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

Incredible Charts
Free Charting Software

Auto-Fit Trendlines

Trend Channels

Linear Regression
Channels

Raff Regression
Channels

Standard Deviation
Channels

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


Friday, July 13, 2007 - 12:23 pm:

Colin

I’ve just been reading through this thread. In one of your posts you write “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%”

Is this from a book? Does he provide much background research or data for the figures?

I would be interested in reading further in this area. Do you know of any other published material?

In the absence of adequate understanding I currently assume perfect correlation (extra safety). This however often leaves me with un-utilised capital simply to moderate market risk.

My search for uncorrelated positions hasn’t been very fruitful yet. Knowledge required and performance offsetting have been stumbling blocks.

Regards

Resillent

hilarius

Member

Post Number: 2505
Registered: 04-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 12:53 pm:

David

I am a bit confused

Doesn't independence (= lack of correlation ) provide risk minimisation rather than profit maximisation ?

Am I missing something ?

Hilarius

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

david_louisson

Member

Post Number: 279
Registered: 02-2004

Rating: N/A
Votes: 0


Friday, July 13, 2007 - 10:32 pm:

Hilarius

Referring to my previous post:

a) 10 simultaneous positions @ 2% risk per position with 100% correlation = 20% effective total exposure; but
b) 10 simultaneous positions @ 2% risk per position with 0% correlation = only 2% effective total exposure.
c) 10 uncorrelated simultaneous positions will therefore generate 10 x the theoretical return of a single position, at no added risk.

The maximization of profit stems from the fact that we can now open 10 positions, instead of one, hence the profit is increased 10 times. If, the correlation = 0% (and I stress this is theoretical, as there is always some correlation in the markets), then the risk of having 10 simultaneous positions open is no greater than having just one open (refer to point (b) above).

I agree this is difficult to get one's head around. I had to read the relevant chapter in Ryan Jones' book at least twice before the concept finally made sense to me.

David

msparks

Member

Post Number: 1022
Registered: 10-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 12:35 am:

Hi David
Welcome back, i always enjoy reading your theories.

Why can't you design a system whereby stocks with high volatility and large ranges are purchased with a guaranteed stop loss of 5% and take equal long and short positions.

Market risk is eliminated
Loss is guaranteed and LIMITED
Gain is UNLIMITED

Is that not a positive expectancy and guaranteed way to make money ?

hilarius

Member

Post Number: 2515
Registered: 04-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 03:57 am:

Good Morning David

I can well understand how being in uncorrelated markets would minimise risk by offsetting losses with gains, so minimising risk

It thus seems to me that if there is perfect uncorrelation the opposite must also apply ... i.e. gains will be offset by losses, so minimising gains

Clearly my simple brain needs to be expanded

So far I can only see that uncorrelation would ensure that gains and losses would cancel each other out

If there is perfect uncorrelation would it not be simpler to do nothing and let the cash earn interest?

Hilarius

PS My definition of perfect uncorrelation is an equal number of markets rising as falling, by equal average percentages for the rises and falls

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

hilarius

Member

Post Number: 2516
Registered: 04-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 04:14 am:

I also understand Friar Mark's point that if one has a mechanism for letting losses go quickly and retaining profits for longer for greater percentages that could be the answer to my question

The brain is rather slow on first waking, even though you may have eaten your breakfast by now

Over here the birds are not yet sungung (I'm surrounded by Kiwis on the Gold Coast so I'm learnung to talk like thum)

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: 2976
Registered: 09-2002

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 12:33 pm:

Mark,

Taking equal long and short positions may be achievable in stage 1 or 3 (market bases and tops), but difficult in stage 2 (bull market) or 4 (bear market).

Regards,
Colin

ohkoolnutz

Member

Post Number: 618
Registered: 10-2005

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 02:50 pm:

"Why can't you design a system whereby stocks with high volatility and large ranges are purchased with a guaranteed stop loss of 5% and take equal long and short positions?"

The volatility would trip both sides' stop losses and you would end up with no favorable result.

