Risk Management: The 2% (or 1%) Rule

Chart Forum » Trading - Systems » Risk Management: The 2% (or 1%) Rule

«« Previous Next »»

Author

Message

colin_twiggs

Member

Post Number: 2899
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, June 01, 2007 - 05:10 am:

Mahogany,

I am interested in how you limit your exposure by sector. I have been doing some work on this myself.

Regards,
Colin

mahogany

Member

Post Number: 10
Registered: 01-2005

Rating: N/A
Votes: 0


Friday, June 01, 2007 - 10:54 am:

Colin
I have a very simple rule as part of one of my strategies. No more than 40%, of my total number of stocks in that portfolio at any time, can be in the same sector. Not a
whole lot of rocket science behind it I'm afraid, but it is just an attempt to put a little bit of conservatism in an otherwise medium risk type strategy.
I currently have 17 stocks in this portfolio of which 6 are in Materials and another 6 in Industrials. If a further stock in either of these sectors came up for
selection I would sell down one that was lower in my selection hierarchy, before I add the new selection.
Regards, M

colin_twiggs

Member

Post Number: 2900
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, June 01, 2007 - 11:43 am:

Thomas Dorsey in Point & Figure Charting gives an example of the risks affecting a typical stock:
Market risk 66%
Sector risk 24%
Stock risk 10%

Now these risks vary from stock to stock, but I carry this around as a rule of thumb:
Market risk 50%
Sector risk 25%
Stock risk 25%

In other words it is more important to get your market and sector timing right, than your actual stock timing. If you had to limit your capital at risk to 1% (some traders use 2%) in an individual stock, then it may be advisable to limit your sector risk to no more than 2 or 3 times that, at any one time.

How the 2% rule works is that your investment in any sector can be as high as 40% or 50% - so long as only 2 or 3% of your capital is at risk.

Regards,
Colin

(Message edited by colin_twiggs on June 01, 2007)

mahogany

Member

Post Number: 11
Registered: 01-2005

Rating: N/A
Votes: 0


Friday, June 01, 2007 - 01:20 pm:

Colin

I have only used "Market risk" to determine the amount of capital that I allocate to this strategy (currently my
assessment of the risk in the market has resulted in me allocating 25% of my total trading capital to this portfolio).

I then assess the current risk in each sector as Low, Medium or High. I use this to determine how much to risk
for each individual trade. (a LR = 1% of portfolio trading capital risked on entry & HR = 0.4%)

At the same time applying the rule that there can be no more than 40% of the total portfolio stocks in any one sector.

Regards, M

colin_twiggs

Member

Post Number: 2901
Registered: 09-2002

Rating: N/A
Votes: 0


Saturday, June 02, 2007 - 09:03 am:

Thanks Mahogany,

Most risk management strategies are based on the assumption that stock returns are normally distributed (in a bell-curve around a mean). One of the key requirements for a normal distribution is that individual outcomes are random and independent of each other. We know that this assumption is false for the stock market as probably the biggest threat to our portfolios is co-variance: the tendency of stocks, sectors and markets to all move in the same direction at the same time. Individual outcomes are not independent.

I have been working on strategies to address co-variance for several years and would appreciate the views (and risk management strategies) of other readers.

Regards, Colin

natro

Member

Post Number: 3
Registered: 07-2007

Rating: N/A
Votes: 0


Friday, July 06, 2007 - 10:26 am:

I personally limit myself to 2-3% total loss risk on my account (ie $2-3 per every $100 in account)...Anything less than that would be too limiting on potential gains due to an overly restrictive risk/reward ratio per trade....I'd get stopped out way too much!

david_louisson

Member

Post Number: 276
Registered: 02-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 08:48 am:

This post may seem off-topic to begin with, but please bear with me.

Let's assume, by way of analogy, we are playing a coin tossing game where (for example) heads represents a win of double our stake, and tails loss of our stake (hence we have significantly positive expectancy). We begin with a fixed bankroll, and the coin is to be tossed (for example) 10 times.

Here is the question: assuming consistent bet sizing (e.g. risk 2% of our bankroll per toss), does the risk to our total capital change whether the coin gets tossed once per hour, or once per minute? Hopefully it's obvious that the answer is no, because the same probability distribution is involved; it's just that in the latter case the same (average) outcome will be reached 60 times more quickly. What happens if the coin gets tossed once per second? Same answer. Extrapolating further, what happens if we toss 10 coins simultaneously? Same answer again. We are leveraging time to accelerate the outcome (i.e. the win rate, given that the expectancy is positive), while leaving risk completely unaffected.

That means that we are just as safe risking a total 20% of our capital spread across 10 simultaneous tosses as we would be risking 2% on any one toss. But we will profit 10 times as quickly. (Is it any wonder why casinos cram as many gaming tables under their roof as they possibly can?)

Hopefully it's obvious how this could be applied to trading, albeit with the one hugely significant proviso that Colin has clearly mentioned. The coin tosses are completely independent events, while markets are correlated.

If there is 100% correlation across 10 simultaneous events, then the total risk to the account (at 2% per event) is 20%. If there is 0% correlation, then the total exposure is 2%. If, as Colin says, market risk due to correlation is 50–60%, then trading 10 stocks simultaneously (in different sectors) at 2% risk per stock, would amount to an effective total exposure of maybe 10–12% (as opposed to 20%).

