Money Management vs. % of Success in Trading Systems | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Top 100 best Trading Systems: post good Trading Systems here 62 replies

Alternative Trade Management/Money Management Methods 3 replies

Money Management / Risk Management 24 replies

Systems, Systems and Systems... 1 reply

What’s Differences Between Robot and Manual Trading Money Management Systems 0 replies

Attachments: Money Management vs. % of Success in Trading Systems
Exit Attachments

Money Management vs. % of Success in Trading Systems

  • #41
  • Sep 6, 2006 3:22pm Sep 6, 2006 3:22pm
Quoting Bemac
Hi Phil,

Now it has a + expectancy.

I don't have Mr. Jones' book. If there is a breife explanation that you could post I would appreciate that.
I hope this helps...it is from page 20/21 and took me a long time to type

The main characteristics of antimartingale methods are that is causes geometric growth during positive runs and suffers from what is called asymmetrical leverage during drawdowns. Asymetrical leverage simply states tht as an account suffers losses, the ability to make up those losses decreases. If a 20% drawdown is suffered, a 25% gain is required to get back to even. A 10% drawdown requires a 11.11% gain to get back to even. The forumula for this is:

[1/(1-%loss)] -1 = Required % gain

In many cases, asymmetical leverage does not affect trading. For example, ifa trader trading the absolute minimum available in the bond market suffered a 20% drawdown, the requjired gain would still be 25% of the new account balance, but the ability to acheive the extra 5% has not diminished. This occurs because even though the percentage required to recoup the precentage loss of hte account increases, the amount of capital to recoup the amount of captial lost remains the same. Therefore, asymmetrical leverage does not play a role in the performance of the account. [This concept applies to your spreadsheet]

On the other hand, it plays a huge role when traders apply certain money managment techniques. For example, if a trader decides to trade one contract for every $10,000 in the account, then a single contract would be traded from $10,000 through $19,999. At $20,000, contracts would increase from one to two. Suppose that the very first trade after increasing to two contracts is a loser for $1000. Since there were two contracts on this trade, the actual loss comes to $2000 and the account goes back to $18,000. According to the money management rules a single contract has to be traded once again. The trader must now incur two, $1000 winning trades to get the account back to where it was just prior to suffering a $1000 loss with two contracts. Here, the amount of capital required to bring the account back to even remains the same, but the ability to achieve that amount has decreased by 50%. That is asymmetrical leverage and it can be detrimental.


I hope this helps explain my point. There is nothing wrong with your spreadsheet but it does not account for variations in wins and losses, uneven distributions of wins and losses, the sequence of wins and losses nor the affects of asymmetrical leverage. Earlier in this thread I posted to equity curves of the same system. From them you can see that they both employed a 3% fixed ratio money management technique. Due to the reasons discussed above, one looked great, and one when bust.

Phil

PS Sorry about the spelling...I tried to type while reading and not looking at the keys.
  • #42
  • Sep 6, 2006 3:37pm Sep 6, 2006 3:37pm
Quoting LazyPawn
By the way, don't you find it unfair that having a 2% loss followed by a 2% win, or having a 2% win followed by a 2% loss both lead to a loss overall?! It means that if the system we are using has a 50% expectancy and we use a fixed ratio position size we are in fact losing money. Ugly :-(
True, simple math can look ugly
Try hard, think fast, die young
  • #43
  • Sep 6, 2006 3:39pm Sep 6, 2006 3:39pm
Ahhh. I think you guys are all talking apples and oranges. Bemac and Pawn aren't talking about Fixed Ratio methods, they are talking about Fixed Percentage methods of risk profiling. There's the big difference and the reason that Phil's equity curve looks so nasty. However, with fixed percentage methods, over the course of n trades, you could distribute wins and losses randomly and still end with the same balance.
  • #44
  • Sep 6, 2006 3:41pm Sep 6, 2006 3:41pm
Gotcha.

Thanks for taking the time to do that. I hope others have a clearer picture of what you are saying.

