random too?

I wondered if Asymmetric Leveraging could be somehow eliminated. The only logical way of "fighting" against it, is by having a fixed position size/percentage/whatever!

[color=black][font=Verdana]But say I use a perfect die so each...
X = 2->5

If you choose to use 5, I assume that if you have a 55% expectancy there is no real chance you'll have 20 losing trades in a row. I've simulated a million random 50% win expectancy systems on Matlab, and I've never had more than 13 losses in a row.
Hi ozwilson:

There is no need to, IMHO.

RISK = STOP LOSS * POSITION SIZE.

If you determine your risk and stop loss, then you calculate your position size.

Of course, if you randomize your position size, you will have to adjust your stop loss.
If you only trade long, then you can't always be wrong.
FrankenPip - What I want is to achieve the goal of eliminating Asymmetric Leveraging... Maybe it's not a big deal (or noble cause) but I wanted to see how you people think. So do you think it's could be done?

x1810 - I don't understand what your saying.. If you can explain what you mean It will be appreciated!

So I'll ask again - Is it theoretically possible to eliminate AL by the "fog" of randomness?
Quoting ozwilson
FrankenPip - What I want is to achieve the goal of eliminating Asymmetric Leveraging... Maybe it's not a big deal (or noble cause) but I wanted to see how you people think. So do you think it's could be done?

x1810 - I don't understand what your saying.. If you can explain what you mean It will be appreciated!

So I'll ask again - Is it theoretically possible to eliminate AL by the "fog" of randomness?
As I understand it, Asymmetric leveraging is a techniquie used to adjust your position sizes according to your profits or losses (mainly profits). If this is the case, I don't understand what you are trying to eliminate. If you mean something else by Asym. Lev., please explain and I will try to answer.
Quoting FX-Petra
As I understand it, Asymmetric leveraging is a technique used to adjust your position sizes according to your profits or losses (mainly profits). If this is the case, I don't understand what you are trying to eliminate.
Ok I guess my post was confusing... I'm NOT taking about eliminating Asymmetric leveraging as a techniquie. I meant to eliminate the detrimental effect of Asymmetric leveraging as a phenomenal -
Example:
If I use 1 mini lot for every 1000$ I have in my account, and my account grows to 2000$ (so by my MM rules) I need to now use 2 mini lots for my first time and take a loss. My account is now below 2000$. I need go back to 1 mini lot and win 2 times (theoretically) to compensate for my 2 mini lots trade loss.

So by my logic, if my money management is random (as is the time I take a lose or win, even with a positive expectancy - I still can't predict when I'm wrong or have a losing streak) then AL doesn't affect me which means I eliminated AL by randomness.

What do you think and thanks for helping FX-Petra,
Oz
Quoting ozwilson
Ok I guess my post was confusing... I'm NOT taking about eliminating Asymmetric leveraging as a techniquie. I meant to eliminate the detrimental effect of Asymmetric leveraging as a phenomenal -
Example:
If I use 1 mini lot for every 1000$ I have in my account, and my account grows to 2000$ (so by my MM rules) I need to now use 2 mini lots for my first time and take a loss. My account is now below 2000$. I need go back to 1 mini lot and win 2 times (theoretically) to compensate for my 2 mini lots trade loss.

So by my logic, if my money management...
Ok, we are getting closer. I understand the first paragraph...still unsure about the second. If your money management is "random" as you say, then there is no symmetry to it...neither is it asymmetrical. It is random and has no order to it. I'm not sure this answers your question. If not, try to explain once again. I am dense sometimes.

Regards
Quoting FX-Petra
If your money management is "random" as you say, then there is no symmetry to it...neither is it asymmetrical.
Keep in mind that for the example I gave in my last post my money management was NOT random (every 1000$ I use 1 more mini lot)! That is exactly my question - If my money management was random, would it really eliminate AL because is has no order to it.... Do you understand what I mean? If so what is the negative side of using random MM?

As I understood it the only way to minimize the effects of AL is by using micro accounts because of finer granularity. Nowhere was there mention about how to eliminate it. I found an old post that might help (picture in post #3) -
http://www.forexfactory.com/showthre...?t=6948&page=9
Quoting ozwilson
Does Randomness eliminate Asymmetric Leveraging?
.............
IMHO AL is a small price to pay for the benefit of compounding, assuming you have a +EV method. One could equally argue that fixed frac has the benefit of reducing pos size, and hence losses (in absolute terms), when drawdown is the result of market conditions being (temporarily) unfavorable to your method. I see it as a case of viewpoint, whether your thinking 'scale' is 'arithmetic' or 'geometric' (exponential), and/or whether your primary objective is rate of recovery, or further reduction of (absolute $$) risk.