---
ohk

Lies, Damn Lies and Technical Analysis

david_louisson

Member

Post Number: 280
Registered: 02-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 03:19 pm:

Hilarius

By "uncorrelation" I mean "no correlation" or independence, i.e. where the outcome of one event has no bearing on the outcome of another event.

As we know, the markets are correlated to a certain extent, i.e. if the Nasdaq has an up-day, then the FTSE, Nikkei, ASX, etc likewise tend to rise, more often than not (perhaps the good Cap'n would confirm – is he still around here? – that "all boats rise with the tide" :-). So the opportunities to find "uncorrelated" trades will never be perfect, but one can try one's best by picking two equities in different countries, different markets, different industrial sectors, etc.

What you are describing, I would call "negative correlation" or "hedging". The "perfect" example would be taking both a long and a short position in the same equity. As you point out, the gain in one position would exactly offset the loss in another, hence it would be pointless.

Hope this explains.

David

david_louisson

Member

Post Number: 281
Registered: 02-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 03:23 pm:

Msparks

Thanks for the welcome.

If I understand what you're saying:

Picking stocks, in order to take long positions in some, and shot in others, may alleviate, and even eliminate market risk, but I expect we would need to go long in the strongest stocks, and short the weakest (as opposed to a purely random selection) for the system to have best chance of profit. Then if the market rises, the strong equities should also rise at a rate greater than the weaker ones, that are swimming against the rising tide; hence the profit on the long positions should exceed any losses on the shorts. And vice versa if the market falls. If (and this is the big if) this could be achieved to the point that costs are overcome, consistently enough, then we have a profitable system.

We would need to trade CFDs or options to provide the leverage needed to hold the multiple simultaneous positions. In the case of CFDs (I don't understand options) we would also need to factor in the effect of interest debited on long positions, and credited on shorts, and any synthetic "dividends" that are also debited/credited.

From memory, Ed Thorp (the guy whose book "Beat the Dealer" was a major catalyst for all of the Blackjack card counting systems that emerged during the 1960s – not to be confused with Van Tharp :-) operates a system similar to this, with some success. See http://webhome.idirect.com/~blakjack/edthorp.htm

David

captain_chaza

Member

Post Number: 2549
Registered: 02-2003

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 06:47 pm:

Ahoy Officer David

Yes I am passing through
For how long? The Gods' only know!

The secret to Sailing the Global Exchanges and in my case the ASX because of my addiction for sleep is in SHIP DESIGN

Please note
No TWO sails are ever cut the same
Ask any ship's Bosun

We are not dealing with pennies here Dave where it is assumed both sides of the penny are equally weighted as in TWO UP

We are not even dealing with cricket balls that are polished on one side and scratched on the other to produce swing

It would be a very rare event if it is at all possible to find 5 sails of equal weighting to prove any of your strategies

To me it all a lot of Landlubber Gibberish

To think that a Risk strategy of 2% is valid for a ship worth $25,000 is the same as one worth $200,000 is rubbish

The secret is in SHIP DESIGN from the outset

Re your comment "All boats rise in the tide"
Well, I used to believe that until 20 years ago
when I discovered it was just an old broker's/landlubber dribble

Yes some of those Clumsy Heavy-weighted Mainsails do go that way and some are designed to HOLD UP in the adverse conditions

On Hedging /Betting against winning or losing to me is like betting on the Green and the Red at the casino and trying to make the casino's profit

It makes no sense!

Nowadays I go on the belief that
"The Wind Calls the Tune"

I hoist whatever the charts tell me in the ASX 3 letter code

There is plenty of time to do the FA research after you are in!
If you are that way inclined?

I must admit that Sailing with absolutely NO idea of the Sweet FA's does force you to venture onto some very swift ship designs!!!

Sailing gets really stressful when you discover you are sailing in only ONE or TWO classifications of sail

But what the heck
"Speed is EVERYTHING"

It seems I was indoctrinated to think that
"Diversity is Good"
LIE! LIE! DAMN LIE! #1

Once you accept the fact that
"The Wind Calls the Tune" and not a few tidal waves many thousands of miles away
The Next secret to learn/conquer is the "ATR"

Nothing on this subject makes any sense 'til you do!