A- on the Spelink btw.
  • #45
  • Sep 6, 2006 3:43pm Sep 6, 2006 3:43pm
Quoting stockwet
Ahhh. I think you guys are all talking apples and oranges. Bemac and Pawn aren't talking about Fixed Ratio methods, they are talking about Fixed Percentage methods of risk profiling. There's the big difference and the reason that Phil's equity curve looks so nasty. However, with fixed percentage methods, over the course of n trades, you could distribute wins and losses randomly and still end with the same balance.
Yes. But what Phil is saying is it will require a different amount of Net Pips depending on the order {arrangement} of Wins, Losses.
  • #46
  • Sep 6, 2006 3:52pm Sep 6, 2006 3:52pm
Quoting Bemac
Yes. But what Phil is saying is it will require a different amount of Net Pips depending on the order {arrangement} of Wins, Losses.
Exactly. Because the NET amount that's risked over the life of the comparitive tests is different using Fixed Ratio than with Fixed Percentage.

If I run two tests, side by side, and use Fixed Percentage, then over the course of that test I would have risked the same amount regardless of the distribution of wins and losses. That's not the case with Fixed Ratio where the distribution does matter. In one test, I could have risked more overall than in another.
  • #47
  • Sep 6, 2006 3:52pm Sep 6, 2006 3:52pm
Quoting stockwet
Ahhh. I think you guys are all talking apples and oranges. Bemac and Pawn aren't talking about Fixed Ratio methods, they are talking about Fixed Percentage methods of risk profiling. There's the big difference and the reason that Phil's equity curve looks so nasty. However, with fixed percentage methods, over the course of n trades, you could distribute wins and losses randomly and still end with the same balance.
To some degree we weren't talking about the same thing. But to be clear, the two curves I posted earlier used fixed fraction which is the same as fixed percentage not to be confused with fixed ratio. i can select between 9 different money management techniques but the results could be just as disasterous.

Asymmetrical leverage applies to all money management systems, not just a particular algorithm.
  • #48
  • Sep 6, 2006 3:53pm Sep 6, 2006 3:53pm
Quoting stockwet
Exactly. Because the NET amount that's risked over the life of the comparitive tests is different using Fixed Ratio than with Fixed Percentage.

If I run two tests, side by side, and use Fixed Percentage, then over the course of that test I would have risked the same amount regardless of the distribution of wins and losses. That's not the case with Fixed Ratio where the distribution does matter. In one test, I could have risked more overall than in another.
I guess what I'm saying is that we're all agreeing, but, I think Phil and Pawn were talking around each other. Both had valid points and the difference was in the risk method.
  • #49
  • Sep 6, 2006 3:55pm Sep 6, 2006 3:55pm
Quoting philmcgrew
To some degree we weren't talking about the same thing. But to be clear, the two curves I posted earlier used fixed fraction which is the same as fixed percentage not to be confused with fixed ratio. i can select between 9 different money management techniques but the results could be just as disasterous.

Asymmetrical leverage applies to all money management systems, not just a particular algorithm.
Well, now, that's really interesting. I also ran a test using fixed fraction and didn't come up with the same equity curve as you. So, maybe you could shed some more light on your calculations.

The example from the book, though, was NOT a fixed fractional example. It was a fixed ratio example.
  • #50
  • Sep 6, 2006 4:01pm Sep 6, 2006 4:01pm
Quoting stockwet
Well, now, that's really interesting. I also ran a test using fixed fraction and didn't come up with the same equity curve as you. So, maybe you could shed some more light on your calculations.

The example from the book, though, was NOT a fixed fractional example. It was a fixed ratio example.
How in hell could you come up with a curve that was generated from my system? As explained earlier, the software took the 378 trades, applied a fixed fraction of 3% and then jumbled the sequence using monte carlo analysis. The point is that when using asymmetrical leverage with any money management system is applied the outcome is absolutely dependent upon the sequence in which trades occur.

You do not have enough information to definitively say that the Ryan Jones example was a fixed ratio example.
  • #51
  • Sep 6, 2006 4:07pm Sep 6, 2006 4:07pm
My bit.
Attached Image
  • #52
  • Sep 6, 2006 4:17pm Sep 6, 2006 4:17pm
Bemac,

One thing that occured to me is that your spreadsheet, while completely valid, does not account for contract sizes and that is where the problem that Ryan Jones explained comes into play. Let's say you could trade 1 mini lot per $1000 in your account. In that case, the results of your tests would be completely dependent upon the sequence of wins and losses.

Phil
  • #53
  • Sep 6, 2006 4:36pm Sep 6, 2006 4:36pm
Quoting philmcgrew
Bemac,

One thing that occured to me is that your spreadsheet, while completely valid, does not account for contract sizes and that is where the problem that Ryan Jones explained comes into play. Let's say you could trade 1 mini lot per $1000 in your account. In that case, the results of your tests would be completely dependent upon the sequence of wins and losses.