Using variable pos sizes (and random sizes means without the prerequisite of adjusting size relative to each setup's expectancy) simply means that if you get lucky, your higher sizes will coincide with your winning trades; unlucky, vice versa. Hence I see it as being neither better nor worse, just an unnecessary complication that brings another level of 'luck' into the equation.

I might well be missing something, but (1) I don't see AL as a great problem in the overall scheme of things, and (2) I don't see random sizing as a cure.

If a method experiences prolonged drawdown, then I'd be more concerned with its EV, than AL (but of course this doesn't apply to you, given your EV is 55%).

I discuss the pros and cons of fixed frac versus flat betting in posts #23 and #25 here. There are also downloadable XLSs in later posts that allow experimentation with different MM systems, and their effect on expectancy.

You might also be interested in Tdion's approach to MM, in the same thread, and also here. If I understand correctly, he increases pos size (both return and risk) to accelerate the recovery to breakeven and therefore alleviate the effect of AL.
Up to an account size of circa $200K you're better off using micro lots. There is a greater degree of granularity with the smaller lot size.

With a mini account the loss may require you go back to 1 mini but with micro lots you could be trading 157 micros as an example.

I think I have that right, if not Hanover will put me right, no doubt .
Quoting PeterFM
Up to an account size of circa $200K you're better off using micro lots. There is a greater degree of granularity with the smaller lot size.

With a mini account the loss may require you go back to 1 mini but with micro lots you could be trading 157 micros as an example.

I think I have that right, if not Hanover will put me right, no doubt .
Peter, I believe you're correct when you equate 'asymmetric leverage' with the granularity considerations. That, as I understand it, is the traditional definition of AL, and was particularly relevant to futures traders whose minimum contract size was something like 30,000 units. Of course the availability of micro (and nano) lots has helped to alleviate this, in spot forex.

However, it now seems fashionable to describe the problem outlined here as 'asymmetric leverage'. I assumed - rightly or wrongly - that this was what the OP was referring to. But maybe I've missed the point. Either way, any lack of granularity (caused by the decreasing account size) would make the recovery-to-breakeven process even more difficult.
Quoting hanover
I might well be missing something, but (1) I don't see AL as a great problem in the overall scheme of things, and (2) I don't see random sizing as a cure.
1) I agree with you, but this is a theoretical question more then solving a trading problem!

2) IMO it is as much of a cure for AL as a fixed dollar amount "scale"!
There's one thing you didn't address in your post:
We can both agree that my primary objective in this thread is rate of recovery and compounding more money by using a (random) geometric progression "scale" (like Tidon's approach to MM), right?
But because I want to adjust the trade size relative to my account balance on a random basis it means that the effects of AL will be random as well! So if I have more random options (more and more sides in the die) I will be compunding money in a geometric progression "scale" but have the effects of AL like in an arithmetic progression "scale". right?
  • Post #13
  • Quote
  • Last Post: Sep 12, 2009 3:13pm Sep 12, 2009 3:13pm
Quoting ozwilson
But because I want to adjust the trade size relative to my account balance on a random basis it means that the effects of AL will be random as well! So if I have more random options (more and more sides in the die) I will be compunding money in a geometric progression "scale" but have the effects of AL like in an arithmetic progression "scale". right?
Well, I'm no expert, but it seems to me that the 'geometric' element stems from the fixed frac nature of the approach, i.e. the fact that a percentage of the account is being risked each time instead of a flat amount. As opposed to how the percent value is being varied. The percentage itself might be being varied arithmetically, but the effect on the account balance remains geometric/exponential.

Alternatively, you could perhaps describe your approach as some kind of random progression. But then we'd be merely talking terminology.

Or you could choose to look at it this way: With a six sided die, there'd be an even probability of your risking 1%, 2%, 3%, 4% 5% or 6% of your account on each trade. Over a large number of trials, that would be something like risking the average, i.e. 3.5%, per trade. So that would be a kind of 'geometric' fixed frac.

If the 6% trades turn out to be winners, and the 1% trades losses, then bravo! Vice versa, disaster. As every Blackjack pro knows, the key is to increase bet size when expectancy is at its highest. But unless a forex trader knows in advance which trades have a higher probability of profit than the others, there's no valid basis for varying his position size.

Re increasing the number of faces on the die: As with Tdion, increasing the percent risked will either accelerate recovery to breakeven, or the downward spiral toward ruin, all depending on how the next sequence of trades happens to pan out.

Some years back I looked at alternatives to fixed frac, e.g. optimal-f, Ryan Jones' fixed ratio. I can't remember exactly how all these work now, but my conclusion was that every MM method ultimately involves trading off some aspect of return for some aspect of risk. If one promised a significantly better outcome (in the long term, statistical sense) than the others, then no doubt this would be widely known, and everybody would use it.

Anyway, this is all fairly abstract, and I'm not very good at thinking outside the traditional squares. Perhaps you're seeing something that I'm not.
0 traders viewing now