concurrent uncorrelated positions | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Reasonably uncorrelated basket trading 6 replies

Trend basket by linear regression of uncorrelated pairs 12 replies

Should I be concerned about uncorrelated markets? 7 replies

uncorrelated pairs for risk control and focus 3 replies

Uncorrelated pairs 1 reply

Attachments: concurrent uncorrelated positions
Exit Attachments

concurrent uncorrelated positions

  • #41
  • May 6, 2014 2:56pm May 6, 2014 2:56pm
Quoting PipMeUp
{quote} An alternative is to risk a percentage of the square root of your money. The reason is that the square root increases faster at the beginning. It keeps increasing forever but less and less. The less money you have the more the percentage will be. The more you accumulate the more you risk in absolute value but the less you risk in percentage. During a DD you give back less in % since your risk was less in %. This way you preserve the accumulated profits. Corollary: It is easier for rich people to keep their money.
Very clever. Thanks for posting.

Given that SQRT(X) = X^0.5, we could use values other than 0.5 to fine tune this further.
  • #42
  • May 6, 2014 3:14pm May 6, 2014 3:14pm
Any function which is strictly positive in ℝ+, monotonically increasing and which has a asymptotically derivative decreasing to zero will do.
No greed. No fear. Just maths.
  • #43
  • May 6, 2014 3:30pm May 6, 2014 3:30pm
Quoting PipMeUp
{quote} The correlation tells you that if you trade your system on E/U and U/CHF at the same time, you will almost always get the same signal in each at the same time (actually one LONG and one SHORT). The outcome of the trades will almost always be the same: both winners or both losers. You could simply bet twice the risk on E/U and not trade U/C at all. But if you double the risk you are over leveraged according to your MM. Now if you trade N/U and E/G that's a different story. The signals aren't correlated. You're trading your system twice more...
That's the reason why I only use correlation to confirm what's happening and not just to trade on it. In my view the only thing it tells you is that USD is behaving as expected whether it is weak or strong, but nothing else really. The other issue with correlation is that the risk on both trades is not nearly as linear as we like to believe, in a perfect world yes. But after it's all said and done, one of those pairs would still be better to trade than the other, it might be the difference of gaining an extra 5 pips or 50 depending on your time frame, or losing the same amount. So my question is, wouldn't it be better to trade the better pair and forget about trading both in a correlated less efficient manner?
Being a trader is lonely, but being a great trader is lonelier still
  • #44
  • May 6, 2014 3:45pm May 6, 2014 3:45pm
Quoting saneblane
{quote} That's the reason why I only use correlation to confirm what's happening and not just to trade on it. In my view the only thing it tells you is that USD is behaving as expected whether it is weak or strong, but nothing else really. The other issue with correlation is that the risk on both trades is not nearly as linear as we like to believe, in a perfect world yes. But after it's all said and done, one of those pairs would still be better to trade than the other, it might be the difference of gaining an extra 5 pips or 50 depending on your...
I have thought about that and this is my line of thinking.

In general, it is usually best just to trade the better pair. But we operate in the specific. aaa/bbb may generally be better than aaa/ccc, but today, ccc may be dropping like a rock while bbb just chugs along. Int that case today, aaa/ccc is better. How do you know this before it has already happened? Today may be the day they do not correlate at all.

Correlation calculations are on the past. Their ability to tell the future is limited. Correlations change over time. Their changes may be trending, Hm? or not. Have we just hit a correlation reversal? No I am not just having fun--these things happen.

I agree totally with your first sentence:"That's the reason why I only use correlation to confirm what's happening and not just trade on it."

  • #45
  • Edited 4:22pm May 6, 2014 3:58pm | Edited 4:22pm
Quoting Rap Skallion
{quote} I have thought about that and this is my line of thinking. In general, it is usually best just to trade the better pair. But we operate in the specific. aaa/bbb may generally be better than aaa/ccc, but today, ccc may be dropping like a rock while bbb just chugs along. Int that case today, aaa/ccc is better. How do you know this before it has already happened? Today may be the day they do not correlate at all. Correlation calculations are on the past. Their ability to tell the future is limited. Correlations change over time. Their changes...

