The 2% Rule and leverage

[im sparticus]

tech do you have any opinions on the 35% margin to equity ratio (adaptive analysis) in managing the number of open positions or total exposure?

 

[Nick Radge]

Great topic TH and one that many just don't grasp. There are a few of these key issues that should be grasped before people start trading, but they're for another day.

The level of exposure to the market is important if one is trading highly correlated instruments and one has many open positions in one direction. TH may not get into this scenario trading intra day in the Hang Seng and SPI, but a CFD trader may well do so.

In the professional funds management business, as a rule of thumb, a margin to equity ratio of 30% is a level at which leverage starts to be seen to be excessive. Now that's for exposure in one direction and its a conservative estimate.

For retail traders such as us we can take that higher, say 50% or so.

However. when one 'balances' their book, that is, holds a balance of longs and shorts then this level can be taken higher.

As an example I currently have 16 open positions across ASX and US equities, but I have 8 longs and 8 shorts. My margin to equity ratio is high, although not extreme, but I am comfortable in knowing that any major market move will more or less see my account be hedged.

Without a significant market move each positions does its own thing within its own trend which is the idea but I will be relatively protected in a collapse as Kauri points out.

The problem with having a balanced book is that its hard to achieve trading the ASX and secondly all position on the ASX tend to move in tandem, unlike the US.


This post may contain advice that has been prepared by Reef Capital Coaching ABN 24 092 309 978 (“RCC”) and is general advice and does not take account of your objectives, financial situation or needs. Before acting on this general advice you should therefore consider the appropriateness of the advice having regard to your situation. We recommend you obtain financial, legal and taxation advice before making any financial investment decision.

Past performance is not a reliable indication of future performance. This material has been prepared based on information believed to be accurate at the time of publication. Subsequent changes in circumstances may occur at any time and may impact the accuracy of the information.

 

[im sparticus]

Thanks for the reply nick,

So basically if i give my shorts a negative number and my longs a positive as long as my total doesnt exeed 30% (or -30% if im heavier on the short side) im still within the boundary?

Also TH braught up a good point of different providers offering different margin requirements (some as low as 1%) how would this affect the 30% are there round figures on margin requirements we should stick to ie 5% for an index and 10% for a top 20 and ignore how much our provider is "actually" holding as margin?

 

[Nick Radge]

Personally I don't look at the margin/equity ratio and I never look at the individual margin. Common sense says that too many positions in one direction will hurt you eventually.

Don't read too far into this stuff. Just understand the common sense approach. If you're carrying 10 long positions and the Dow tanks 500 you're going to feel it.

 

[im sparticus]

Anyone ever thaught of using the "kelly criterion" for working out the amount to place at risk, sound like very agressive position sizing method but apparently its the best mathematicly proven way to get max return from your investment when playing a game where you have a positive expectancy, apparently using the method its impossible to do your account though it does add for a bumpy ride might work well with a gsl(obviously you will still bust without a gsl if the price gaps beyond your max amount at risk)?? any thaughts.

 

[im sparticus]

you can play with this excel sim

Put in 6% risk per trade and run the sim 5 times. You will blow up on my testing 2 out of 5 trials even with a positive expectancy.

this is mathematically impossible as everything is infinetly divisible even with a 100% negative expectancy!

 

[tech/a]

You may wish to download and play with this matrix.

http://www.artoftrading.com.au/tools/DownloadView.asp?iddownload=44&idCategory=5&topcategory=5

 

[im sparticus]

any thaughts on kellys or half kellys criterion tech? i know its very popular with blackjack card counters for money managment and there whole strategy is exploiting a very tight edge (around 1%) just though money managment (there play remains mostly unchanged with changes in position sizing once the edge presents itself accounting for 70-90% of the profits) imagine what it could do for someone with the sorts of positive expectancys people around here quote. (kellys criterion allows them to put up alot more % wise than that matrix would have you believe would send you bust per trade)

 

[tech/a]

Had to look it up myself!

http://en.wikipedia.org/wiki/Kelly_criterion

But would be highly correlated to strings of losses.

While it has the advantage of maximising gain it has the drawback of deminishing the capital base in large chunks.

Personally I look at R/R and if I can maximise Reward to Risk then Compounding and Leverage will maximise my return---based upon R/R alone.