Phil
I think it that spreadsheet, he's working with the stop loss.
If you have fix 2% and fix 10$/pip... you have to adjust your SL.
Try hard, think fast, die young
  • #54
  • Sep 6, 2006 4:41pm Sep 6, 2006 4:41pm
Whooah, Nelly. This isn't a fight.

OK, let's get back to the orginal scenario, which was looking at a set of trades with 4:1 win loss ratio with x% success rate and randomly distributed wins and losses. That's how this whole thing got started. The only way that two scenarios, using fixed fraction mm would not be equivalent at end would be if any one of these variables were changed during the course of the test. If your win loss ratio is variable between the tests, then, you could get your equity curve. So, you're right, I can't replicate YOUR 378 trades, but, I can comment on the scenario and confirm what Pawn and Bemac were saying, which is that order does not matter so long as all other variables remain the same.

As far as not having enough information. Well, you're right. Sorry that I made an assumption that the Ryan Jones example was fixed ratio, but, given the excerpt from the book you posted, you can certainly understand my confusion.

Let's just chalk this up to me being completely confused and move on to another exciting topic!

Quoting philmcgrew
How in hell could you come up with a curve that was generated from my system? As explained earlier, the software took the 378 trades, applied a fixed fraction of 3% and then jumbled the sequence using monte carlo analysis. The point is that when using asymmetrical leverage with any money management system is applied the outcome is absolutely dependent upon the sequence in which trades occur.

You do not have enough information to definitively say that the Ryan Jones example was a fixed ratio example.
  • #55
  • Sep 6, 2006 4:44pm Sep 6, 2006 4:44pm
Quoting dof
I think it that spreadsheet, he's working with the stop loss.
If you have fix 2% and fix 10$/pip... you have to adjust your SL.
Which, if you modify the SL, thereby changes the win/loss ratio. As I said, I can't see how you get different equity results using fixed percentage unless you modify one of the other variables in the scenario. But, alas, my head is spinning.
  • #56
  • Sep 6, 2006 5:03pm Sep 6, 2006 5:03pm
Look at the Qty of Pips Required from the 2nd to & Last Trade to aceive the same balance.

Sory. There was an error in the Pip Totals in previous image.
Attached Image
  • #57
  • Sep 6, 2006 7:31pm Sep 6, 2006 7:31pm
Geez, I get busy for an afternoon and come back to find THIS!

Heheh.

Anyway, look fellas, everyone's got to keep in mind that forum postings, like emails, can be very easily misinterpretted or misread. That's why its important to over-explain everything. Talking about results - and this is directed at no-one in particular - without A) showing some kind of graph, spreadsheet, chart, or whatever, and B) explaining and over-explaining exactly how you reached those results, doesn't really do much for anyone. Assume first the people you are talking to don't know anything about FX, and we'll all go far.

What we're all doing here is abstract, theoretical, quasi-scientific studies based on mathematical formulae that are analyzing erronous historical data via highly imperfect methods. If we don't take the approach that at the end of our theses someone else should be able to closely reproduce our own work - or at least understand why they can't due to our own failings in approach - then misunderstanding will inevitably follow. Thus the reason I write such painfully long posts, like this one.

Beyond that, question yourself first and others later. I've found it to be a good working model for life in general. Even math disproves itself (read "Godel," Escher, Bach" for a quick summation), so as I said in one of the earliest posts in this thread:

...please, let's keep the thread clean, be respectful and respectable, and have an educational, informed conversation.

Is it OK to quote oneself? Heheh. I'm not trying to lecture anyone here... just want to keep the value as high as I feel it has been.

I made a mistake earlier when I posted that "fixed fractional" spreadsheet - I didn't explain exactly how I built it, and on what premises. I regret that... it was my mistake to post something while knowing that I didn't have the time to fully explain what I did there. So learn from my mistake.

TO CLARIFY THE "FIXED FRACTIONAL" SPREADSHEET:

My understanding, most probably incorrect, was that 'fixed fractional' meant that you took, say, a 2% risk on each transaction - 2% of your CURRENT account balance. So if you had $10,000, you'd risk $200. If you won and your balance was now $10,800 (assuming a 1:4 RR here), you'd then risk $216 on your next trade (2% of $10,800, not the original $10,000). I've actually read that's not a good idea, but regardless, that's what I assumed in that spreadsheet.