Salute and Gods' speed

(Message edited by captain_chaza on July 14, 2007)

(Message edited by captain_chaza on July 14, 2007)

"While we stop and think, we often miss our opportunity." Publilius Syrus, 1st century B.C.

"I believe the future is only the past again, entered through another gate."
Sir Arthur Wing Pinero 1893

"There are two times in a man's life when he should not speculate: When he can't afford it, and when he can." Mark Twain, 1897

captain_chaza

Member

Post Number: 2551
Registered: 02-2003

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 08:32 pm:

Ahoy David

Are these theories you are referring to based on share trading
OR
an academic's view / mathematician's answer to avoid doing the hard yards in learning all about TA or FA

It seems to me that no mathematician has ever succeeded in any great sport ???

Salute and Gods' speed

ingot54

Member

Post Number: 1895
Registered: 05-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 08:49 pm:

Ahoy Captain

For the things of this world cannot be made known without a knowledge of mathematics ~ Roger Bacon

I believe one of the greatest yacht designers ever was Ben Lexcen had not a little mathematical knowledge!

Roger Bacon - a mathematician and sometime philosopher like his namesake Francis Bacon, also had this to say:

Argument is conclusive, but it does not remove doubt, so that the mind may rest in the sure knowledge of the truth, unless it finds it by the method of experiment.

Without the Davids of the world, and the Captains, where would the Ingots learn anything?

Keep Smiling - Don't look back

Trading style: Chartist Artist _ Breakouts and Shakeouts.

msparks

Member

Post Number: 1024
Registered: 10-2004

Rating: N/A
Votes: 0


Saturday, July 14, 2007 - 09:08 pm:

Hi Colin
There are always risers and fallers.
The main point was the guarantee of a loss of 5%.

People talk about tossing coins etc, etc.
There is an opportunity to GUARANTEE the loss is 5%, while having unlimited potential gain LONG or SHORT.

Tossing coins is even money, red , black .
Tossing coins where a loss = 5% and a win = 100% (or whatever) ,seems like a casino's nightmare.

Thanks for the lesson skippa, those atr's are such a good tool but i hate those BB's on my charts.

Could we have a selection button ? On or Off
AND an trailing ATR stoploss line ?

lafee

Member

Post Number: 580
Registered: 04-2003

Rating: N/A
Votes: 0


Sunday, July 15, 2007 - 01:53 pm:

Unfortunately the fat tail enters the picture in both the magnitude of a loss AND the potential number of losses in a row.

Cheers Lafee

Don't ask an academic if what he does is relevant

msparks

Member

Post Number: 1025
Registered: 10-2004

Rating: N/A
Votes: 0


Sunday, July 15, 2007 - 05:05 pm:

Lafee

Did you just insult me or was your post a witty cryptic answer of the type that only a professional trader would understand ?

The magnitude of the loss can be managed, the number of loss's in row needs to be managed via entry setup,and i suggest position size reduction after several "in a row loss's".

5% is probably a little too tight also but whatever loss is optimal, can be guaranteed, however i do not use them now, the spreads are too much for short term, swing and longer term positions guarantee is OK.

lafee

Member

Post Number: 581
Registered: 04-2003

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 12:03 am:

Msparks,

I think 'fat tails' were mentions above? I was not trying to insult you.

I am just saying that while a g stop may save you from a (1) rare event (fat tail) your system performance over x period is also subject to'fat tails'. Say your reasearch indicates a particular trade has a 70% chance of success and that you have never had more than 4 losses in a row. Now keep in mind that any odds you calculate are not real - you are 'discoving' the odds as time goes forward. It is much more likely than the stat professor will tell you that your system will one day face 8 or much more losses in a row.