Hence the goal is to find markets that are as uncorrelated as possible. By doing so, exposure could be increased beyond what is "normally safe", accelerating return while mitigating risk.

The purpose of the post is to explain the mathematical relationship between correlation and exposure. It is better explained in Chapter 8 of Ryan Jones' book "The Trading Game: Playing by the numbers to make millions" (Wiley, 1999)

David

(Message edited by david_louisson on July 13, 2007)

hilarius

Member

Post Number: 2501
Registered: 04-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 09:13 am:

Hello David, Welcome Back !

Are you simply saying that a bigger risk (amount staked) will provide a bigger gain or loss than a smaller risk?

Regarding the coin toss the risk is success or failure in the blink of an eye

In the markets the risk is more a question of failing to recognise failure quickly and failing to hang on to success

These opportunities do not exist with a coin toss

In addition there is a mass of fundamental data to support choices with better than average chances of success

I know your views about the randomness of price action ... but may I put it to you that the positive correlation between prices and earnings is far from a random event?

With Best Wishes

Hilarius

I come in peace to share my thoughts and to shine my candle light on possible long term opportunities

colin_twiggs

Member

Post Number: 2974
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 10:47 am:

David,
I agree that the more trades you place, the more you will make. The constraint is your capital.

The major risks of the 2% (or 1%) rule are:
(a) stocks do not move independently of each other;
(b) markets are not always independent; and
(c) stop loss orders may fail to protect you.

To minimize this we should:
(a) attempt to trade stocks, commodities and futures that are as independent as possible (e.g. RIO & WOW rather than RIO & BHP)
(b) look for markets that are as independent as possible (this may be hard to find)
(c) trade both short and long when the opportunity exists (i.e. market phase 1 or 3 - bases or tops)
(d) use guaranteed stops or derivatives to protect large positions (the 2% rule does not limit the amount that you invest in a stock - only your "capital at risk")

Regards,
Colin

resillent1

Member

Post Number: 92
Registered: 10-2006

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 11:09 am:

Great post David.

The expectancy from a trading system will increase if you can do better than 50% win ratio and have bigger winners than losers. (size of win/loss ratio is more important than win ratio)

The mathematics of probability still remains the same but the variables change.

Once you have a system with positive expectancy the idea is to utilise it as much as possible, but in doing so you want to avoid positive correlations in your positions, so that one event (market downturn) doesn’t cause multiple losing positions.

Market risk becomes a limit to utilisation of a positive expectancy system unless you can find truly uncorrelated positions.

david_louisson

Member

Post Number: 277
Registered: 02-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 11:31 am:

Hilarius, many thanks for the welcome. Been away for so long that I'd forgotten what a friendly community the IC forum is. :-)
I trust all is well with you.

I'm in no way trying to equate trading with coin tossing, nor am I suggesting that the same element of randomness in a coin toss exists with price movements. I am not considering price movements, nor their relation to entries or exits. The only thing I'm considering is position sizing (percentage of capital at risk).

My point is to simply illustrate the mathematical reasoning behind why:

10 simultaneous positions @ 2% risk per position with 100% correlation = 20% effective total exposure; but

10 simultaneous positions @ 2% risk per position with 0% correlation = only 2% effective total exposure.

10 uncorrelated simultaneous positions will therefore generate 10 x the theoretical return of a single position, at no added risk. While the correlation inherent in the markets undermines this significantly, the principle is still valid, albeit watered-down by the extent of the correlation.

Colin, I agree totally with you, especially re independence. The goal is to try to find markets that are as uncorrelated as possible. I guess I've been frequenting the forex forums too long now, where 100:1 leverage, and automatically guaranteed stop losses, get ingrained into one's thinking. Without leverage, (lack of) capital is certainly an insurmountable constraint.

David

david_louisson

Member

Post Number: 278
Registered: 02-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 11:53 am:

Thanks, Resillient1

I've been demo trading forex instead of CFDs for almost a year now, but hopefully some of the basic principles hold good.

Expectancy or "profit factor"
= dollars returned for every dollar risked
= (# wins / #losses) x (avg net win size / avg net loss size)

Variables can be adjusted to suit, e.g. tighter protective stoploss will reduce avg loss size, but the more frequent stopouts lowers the win rate.

Position size operates independently of expectancy, but magnifies or diminishes both return and risk (win and loss sizes) in like proportion.

Agreed, correlation is an enemy. The irony is that forex provides the leverage needed to trade multiple simultaneous positions, but because there are only a handful of tradable currency pairs, and all the majors involve USD, correlation makes the strategy unworkable. With stocks and derivatives, there are different markets (e.g. Nasdaq, FTSE, ASX, etc), different industrial sectors, and thousands of equities to choose from, but minimal leverage.

David

colin_twiggs

Member

Post Number: 2975
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 12:02 pm:

David,

Referring to it as a capital constraint may be misleading. The limit is the number of positions that one can enter simultaneously while applying the 2% rule. This would be a lot less than 50 without leverage.

Regards,
Colin

resillent1

Member

Post Number: 93
Registered: 10-2006

Rating: N/A
Votes: 0