Best Thread Position Sizing & Compounding: the Keys to the Magic Kingdom !

BSD

Veteren member
Messages
3,819
Likes
988
I was reading a journal by an obviously extremely talented trader here who isn't maximising his potential anywhere near to the kind of money he could and actually should be printing.

The problem very obviously is not his trading method which is very successful, putting it mildly.

No.

The reason why he is nowhere even close to where by all rights he should be is due to a different reason altogether.

It's a lack of proper position sizing combined with a lack of compounding vs the excellent expectancy of his method.

That's why I decided to pen this, market volume having dried up as it is anyway what with Christmas round the corner and everybody closing their books.

So here goes.

Provided you have a method or system with a positive expectancy then the single most important determinant of your overall success will be position sizing and compounding, two of the most overlooked sizes in most peoples neverending quest for a holy grail of perfect entries and exits.

Let's quickly go through the success relevant factors driving your P&L:

Expectancy of your system quite simply consists of the following ingredients:

I: Divide average size of your winners vs average size of your losers = Win / Loss Ratio

II: Percentage of your trades that end up > breakeven = Win Probability

III: Transaction costs = Commissions + Slippage

IV: Trading opportunities = how many signals per month, week, day, hour, etc your system generates.

You can play around with some of those parameters here and see how your expectancy stacks up:

Random Equity Curve Simulator of a trading system. Learn it before you trade

A positive expectancy is obviously a prerequisite, no positive expectancy, no net profitability.

But provided we have a system with net profitability - and don't forget that that is achievable with a system that on average has 33% winners, and a risk / reward ratio of 1 / 3 - then the next step will be position sizing and compounding.

This alone will determine the extent of your success.

Keeping in mind that risk and reward are correlated, that the slope of your equity making new highs will have corresponding drawdowns, it's down to you and your pain threshold to determine the amount of money you will be printing from now on, where you will be in the next year or in ten years.

Of course everything has an optimum, and many will miscalculate what they are willing to endure.

Backtesting something on paper and seeing that if you are willing to risk X % per trade that that will generate XYZ in returns one year down the road, but also means you will have to endure a 50% drawdown, is ONE thing in theory, and a totally different matter in practise, depending on what you can or, more likely, cannot handle.

Another thing that needs to be kept in mind is that outlier events are not a question of IF, but only of WHEN.

You WILL have losing streaks you would never have believed possible if you survive long enough in this game to experience them, if you haven't blown up or given up before.

Risk of ruin isn't just a theory, it's a very real threat.

Cliché, but true, there are old traders, there are bold traders, but no old AND bold traders...

Enough about the risks then, lets move on to the position sizing options themselves.

The key factor behind the money you will make isn't an eternal search for fail proof entries or woulda-coulda-shoulda exits that took you out at the top hehe, it's simply how big you trade and if you can keep compounding your position size in an anti-martingale style - ie where you get bigger as your equity grows, and smaller when it shrinks - with every new position you put on.

The mechanics behind fixed fractional position sizing are straightforward enough:

First you determine where you want to enter.

Then you determine where you will be stopped out with a loss if the market goes against you.

And the distance between your entry price and your stop loss will then always equal a position size that constitutes the exact same percentage amount of your account.

Meaning that if your trade gets stopped out with a loss that will then always equal (excluding slippage or gaps during adverse or fast market conditions) whatever % of your equity you are willing to lose per trade, 0.25%, 0,5%, 1%, 2%, 3%, or whatever seems justifiable per your systems expectancy and your ability to handle drawdowns.

In absolute terms total amounts of money won or lost will obviously vary along with growing or shrinking equity levels in your account as you go through losing or winning streaks, but the percentage amount which is really all that counts until eventual liquidity issues will force you to start trading smaller and smaller percentage wise will always be exactly the same (adverse market conditions apart as mentioned).

The inherent beauty and amazing power of this method of fixed fractional position sizing is that it adapts perfectly to your current situation, when things are going your way you get bigger and bigger with every single trade, harnessing the magic of compounding to the hilt, whilst when things are going against you, you are reducing your position size with every trade thereby ensuring your survival - your most important objective in trading - and that you will be around to capitalize on better market conditions for your method when they arise.