I also assumed a FIXED SL and TP in pips, not dollars. My feeling was, "If I've established that a 28 pip SL works and a 112 pip TP is attainable within my system on currency X, then I'm not going to mess with those numbers.

Instead, I varied the number of lots I traded in order to attain the risk in dollars based on a 28 pip SL... and then just calculate the reward in dollars by multiplying the risked dollars by 4 (again, 1:4 RR).

So, within my understanding, I presented there a system that was "fixed fractional," as 2% of the current balance was risked on every trade, and also "fixed ratio," as the R/R remained constant for every trade (1:4 or 1:3, depending on whether you looked at Sheet 1 or 2). It also maintained fixed SL and TP in pips. The only true variable employed was changing the number of lots traded as the current balance increased or decreased. And then the whole thing was applied to an absolutely impossible model where you always lost three trades and then won a single trade.

One of the most interesting points I've seen since I posted that was stated by Darkstar:

Now I’m no statistician, but the probability of hitting a 50% drawdown (8 consecutive trades) is SIGNIFICANTLY higher on a 25% system then it is on a 75% system, so the trader is all but forced to lower his risk-per-trade rate.


... followed right thereafter by Capt. Piptastic's 'probability of seeing X consecutive losing trades" chart. Both of these statements / presentations are mathematically irrefutable fact... for what that's worth. Heheh.

And allow me to clarify that *I* don't believe that % Success is irrelevant, which would seem an easy thing to misunderstand due to the nature of the original posting. Instead, my original point - and the point that I continue to believe - is this:

Before you spend time trying to find a 'tradable system' based on % Success, understand Money Management to the best of your ability.

... after you've done that, it's definitely time to research trading systems that produce a high probability of success... be that due to mathematical studies on historic data, easiness to actually work the system considering the time you have available, psychological factors as to how much drawdown you can handle, or anything else (all of which I believe to be very important factors and must be researched and understood in order to be successful).

This is a complex game, my friends. Unfortunate that its been sold as 'easy money' to the ignorant. If we all work hard, study hard, and share what we've learned, we can all get there... and the payoff is brilliant.



-miscon777
Et ecce equus pallidus...
  • #58
  • Feb 12, 2007 11:38pm Feb 12, 2007 11:38pm
Excellent post miscon,

The more I read informative posts from generous people such as yourself that are willing to learn while teaching others the more I realise that forex trading is a game of numbers and probabilities, and a thorough understanding of how these factors are related to real life trading is imperetive to success (more so than special indicators and systems).

I appreciate your thorough posts and look forward to reading more.

Regards

Dingo


Quoting miscon777
Geez, I get busy for an afternoon and come back to find THIS!

Heheh.

Anyway, look fellas, everyone's got to keep in mind that forum postings, like emails, can be very easily misinterpretted or misread. That's why its important to over-explain everything. Talking about results - and this is directed at no-one in particular - without A) showing some kind of graph, spreadsheet, chart, or whatever, and B) explaining and over-explaining exactly how you reached those results, doesn't really do much for anyone. Assume first the people you are talking to don't know anything about FX, and we'll all go far.

What we're all doing here is abstract, theoretical, quasi-scientific studies based on mathematical formulae that are analyzing erronous historical data via highly imperfect methods. If we don't take the approach that at the end of our theses someone else should be able to closely reproduce our own work - or at least understand why they can't due to our own failings in approach - then misunderstanding will inevitably follow. Thus the reason I write such painfully long posts, like this one.

Beyond that, question yourself first and others later. I've found it to be a good working model for life in general. Even math disproves itself (read "Godel," Escher, Bach" for a quick summation), so as I said in one of the earliest posts in this thread:

...please, let's keep the thread clean, be respectful and respectable, and have an educational, informed conversation.

Is it OK to quote oneself? Heheh. I'm not trying to lecture anyone here... just want to keep the value as high as I feel it has been.

I made a mistake earlier when I posted that "fixed fractional" spreadsheet - I didn't explain exactly how I built it, and on what premises. I regret that... it was my mistake to post something while knowing that I didn't have the time to fully explain what I did there. So learn from my mistake.