Hahah. I guess we do think alike. All those things you said there is the reason why i have said that the chart in Forex is the biggest magic trick in the world. To answer your question about how do we know these things you ask about. You can't know all the time but the times that you are right can be increased more often than not if you know for certain that you are wrong or at least to a high degree. I wish I was as articulate as Hanover but some of my thoughts I can't even put them down. But this has been a good thread.

All the information that is needed to have an 80 percent probability trade is presented to us buried in "code" across the entire market, it's a lot of research to get it, but I will say this much, it takes a hell lot of research to get that info, and I understand why a lot of traders just like to trade in blind ignorance and hope for the best.
Being a trader is lonely, but being a great trader is lonelier still
  • #46
  • May 6, 2014 5:58pm May 6, 2014 5:58pm
Quoting saneblane
{quote} Hahah. I guess we do think alike. All those things you said there is the reason why i have said that the chart in Forex is the biggest magic trick in the world. To answer your question about how do we know these things you ask about. You can't know all the time but the times that you are right can be increased more often than not if you know for certain that you are wrong or at least to a high degree. I wish I was as articulate as Hanover but some of my thoughts I can't even put them down. But this has been a good thread. All the information...
There is a whole class of people I have observed that spend a minute or two of thought and then say " This is too hard for me, I give up. Somebody tell me the answer." They pick a guru and use his rules (which could even be very good rules) without understanding the reasoning behind them.
I have counted some people like that as very good friends, very nice people -- but they should not get themselves into trading.

Each new day the forex is a new problem, with new wrinkles, and incomplete information, and calculations can help but in the end, if you lose less than you make, you can say you have done well.

There is no completely clear explanation of anything of importance, involving all the ramifications, anywhere, unless you think tautologies are important. but there are partial insights observations which help us to be right a little more than we are wrong, and what to do about it.

For me, I think you belong in this business.
  • #47
  • May 6, 2014 6:32pm May 6, 2014 6:32pm
Quoting PipMeUp
Any function which is strictly positive in ℝ+, monotonically increasing and which has a asymptotically derivative decreasing to zero will do.
Aw c'mon. You know how many people read this who have had no calculus.

He means any curve that can be described by a formula with unique solutions, doesn't he?

Have I made myself clear?

  • #48
  • May 6, 2014 7:34pm May 6, 2014 7:34pm
Quoting PipMeUp
{quote} An alternative is to risk a percentage of the square root of your money. The reason is that the square root increases faster at the beginning. It keeps increasing forever but less and less. The less money you have the more the percentage will be. The more you accumulate the more you risk in absolute value but the less you risk in percentage. During a DD you give back less in % since your risk was less in %. This way you preserve the accumulated profits. Corollary: It is easier for rich people to keep their money.
The things I've learned from you, PipMeUp. Thank you for sharing
"To hold, you must first open your hand. Let go." - Lao Tzu
  • #49
  • May 7, 2014 1:53am May 7, 2014 1:53am
Quoting Rap Skallion
Aw c'mon. You know how many people read this who have had no calculus.
You're right . I mean a function that keeps growing forever but grows slower and slower.
keeps growing forever => you keep risking more and more in absolute value
grows slower and slower => the more you have the less you risk in percentage

E.g. with the square root:
√$100 = $10 -> 10%

√$200 = $14 -> 7%
√$1,000 = $32 -> 3%
√$10,000 = $100 -> 1%
√$1,000,000 = $1000 -> 0.1%

You shift from growing quickly money you don't care losing to preservation of money you not longer accept losing.
No greed. No fear. Just maths.
  • #50
  • May 7, 2014 3:33am May 7, 2014 3:33am
Quoting PipMeUp
{quote} You're right . I mean a function that keeps growing forever but grows slower and slower. keeps growing forever => you keep risking more and more in absolute value grows slower and slower => the more you have the less you risk in percentage E.g. with the square root: √$100 = $10 -> 10% √$200 = $14 -> 7% √$1,000 = $32 -> 3% √$10,000 = $100 -> 1% √$1,000,000 = $1000 -> 0.1% You shift from growing quickly money you don't care losing to preservation of money you not longer accept losing.
PipMeUp