It is a rather disturbing fact that we face - all systems will eventually break down. And you cannot predict when it will happen. I have experienced it many times. It was never terminal because I trade a number of different systems at any given time. I have diversification in the systems :-)

I think I wrote a while back about the ship analogy. A ship is made up of many sealed compartments so if one area of the ship springs a leak and starts to fill with water the whole ship will not sink. If you have say 10 systems and allocate 10% of your trading capital to each of them you will have much the same protection.

You can then go about 'trend following' your systems. Scrapping the ones that show unusually bad performance and holding on to the ones that do well.

Cheers

Lafee

Don't ask an academic if what he does is relevant

david_louisson

Member

Post Number: 282
Registered: 02-2004

Rating:
Votes: 1


Monday, July 16, 2007 - 04:55 am:

Despite being an abject landlubber, I totally agree with the Cap’n: successful entries and exits are all about application of sound TA/FA, good ship (system) design, plus perhaps a suitably seaworthy mind set, and some basic risk management.

The wind indeed calls the tune – it blows the markets along however it pleases, and while we can’t predict the strength and frequency of the gusts, we can hoist our best sails, to allow us to be caught up in, and blown along by, the prevailing trends. Right, Cap’n?

One of the reasons I enjoyed reading – and learned from – Jack Schwager’s Market Wizard books, was that the world’s greatest traders of the 1970s/80s were an extremely diverse bunch, each of whom brought his/her own unique personality, skills and approach to the table, and somehow made it work, spectacularly, for him/her. There were both mathematicians and non-mathematicians among them.

I wasn’t trying to compare the nature of price movement with casino games. I was using coin tossing as an analogy to propose an (apparently little known?) mathematical principle, relating to position sizing, that I learned from Ryan Jones’ book. Exactly how well it can be applied to the markets is another matter, open to debate and experimentation.

Lafee’s point about all systems breaking down interests me. As a software developer, I came to the market with the expectation (i.e. hope) that I could use statistics to predict, with reasonable probability, future price direction. I have come to believe that this is not so, and that discretionary systems, based upon simple, robust principles (buy into strength as it pulls back, sell weakness when it rallies; enter high volume breakouts from congestion; let profits run, cut losses short; etc) have a much better chance of long-term success than those whose MO revolves around optimized parameters. Optimization ultimately amounts to trying to fine tune the timing of entries and exits, and given that markets consist of a diverse bunch of participants, all with differing systems and goals, and entering and exiting at times that can not be accurately predicted, variations in timing – given a large enough sample of trades – will eventually sum to something approaching zero. Hence finesse serves little purpose, and I could have saved myself years of back-testing different systems had I (heeded Bundy, et al, and) grasped that earlier.

It’s true that, while we can use stoplosses and position size to limit risk on each individual trade, we have no control over the number of consecutive losses we might encounter. The widely implied idea that risk is completely manageable is therefore a myth. I’ve moved on to trading forex these days, and in an attempt to maintain the maximum possible emotional detachment, while recognizing that my system may ‘break down’, I have neutral expectations regarding bottom line, and play (and I use that word deliberately) only with money that I’ve already mentally written off as R&D/education capital. The best I can do is focus on applying sound analysis and trade management, to try to swing whatever probabilities might exist as much in my favor as my knowledge and experience allow; but bottom line ultimately falls outside my control. Because no system can hope to handle all possible market phases and conditions, drawdowns of any duration are potentially possible. To whatever extent losses tend to cluster, Msparks' suggestion of reducing position size during a losing sequence is likely very sound.

David

[Edit: apologies, Colin, we've drifted somewhat from your original topic]

(Message edited by david_louisson on July 16, 2007)

hilarius

Member

Post Number: 2532
Registered: 04-2004

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 07:41 am:

Lafee is correct

Why do planes have 4 engines ?

It quadruples the chance of engine failure

It also increases the chances that the other 3 engines will get you home

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: 2977
Registered: 09-2002

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 10:06 am:

Hilarius,

The comfort of flying in a twin-engined plane is that if one engine fails, the other will carry you to the scene of the crash. Don't know if the same principle extends to four engines.