TO CLARIFY THE "FIXED FRACTIONAL" SPREADSHEET:

My understanding, most probably incorrect, was that 'fixed fractional' meant that you took, say, a 2% risk on each transaction - 2% of your CURRENT account balance. So if you had $10,000, you'd risk $200. If you won and your balance was now $10,800 (assuming a 1:4 RR here), you'd then risk $216 on your next trade (2% of $10,800, not the original $10,000). I've actually read that's not a good idea, but regardless, that's what I assumed in that spreadsheet.

I also assumed a FIXED SL and TP in pips, not dollars. My feeling was, "If I've established that a 28 pip SL works and a 112 pip TP is attainable within my system on currency X, then I'm not going to mess with those numbers.

Instead, I varied the number of lots I traded in order to attain the risk in dollars based on a 28 pip SL... and then just calculate the reward in dollars by multiplying the risked dollars by 4 (again, 1:4 RR).

So, within my understanding, I presented there a system that was "fixed fractional," as 2% of the current balance was risked on every trade, and also "fixed ratio," as the R/R remained constant for every trade (1:4 or 1:3, depending on whether you looked at Sheet 1 or 2). It also maintained fixed SL and TP in pips. The only true variable employed was changing the number of lots traded as the current balance increased or decreased. And then the whole thing was applied to an absolutely impossible model where you always lost three trades and then won a single trade.

One of the most interesting points I've seen since I posted that was stated by Darkstar:

Now I’m no statistician, but the probability of hitting a 50% drawdown (8 consecutive trades) is SIGNIFICANTLY higher on a 25% system then it is on a 75% system, so the trader is all but forced to lower his risk-per-trade rate.


... followed right thereafter by Capt. Piptastic's 'probability of seeing X consecutive losing trades" chart. Both of these statements / presentations are mathematically irrefutable fact... for what that's worth. Heheh.

And allow me to clarify that *I* don't believe that % Success is irrelevant, which would seem an easy thing to misunderstand due to the nature of the original posting. Instead, my original point - and the point that I continue to believe - is this:

Before you spend time trying to find a 'tradable system' based on % Success, understand Money Management to the best of your ability.

... after you've done that, it's definitely time to research trading systems that produce a high probability of success... be that due to mathematical studies on historic data, easiness to actually work the system considering the time you have available, psychological factors as to how much drawdown you can handle, or anything else (all of which I believe to be very important factors and must be researched and understood in order to be successful).

This is a complex game, my friends. Unfortunate that its been sold as 'easy money' to the ignorant. If we all work hard, study hard, and share what we've learned, we can all get there... and the payoff is brilliant.



-miscon777
  • #59
  • Last Post: Feb 13, 2007 4:58am Feb 13, 2007 4:58am
This is an interesting and honourable thread to which I would like to add my personal observations. I don’t think that anyone is dissin’ any particular aspect of money management. What I think we can all conclude and is that it is a vital element in trading system profitability. % success is important as is risk reward ratio and position sizing. The key is to get the right balance that suits your method of trading and is, most importantly, profitable. We can theorize and calculate as much as we like, but we all know that a trader’s take profit level is not always hit, so the traders target TP and the system performance, over a number of trades, are likely to be different.

What we need to do is to constantly monitor % success, risk reward ratio, position sizing and reduce the effects of human error involved.

I’m personally happy if my system achieves the following;

  1. %success = >50% (preferably 60%+)
  2. Risk reward ratio = >1:1 (preferably 1:1.5+)
  3. Position size = 1% of account balance per trade (max 3% at any one time) I trade the nearest available position size to my 1% max.

The effects of human error can be reduced by;

  1. sticking to your entry rules (if you miss an entry, wait for the next one),
  2. sticking to your stops, or get out before your stop once you realize your position is wrong (reduce the cost of the error).

In my experience, so far, the cost of human error is the biggest factor in reducing the overall profitability of a trading system. I’m not yet convinced that automated trading is best since it’s difficult to factor in news releases and major S&R levels.

Also theoretical calculations do not take into consideration timing. What may appear to be a poor decision short term may turn out to be a very good decision longer term and visa versa.

I read somewhere that trading system profitability can be assessed using a profit factor calculation which is;

Profit Factor (PF) = Av gain from trades X % winners
Av loss from trades % losers

Ideally a PF of 1.0 (break even) is the lowest acceptable and 1.5 or higher is desirable.

I know that it is possible to make consistent profits from currency trading. Knowing when not to trade is as important as knowing when to trade. It’s important to keep learning and more important to apply what you learn.

Follow the money. Follow the money. Follow the money.
0 traders viewing now