You just open one more new door ..... in my little brain .... thank's and keep going ...it's really helpful

Landsat
Logic will take you from A to B Imagination will take you anywhere (A.E)
  • #51
  • May 7, 2014 4:26am May 7, 2014 4:26am
Quoting PipMeUp
{quote} You're right . I mean a function that keeps growing forever but grows slower and slower. keeps growing forever => you keep risking more and more in absolute value grows slower and slower => the more you have the less you risk in percentage E.g. with the square root: √$100 = $10 -> 10% √$200 = $14 -> 7% √$1,000 = $32 -> 3% √$10,000 = $100 -> 1% √$1,000,000 = $1000 -> 0.1% You shift from growing quickly money you don't care losing to preservation of money you not longer accept losing.
I've read about it in money management book. Working good in theory, but it is hard to put it into practice
  • #52
  • May 7, 2014 5:48am May 7, 2014 5:48am
Quoting landsat
{quote} PipMeUp You just open one more new door ..... in my little brain .... thank's and keep going ...it's really helpful Landsat
That sounds so sensible.

thank you for making this clear to me.

I would use it as a guide. there is probably another sensible curve describing the money you withdraw from the account to put somewhere more stable. (The British government is a lot less likely to declare bankruptcy than even the solidest broker, etc.)

For me, I have promised myself the tithe, once I reach 6 figures, going to the charity of choice. (the Carter foundation comes to mind).
  • #53
  • May 7, 2014 6:09am May 7, 2014 6:09am
Quoting leebsurag
{quote} I've read about it in money management book. Working good in theory, but it is hard to put it into practice
no Pain no Gain ..... you want to trade but no homework ..... good luck ....
Logic will take you from A to B Imagination will take you anywhere (A.E)
  • #54
  • May 7, 2014 6:30am May 7, 2014 6:30am
Quoting PipMeUp
{quote} You're right . I mean a function that keeps growing forever but grows slower and slower. keeps growing forever => you keep risking more and more in absolute value grows slower and slower => the more you have the less you risk in percentage E.g. with the square root: √$100 = $10 -> 10% √$200 = $14 -> 7% √$1,000 = $32 -> 3% √$10,000 = $100 -> 1% √$1,000,000 = $1000 -> 0.1% You shift from growing quickly money you don't care losing to preservation of money you not longer accept losing.

Here's the thing. One can always say one is following some rule, just not exactly....

Noisy responses to a noisy market. Good or Bad? Depends...
  • #55
  • May 7, 2014 4:18pm May 7, 2014 4:18pm
Quoting PipMeUp
√$1,000,000 = $1000 -> 0.1%
A pleasant 'problem' for a trader to have.

Quoting PipMeUp
Any function which is strictly positive in ℝ+, monotonically increasing and which has a asymptotically derivative decreasing to zero will do.
It certainly beats Douglas Adams' answer of "42".
  • #56
  • May 7, 2014 4:43pm May 7, 2014 4:43pm
uncorrelated trades. Logic & from my point of view correct.
all trades will be alone in their destiny to the end. profit or loss.
- trades will not be correlated to each other to save one to another ( such as grid )
- uncorrelated instruments will be used to trade

unfo with Forex there is no uncorrelated pair. all is correlated with USD.

