Money Management Mystique

[pedro01]

All

Every now & again I read a post on here about Money Management. Often accompanied with a comment to the effect that

'you can take a random trading strategy and apply good money management tehcniques and become the next buffet in a fortnight'.

Now - another thing I'll say is that if money management is your edge, then there is no reason to keep it a mystery as it's not as if you can cause yourself slippage by revealing your money management techniques. Revealing this kind of information will not erode your edge. You are just being selfish, you know it and if you dont tell me, I'm taking my ball home.

So - I mostly think that the people espousing these views don't have some secret (which of course, is never revealed) but are just regurgitating stuff they have read on the interweb.

On the other hand I have read a cheesily titled book which describes a long-term investment technique called " Automatic Investment Management" (AIM). This is more an alternative to dollar cost averaging than anything else BUT the mathematics behind it was fairly smart and for sure something I wouldn't have considered.

So - here I am - I am very skeptical of all of these people holding onto their big 'money management' secret. Am aware that they chances are I am already practicing what they preach BUT I am aware that maths whizzes can wrap their heads around numbers in a way that I can't.

Any thoughts on this ? Does someone want to show me what's inside the bag ?

 

[virtuos0]

You're way off the mark, most people agree that money management can make you the next Buffet in a week or less. If it takes you two weeks you just haven't properly grasped the power of money management.

 

[dcraig1]

All

'you can take a random trading strategy and apply good money management tehcniques and become the next buffet in a fortnight'.

How ? I've yet to see this convincingly demonstrated.

And Van Tharpe does not cut it. All the risk/reward stuff makes a probably unjustifiable assumption that R:R is constant over time and runs of wins/losses are normally distributed.

For me, a far more interesting topic is diversification, but this is a genuinely difficult problem which is probably why it is hardly ever talked about.

 

[pedro01]

How ? I've yet to see this convincingly demonstrated.

You - STOP THAT - I'll ask the questions around here. Basically though - that is my question.

Could this 'random entry + money management' mantra be just another trading 'truism' ?

And Van Tharpe does not cut it. All the risk/reward stuff makes a probably unjustifiable assumption that R:R is constant over time and runs of wins/losses are normally distributed..

I have the same issue with the Optimal f - effectively position sizing based on historical risk. How is this better than figuring out where your stop loss is & using that as your max risk and as a basis for your position size ? That is your actual risk per trade.

I would agree that stop-loss based position sizing doesn't take into account historical probability of the trade working out.

Saying that though, how do you, in real life, calculate historical probability to the degree of accuracy that it can be used to evaluate position sizes to the point where it really makes a positive outcome to your results ? I doubt any value that tries to assess validity/probability of a trading decision. Unless of course we are discussing automated trading, which takes away a lot of the human element. I am not, however talking about automated trading.

Saying that - I have seen a mathematical model that both suprised me AND made a lot of sense, so I cannot say that such things don't exist.

For me, a far more interesting topic is diversification, but this is a genuinely difficult problem which is probably why it is hardly ever talked about.

True enough.

 

[Trader333]

'you can take a random trading strategy and apply good money management tehcniques and become the next buffet in a fortnight'.

This is a little exaggerated in my view.

a far more interesting topic is diversification

In what context ?


Paul

 

[amnonco]

All

'you can take a random trading strategy and apply good money management tehcniques and become the next buffet in a fortnight'.

Let's change that a bit.

you can take a random entry trading strategy and apply good money management techniques and have positive expectancy over time.

now can you live with it?

 

[pedro01]

Let's change that a bit.

you can take a random entry trading strategy and apply good money management techniques and have positive expectancy over time.

now can you live with it?

Not unless you can put some meat on that bone, no.

It's become a nice cliche thus far but to add 'over time' is hardly padding it out with detail is it ?

 

[cuotes]

Anyway, we all agree that certain amount of money management is necessary, right? :|

 

[amnonco]

Not unless you can put some meat on that bone, no.

It's become a nice cliche thus far but to add 'over time' is hardly padding it out with detail is it ?

ok, let's take one step at a time.
would you accept that there are "all the time in the market" types of systems, that generate positive expectancy? such as, donchian channels, 2 MA, 3 MA, ATR Breakout etc...?

 

[pedro01]

ok, let's take one step at a time
would you accept that there are "all the time in the market" types of systems, that generate positive expectancy? such as, donchian channels, 2 MA, 3 MA, ATR Breakout etc...?

If you have a point, I think it is within the realm of what myself & the others on this board can understand. There's no need for 'baby steps' here. Just go ahead & state your case. When I read it & don't understand it, I will ask questions :smart:

What I will say in response to your question is that you have diverted from 'random entry'. A breakout system is in no way a random entry, it is a methodology that can be tested and certain conclusions can be made about the probability about a trade being sucessful based on past results. Depending of course, on how much you believe in such things.

On the other hand, a random entry could be 2 darts, 1 in a list of stocks and 1 in a list of entry times.