Market Wizard Ed Seykota of the hundreds of thousands percent returns has a great piece about money management here:

Risk

Not a great fan of this guy who basically leeched off of others he never had any connection with but his summary here is good nevertheless:

Money Management | TurtleTrader

And this guy who trades for a living goes into great depth:

Mechanical Trading Systems - Money Management

Once you have a method with a positive expectancy then position sizing and compounding combined are the only real Holy Grail, the only real Keys to the Magid Kingdom !
 
hands down the best piece of information i've read about working in position sizing and compounding
FX Thoughts
 
Last edited by a moderator:
I'm a medium/long term trader, and whilst the advice above may make sense for day traders, I would disagree with any system that looks to place a fixed % of capital on the next trade.

The most important factor when determining trade size is recent volaility of the underlying and the VaR this implies. Without knowing the details of Markowitz it still makes sense that you should bet smaller amounts in more volatile markets.

As an example, the recent volatility of GBPUSD is around 3 times that of Dec 2006 (by the measure that I use). This means that for the same VaR I can take on a position 3 times smaller than before but my expected return is the same. It would be misleading for me to always use the same trade size and then compare my monthly returns, as at the moment I would be taking 3 times as much risk as two years ago.
 
I'm a medium/long term trader, and whilst the advice above may make sense for day traders, I would disagree with any system that looks to place a fixed % of capital on the next trade.

The most important factor when determining trade size is recent volaility of the underlying and the VaR this implies. Without knowing the details of Markowitz it still makes sense that you should bet smaller amounts in more volatile markets.

As an example, the recent volatility of GBPUSD is around 3 times that of Dec 2006 (by the measure that I use). This means that for the same VaR I can take on a position 3 times smaller than before but my expected return is the same. It would be misleading for me to always use the same trade size and then compare my monthly returns, as at the moment I would be taking 3 times as much risk as two years ago.

IMO correct placement of stops includes an adjustment for increased volatility - put simple, in a thin / choppy market, there is a case that the sensible place for a stop includes an adjustment for the volatile market conditions. Then the position size is amended accordingly for VaR purposes... the increase is already accounted for in the placement of the stop - to adjust for VaR again would compund the "volatitility adjustment".

(... one thing I will add to BSDs excellent post is that another "key to the magic kingdom" is scaling in and out of positions, but this is just my humble opinion).
 
IMO correct placement of stops includes an adjustment for increased volatility - put simple, in a thin / choppy market, there is a case that the sensible place for a stop includes an adjustment for the volatile market conditions. Then the position size is amended accordingly for VaR purposes... the increase is already accounted for in the placement of the stop - to adjust for VaR again would compund the "volatitility adjustment".

(... one thing I will add to BSDs excellent post is that another "key to the magic kingdom" is scaling in and out of positions, but this is just my humble opinion).

I'm sure you're correct, I was just thinking with my long term trading hat on. As I trade systematically I don't use stops - I just trade the signal. In this type of trading volatility adjustment of position size is a natural thing to do.
 
I'm sure you're correct, I was just thinking with my long term trading hat on. As I trade systematically I don't use stops - I just trade the signal. In this type of trading volatility adjustment of position size is a natural thing to do.

Oh I see - sure, if you are position trading without stops, then the process you describe is wholly appropriate.

:D
 
Nice work Markus - good post!

I have tried my hand at day trading U.S. equities full time for the past year and failed in so far as I've not made any money. However, I was determined not to repeat the mistake I made about 4 years ago, whereby I blew 70% of my account on just one trade lasting 24 hours, simply because I failed to implement any basic risk and money management principles. However, I've learnt from that mistake and managed to keep the losses small because I implemented the exact same risk and money management techniques that Markus outlines in his post. When I hit my maximum permitted drawdown I stopped trading. It's not what I wanted and I'm pissed off about it, but the net result is that my account remains largely in tact and I live to fight another day.
:)
Tim.
 
Excellent post Markus.

I have to admit that I don't capitalise on my profitable trades as much as I should do. My strike rate is high and I am a successful day trader. However, I know that to be more successful, I must change gears in accordance with the principles which you dexcribe.
 
IMO correct placement of stops includes an adjustment for increased volatility - put simple, in a thin / choppy market, there is a case that the sensible place for a stop includes an adjustment for the volatile market conditions. Then the position size is amended accordingly for VaR purposes... the increase is already accounted for in the placement of the stop - to adjust for VaR again would compund the "volatitility adjustment".

I don't know what the VaR stat implies, but you should technically carry on applying the fixed fraction of your capital to the trades you are making, provided the volatility lies within the bounds of the trading conditions you based your risk assessment on.

In this sense, you should see the volatility as a part of the pattern you have identified which gives you your edge.

Of course if you were trading on fundamental analysis or with a TA approach irreducible in terms of programming, then you'd want to do what you suggest.
 