logic might be used for real uncorrelated instruments.
such as
EURUSD
APPL
My Best Trade is Next Trade!
  • #57
  • May 8, 2014 4:10pm May 8, 2014 4:10pm
Quoting PipMeUp
{quote} You're right . I mean a function that keeps growing forever but grows slower and slower. keeps growing forever => you keep risking more and more in absolute value grows slower and slower => the more you have the less you risk in percentage E.g. with the square root: √$100 = $10 -> 10% √$200 = $14 -> 7% √$1,000 = $32 -> 3% √$10,000 = $100 -> 1% √$1,000,000 = $1000 -> 0.1% You shift from growing quickly money you don't care losing to preservation of money you not longer accept losing.
Daniel Kahneman wrote in his book "Thinking Fast and Slow" about how our expected utility functions are concave in gains and convex in losses, meaning we are risk-loving in avoiding losses and risk-adverse when seeking gains. So I guess if we use a concave function (like the square root as you have shown) to modify our risk as our account grows, it'll fit the typical human psychology (of being risk-adverse in seeking gains) snugly.

thanks for sharing
Attached Image
Build good relationships with others.
  • #58
  • May 9, 2014 3:48am May 9, 2014 3:48am
For calculus impaired people
convex function = if you link any two distinct points on the graph of the function the segment of line is always above the curve.
cancave function = the opposite.

The fixed fraction MM, like 1%, is the function f(x)=0.01x. This is a straight line. The functions f(x)=a.x are the limit between cancave and convex functions. If you link any two points on the line, the segment is always exactly on the curve.


This shifts the thread aways from the pairs correlation to the MM but... after reading this post from shellsnail and a previous one from hanover, I thought about something.

All these MM have in common that the risk is determined only by the size of the account. It doesn't take care of the history of the size of the account (the wins and losses). On the opposite side, the martingale MM only takes care of the history with no regards to the account size.

Quoting hanover
Some trading coaches believe that you should size more aggressively when in DD, to accelerate possible recovery
This both takes care of the size and the history of the account. If you are in a DD it means that you were recently at an equity high and now you're back to an account size you used to have previously. At that time, your risk was some percentage of your account and you were unwilling to risk more. Yet if you knew that you would reach an equity high you would have risked more. If you increase the risk as long as the DD lasts you lose more and more. You're digging the hole faster and deeper. If you knew this before you would have risked less.

But what about risking slighly more when the account grows and less when the account shrinks? When you grow towards a new equity high you risk more and grow faster. As soon as you enter a DD period, you risk less than usual. Of course you won't recover faster but you will have less to recover! As soon as your DD period is over your equity increases again. So does your risk. Now you recover faster. At the right time.

=> You're LONG your system when it trends up. You're MA crossover trading your own equity curve! Who said MA crossover are useless ?
Attached Image (click to enlarge)
Click to Enlarge Name: equityMAs.png Size: 15 KB
No greed. No fear. Just maths.
  • #59
  • May 9, 2014 9:19am May 9, 2014 9:19am
Quoting PipMeUp
For calculus impaired people convex function = if you link any two distinct points on the graph of the function the segment of line is always above the curve. cancave function = the opposite. The fixed fraction MM, like 1%, is the function f(x)=0.01x. This is a straight line. The functions f(x)=a.x are the limit between cancave and convex functions. If you link any two points on the line, the segment is always exactly on the curve. This shifts the thread aways from the pairs correlation to the MM but... after reading this post from shellsnail...
Haha this will work very well if there's positive serial correlation... hmm but what if the serial correlation is negative wouldn't you shoot yourself in the foot? just wondering, or am I not making sense.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 2.9 MB
Build good relationships with others.
  • #60
  • May 9, 2014 10:07am May 9, 2014 10:07am
Systems tend to have a preferred market condition. For instance a system which loves the trends will hate the ranges. These periods tends to last in time. So I think that the serial correlation has good chance to be positive.

If the serial correlation is negative the equity will probably oscillate around a (preferably up going) mean. If the RR is small the MAs will tend to converge, reducing the effect of over- and under- betting (w.r.t your original MM). If the system has a good RR the sequences win-loss will keep the equity growing because even if the loss is bigger than the previous risk, it will be less than the previous profit. Same for loss-win, the winner will cover the loss despite the smaller lot size.

If there is no correlation at all the effects of over- and under- sizing will tend to cancel each other. If this wasn't the case you could beat a Roulette and I don't believe I found a way to do this just with MM .
No greed. No fear. Just maths.
0 traders viewing now