There is no need to debate non-random entries when the cliche so often espoused is that of a random entry being able to produce positive expectancy when combined with good money management.

Can anyone demonstrate how, with a random market at a random time, a profit can be made purely with the use of money management alone ?

 

[everyonerich]

ok, let's take one step at a time.
would you accept that there are "all the time in the market" types of systems, that generate positive expectancy? such as, donchian channels, 2 MA, 3 MA, ATR Breakout etc...?

YES THERE IS. price will either trend or range. Thats all there to it. It really is that simple but clearly not that easy. These two discrete properties of price require diametrically opposite mind-sets and money management techniques. Knowing when to apply each is what makes trading so difficult.

 

[amnonco]

If you have a point, I think it is within the realm of what myself & the others on this board can understand. There's no need for 'baby steps' here. Just go ahead & state your case. When I read it & don't understand it, I will ask questions :smart:

What I will say in response to your question is that you have diverted from 'random entry'. A breakout system is in no way a random entry, it is a methodology that can be tested and certain conclusions can be made about the probability about a trade being sucessful based on past results. Depending of course, on how much you believe in such things.

On the other hand, a random entry could be 2 darts, 1 in a list of stocks and 1 in a list of entry times.

There is no need to debate non-random entries when the cliche so often espoused is that of a random entry being able to produce positive expectancy when combined with good money management.

Can anyone demonstrate how, with a random market at a random time, a profit can be made purely with the use of money management alone ?

obviously you missed the point.
the claim I make is based on you accepting all the premises.

1. there are "all the time in the market" trading systems, that over time generate positive expectancy.
2. there are money management rules that can rely on such systems. such as ATR based stops/position sizing/pyramiding, or position sizing based on % from X days low/high (donchian based) etc.
3. your first entry can then be random, but
4. because the system is all the time in the market, when it first stops out it will become exactly one of the systems in (1).

therefore,
despite the fact, that you can (and probably will) lose on your first entry, over time, the system will generate a positive expectancy.
OK, so there was only 1 entry. and it was random. and the system then became a profitable one.

hope I'm clear. of course this is academic/philosophical, but so is the nature of this debate.

 

[pedro01]

Good post Amnonco.

First of all, I would question whether there really are 'always in' systems that generate a positive expectancy over time, especially when you consider fees and slippage. Death of a thousand cuts comes to mind. I have seen a few such systems based on a single instrument and they were weak but I admit I have never looked at always in systems that are more exotic/based on multiple instruments.

In any case, even an always in system is in no way a random entry as it has specific rules on when to reverse. The key with random entry is that the expectancy is break even. So - how do you take a break even system & make it profitable with money management. For sure, it is not with the use of commonly used techniques such as volatility based stops.

The problem with money management is that people talk about how great it is and how it is the key to success but in generalities. Therefore either :
- people don't really know about it and are just saying it to sound clever
- they are just doing what everyone else does - i.e. common money management techniques
- they know something so earth shattering (as they imply) that they will not reveal it

I used the example of AIM as it is an unambiguous mathematical model that can be clearly tested and demonstrated. A reasonable description of it is here : Core Position Trading™ (CPT), Robert Lichello's Automatic Investment Management (AIM) Basics (Stocks, ETFs, and Mutual Funds)

AIM is an investment method where you do not put all of your money into the stock. You keep some in the stock & some in cash. As the price fluctuates, you switch more money to the stock or sell stock and put money in the cash reserve. It is the way you manage the cash/stock switching that sees you make more money that you would have done had you been fully invested in the stock IN MANY CIRCUMSTANCES. Note that I have not fully tested it in all circumstances, I think the downside is in the upside or rather that you would make less return on those stocks that are rising without pullbacks.

AIM is a complete money management system. It is certainly something I would not have thought up myself. It is reasonably clever and it demonstrates what can be done with the application of some math theory.

I have yet to see something similar in the field of shorter term trading that would allow us to profit from a break even strategy.

 

[amnonco]

Good post Amnonco.

had to leave that :cheesy:

First of all, I would question whether there really are 'always in' systems that generate a positive expectancy over time, especially when you consider fees and slippage.

If I remember correctly you have something against what van tharp offered. but if you agree that long term trend following works, than 1 random entry, with reversal at the stop - works as well. I back tested this one...in a long term system, slippage and fees are really marginal

The key with random entry is that the expectancy is break even. So - how do you take a break even system & make it profitable with money management. For sure, it is not with the use of commonly used techniques such as volatility based stops.

completely disagree on two points.
1. random entry system doesn't necessarily have break even expectancy. I give you that the average random entry system has a break even W% (i.e 50%)
2. basic money management techniques can and will improve expectancy on any system. never mind if it's a random entry. it's just a matter of understanding the basic principals of the system. remember, a random entry system, is not a random entry and random exit and random rules.