Hi,

I have a question regarding fixed fractional sizing and i cant work out what the best course of action is.

I have some longer term trades (hold for up to 6 weeks) and some shorter term trades (intraday).
As the shorter term trades are opened and closed throughout the day it is simple to work out what position size to put on, however, as my account increases (or decreases) should i be adding to, or reducing my position size on my open trades?

Or, do i only factor account size and therefore position size when i open the trade? My concern is my portfolio compliments itself quite well with shorter term trades and longer term trades running simulataneously, however, if i have a good winning streak on the shorter term trades and therefore increase the size significantly this runs the risk of my longer term trades being dwarfed by the shorter term positions in terms of size therefore losing the balance to my portfolio,

Is there any standard proceedure on this or is it just personal preference?

thanks
 
Hi,

I have a question regarding fixed fractional sizing and i cant work out what the best course of action is.

I have some longer term trades (hold for up to 6 weeks) and some shorter term trades (intraday).
As the shorter term trades are opened and closed throughout the day it is simple to work out what position size to put on, however, as my account increases (or decreases) should i be adding to, or reducing my position size on my open trades?

Or, do i only factor account size and therefore position size when i open the trade? My concern is my portfolio compliments itself quite well with shorter term trades and longer term trades running simulataneously, however, if i have a good winning streak on the shorter term trades and therefore increase the size significantly this runs the risk of my longer term trades being dwarfed by the shorter term positions in terms of size therefore losing the balance to my portfolio,

Is there any standard proceedure on this or is it just personal preference?

thanks


I do believe it is up to you. Personally I do all my calculations when the trade is first opened.
 
You would have to reduce your position size in the long term trade if you were using high leverage, simply because you might wipe yourself out if you don't. If your account can still handle your worst case scenario, then don't change position size. I can only see disadvantages. I would cost extra commission and slippage, it would make analysis of the returns on your long term system much more complicated, the effects would have random results, presuming your system only gives entry and exit signals.
 
I agree totally with your comment on the next level for successful traders after developing a trading plan is capital management. Now some might say that capital management should be apart of your trading plan. Either way, this was a great post.
 
I've enjoyed this thread and I'm trying out Optimal F in a spreadbet account - great for infinitesimal changes in postion sizing.

Procedure is:

Start with 4% of capital, just as an arbitrary assumption.

Run 30 trades and calculate an Optimal F using the solver function in Excel.

Run another 30 trades and recalculate for the entire 60 trade sample.

Continue refining with each sample.

There's another thread somewhere by Mathemagician where he argues it's far more important than entries and exits, so it'll be good to see if it works.

There's a free postion sizing game on the Van Tharp website, for anyone who's interested.

Trading Simulation Game
 
  • Like
Reactions: BSD
Another route for calculating optimal position size is the Kelly Critereon:

1) Express your returns on each trade as Log returns

2) F* = (Mean of returns) / (variance of returns)

3) then use "half kelly" @ F*/2 - F* is the optimal, half kelly is to reduce the risk of total catastrophe - it halves the volatility of your returns but only decreases your growth by 25%


************

there are other ways of calculating F* - the most common alternative is:

F* = [(Avg.Win - Avg. Loss)*P(win)] - P(loss)
(Avg. Win - Avg. Loss)​


EDIT: Note that the Kelly Criterion is the optimal position size to maximise your expected long term growth rate... there might be something else which you want to optimise - for example, you might want to minimize drawdowns, or the variance of your returns, wealth maximisation, and so on...
 
Thanks MrGecko.

I was unsure of whether a binary outcome was required to use the Kelly formula. I'll calculate it and hopefully the results will be in line with the sim result.
 
I think the "half" bit MrGecko refers to is important for another reason - when I tried a spreadbet the other day (S&P 500 Daily) it went 1.4 through my stop!

Bloody criminal, but also dangerous in that the slippage can become huge (as a proportion of your risk) the tighter the stop you use.
 
Last edited by a moderator:
The following results were obtained in October 2012.

29.28 % profit on a $100,000 account using several trading systems .Traders have to devise a system like it to start with.

Compound the weekly profits .
It is possible to compound this on a weekly basis @ 7.5 % per week and turn it into $4m within 1 year , and to over $1bn within 3 years .

Who has a sound method /system to do this consistently?

Warning :New traders should not attempt it , but learn the basics.

The forum is full of successful traders with great methods.
 

Attachments

  • returns.xls
    68 KB · Views: 310
Top