There are ways of adapting a less aggressive strategy and then ofcourse we would have to be specific on each trade relative to the results of all trades current and still open.

If we are only taking closed trades as the base "f" then it is simpler.
If portfolio value then this will vary.
Ive never used it and would suit someone who has a very defined trading system with a very clear blueprint.

The sort of discipline/Model to be able to practically deploy this form of capital maximisation would be rarely found outside Instos,I would suggest.

 

[glenn_r]

A report on a ASX50 system using OT's portfolio simulator, it is set at defaults and is just for illustration purposes only.




The Fixed Trade Size allocation method always attempts to trade a fixed number of shares.

The Fixed Dollar Amount allocation method always attempts to trade a fixed dollar amount per each trade

The Percent of Equity allocation method always attempts to trade a fixed percentage of the account’s current equity. As equity rises, the amount invested per trade also goes up in proportion to the equity amount

The Size to Equity allocation method begins by trading with a fixed trade size in shares or contracts. A series of equity levels defined by the parameters trigger allocation changes as they are crossed. Trade size will increase and/or decrease based on when the gains or losses of the Trigger Amount are realized.

The Kelly Criteria allocation method attempts to find the optimal allocation size (as a percentage of equity) based on J. L. Kelly’s famous formula. The new portfolio simulator implements this calculation on a bar by bar basis, allowing position size to be calculated without hindsight.
Parameters
Initial Amount—Specifies initial allocation amount in dollars. This amount is only used during the warm-up period, while the Kelly Formula calculation is warming up.

Warmup Trades—Specifies the number of trades which must complete before the Kelly Criterion is applied to the trades.

Multiplier—Specifies the multiple of the Kelly allocation which is used for the actual position sizing in the simulation. This allows users to replicate popular techniques such as ‘Half Kelly’.

The Optimal f allocation method attempts to maximize equity growth. The method first finds the optimal allocation (as a fixed percentage of equity) based on all the trades in the simulation. The method will then apply the optimal fixed fraction to the complete series of trades.

 

[tech/a]

Yes I can see myself placing an average trade size of $3.8 million.
Plus a 72% drawdown.

Nice in theory!

 

[glenn_r]

The test was set @ 10% margin if your wondering about the big figures.

 

[Trembling Hand]

The test was set @ 10% margin if your wondering about the big figures.

10 % of what??

 

[Nick Radge]

I'm well versed in Kelly %. One can get quite creative with position sizing but the bottom line however remains the same; the more you bet, the more volatile your account will become.

 

[glenn_r]

10 % of what??

Sorry been away at a work seminar at Glenelg, 10% margin means you only need $10 of your own money to trade $100 of CFD's, Equities etc.

So when you set 10% margin as a parameter in a back test it will magnify the returns and losses of your capital x 10.

 

[Trembling Hand]

Sorry been away at a work seminar at Glenelg, 10% margin means you only need $10 of your own money to trade $100 of CFD's, Equities etc.

So when you set 10% margin as a parameter in a back test it will magnify the returns and losses of your capital x 10.

So its complete rubbish as far as sensible MM

 

[glenn_r]

So its complete rubbish as far as sensible MM

I was just illustrating different position size techniques nothing more or less, personally I use % of equity.

 

[doogie_goes_off]

Glenn r, I am not a trader, I just came up with my own optimum rules, don't you think it's going overboard to find a formula for the optimum trade size/value rather than just finding out what works for you, particularly based on your income (if that's how you fund it?).

 

[WaySolid]

One parameter study I pay attention to is %fixed fractional risk vs Maximum DD on monte carlo runs of existing trades, add a fudge factor to the max allowed drawdown you will tolerate, find your allowed %risk and position size accordingly.

It's as much art and experience as math; as the data series you are working with as a trader is nonstationary and hopefully continuing to get better with some rough patches thrown in.

I can't imagine risking 2% of capital per bet, If you run the Max DD v FF% parameter study and have a look at what type of DD's you need to tolerate then that might give you a heads up, and you can assume reasonably safely that an out of sample draw down will prove to be your largest ever.

 

[Nick Radge]

You can actually calculate your maximum risk per trade based upon your drawdown tolerance.

As an example, if your trading win rate is 50% and your maximum drawdown tolerance is 25%, then your maximum risk per trade to remain within that tolerance is 1.78%.