The problem with money management is that people talk about how great it is and how it is the key to success but in generalities. Therefore either :
- people don't really know about it and are just saying it to sound clever
- they are just doing what everyone else does - i.e. common money management techniques
- they know something so earth shattering (as they imply) that they will not reveal it

how about this - it's not an either or issue. some just saying it to sound clever, some using what we all use, and maybe there are genius traders who use something special.

personally, I believe in simplicity, a small number of parameters, basic money management. I use what works.
but for our little debate, yes, even the basic money management principals can improve a 50% W% system to a positive expectancy system. hell good money management turn a lower than 50% W% system to a profitable one (the basic trend following).

as you can understand, I follow trends

 

[tommog]

this is one of my favorite topics that im constantly thinking about. It is true you CAN have a winning system with "almost" random entry (completely random is quite hard to achieve) and money management alone, however, before you get too excited, the equity curves are appalling, nobody in their right mind would trade them. Money management will improve just about any approach to the markets you choose, so studying it is an absolute must. Unfortunately there is very little good literature on the topic

 

[1Pipped]

Can anyone demonstrate how, with a random market at a random time, a profit can be made purely with the use of money management alone ?

For the system to have positive expectancy it must win more often than:
Average loss / (Average loss + Average win)

Lets say you always risk one unit and always win one unit, to have positive expectancy the system must win more often than:
1/(1+1) = 0.5 .... 50% of the time.

Which might not be impossible in itself, but lets just say you win exactly 50% of the time so you break even (random market, random entries, no costs etc.).

What you are looking for is a Money Management Mystique function that modulates Risk (maybe via some complex proprietry method) so that:
Risk / ( Risk * 2 ) < 0.5 ..... so that your 50% win rate would produce a profit.

If you found what you were looking for, you would have found something that makes a number equal less than half, of itself multiplied by two!



.

 

[amnonco]

Lets say you always risk one unit and always win one unit, to have positive expectancy the system must win more often than:
1/(1+1) = 0.5 .... 50% of the time.


.

why should we say this?
we all agree that random entry strategies have 50% W%
but who said anything about risk:reward ratio?
what if you can, using random entry and simple MM rules, generate more than 1:1.
let's say 1:1.2 is it that difficult to think about?

 

[pedro01]

OK - let me tell you my experience of random entry. I will provide code later to present a totally random system if necessary. I will also provide code that shows a 'non-random' entry that shows similar results to a random system which effectively means the system has no edge.

Also - I will say that the 'always in' system that has a +ve expectancy that can be exploited by the common man is a myth. If it can be demonstrated I'll believe it. If someone wants to say they have a MA xover system that has a +ve expectancy over time, please present it. Otherwise, we are merely discussing theory & not practise. Of course I don't expect anyone to present a working model if they have one, they can jut present the results of executing such a model.

Let's talk about units of target stop loss. They may be ATRs, points,ticks etc. Lets say you have 4 units stop loss and 2 units target. I have seen systems that supposedly have some 'brains' behind them, use this 2x target, 4x stop loss and produce 66% winners. Of course what this means is that the 'brains' behind the system has no edge. The winners are half the size of the losers, so you end up making nothing.

So - when I talk about random entries, it doesn't have to be random. No-one sets out to make random entries. What I have seen is Stop loss:Target ratios that produce winners in the same ratio, effectively having zero edge. Could Money Mangament give such systems an edge ? Well - people imply on this forum that they can. I have my doubts but like I say - I have been impressed by the works of mathematicians enough to know that I can't rule anything else.

Another random system we could us is enter every day at 10:00 am. Stop loss = 2 x 10 min ATRs, Target 2 x 10 min ATRs. Alternate between longs & shorts. If necessary I can put such a model together & present the results. I am sure if run over a week, there may be a profit or loss but run over a year, probability says your trades will come out even

Now - some of you may say "Wow - 66% winners". Problem is if your entry is random and represents no edge, then the expectancy when your stop loss is twice your target is that 2 thirds of your trades will be winners.

Of course the 4x target & 2x stop loss system has 66% winners because of the laws of probability, which I have seen in action and am willing to demonstrate if it will also encourage on the other side to demonstrate that money management can be used to make a random system profitable. My guess is that this is all hot air.

 

[1Pipped]

why should we say this?
we all agree that random entry strategies have 50% W%
but who said anything about risk:reward ratio?
what if you can, using random entry and simple MM rules, generate more than 1:1.
let's say 1:1.2 is it that difficult to think about?

If you are winning 50% of the time at 1:1.2 , and breaking even, you are doing something very wrong, Sir. Those numbers have a positive expectancy.

 

[Splitlink]

YES THERE IS. price will either trend or range. Thats all there to it. It really is that simple but clearly not that easy. These two discrete properties of price require diametrically opposite mind-sets and money management techniques. Knowing when to apply each is what makes trading so difficult.

I can't see that "knowing when to apply" is, then, random entry.