(Message edited by colin_twiggs on July 16, 2007)

colin_twiggs

Member

Post Number: 2978
Registered: 09-2002

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 10:30 am:

Mark,
You can de-select indicators from the Legend - just clear the check box. I will ask our development team if this can be extended to apply to the entire project, rather than just a single stock.

Regards,
Colin

resillent1

Member

Post Number: 96
Registered: 10-2006

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 10:39 am:

In relation to system diversification, I like single engine jets. However flying one requires an intimate knowledge of how the eject system works and when to use it. There is however always the fear of breaking your neck (slippage) when trying to get out.

lafee,

How many different systems do you use at any one time, do you have multiple positions within each system?

Do you allocate funds evenly to each system or skew towards those performing best?

Regards

Resillent1

lafee

Member

Post Number: 582
Registered: 04-2003

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 12:47 pm:

ATM I trade 12 systems but I have about 150 on file. Each system is usually allocated evenly at the beginning and compounded individually. This not a hard and fast rule as different systems and trading instruments have different risk profiles. For instance, I am far less aggressive on stocks than stock index futures.

Regarding multiple positions within the system. Yes, I often add to systems on dips in shorter timeframes. I also do not have a problem with averaging down - conservatively that is.

Cheers Lafee

Don't ask an academic if what he does is relevant

msparks

Member

Post Number: 1026
Registered: 10-2004

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 01:31 pm:

Thanks Colin

Works for the whole project

, is there any mention of a trailing atr stops or is the atr and standard deviation channels sufficient ?

resillent1

Member

Post Number: 97
Registered: 10-2006

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 02:08 pm:

Lafee,

Thanks for the info.

Just clarifying multiple positions. With say a system for stocks, would you apply it to a few different stocks or do you consider multiple positions as different entries on a single stock?

I find in utilising a single system that when it breaks down, I know what is NOT working in the market but not what IS. I can see your multiple systems would overcome this.

Are your systems largely uncorrelated or do you find they tend to breakdown in clusters?

Do you find the benefits of diversifying systems worth the opportunity costs of focusing on the best system?

Are all your systems for shorter terms? Do you think the feedback time on longer term systems would cause issues?

Cheers

Resillent1

(Message edited by Resillent1 on July 16, 2007)

colin_twiggs

Member

Post Number: 2979
Registered: 09-2002

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 03:17 pm:

Mark,
ATR channels are one of the items on the development list for when we have completed drawing tools.

Regards,
Colin

lafee

Member

Post Number: 583
Registered: 04-2003

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 03:39 pm:

Resillent,

I would rank the system performance in each stock across the entire market to determine which stocks I want to trade and take as many of the top past performers as I can handle. I will, by discretion add to my positions.

Each system is its own entity. So you are monitoring the system performance and deciding upon it's fate on it's (the systems) own merits.

There is always some degree of correlation or clustering of results. The structure I am advocating aims to minimise the negative effects. The different systems are trading different setups so less likely they will all fail together. In addition you are trading a variety of markets. Of course nothing is certain just like there is no guarantee I wont be in a car accident tomorrow.

I have no real idea which of the systems will perform best. I know what has performed best in the past but cycles can change very quickly. I want to make sure I get the big wining systems, trading multiple systems improves my chances.

Most of my trades are under 5 days. I am a strong believer that as time passes predictability decreases exponentially. But you do always have new data each day to act upon.

Does that make thing a bit clearer? or have I made it worse :-)

Cheers Lafee

Don't ask an academic if what he does is relevant

resillent1

Member

Post Number: 98
Registered: 10-2006

Rating: N/A
Votes: 0


Monday, July 16, 2007 - 04:10 pm:

Thanks Lafee.

Things are now both clearer and more confused. You have provoked me to think of lots of questions, but most I need to answer for myself.

Thanks for your responses and if you don’t mind I might quiz you further down the track.

Regards

Resillent.

lafee

Member

Post Number: 585
Registered: 04-2003

Rating: N/A
Votes: 0


Tuesday, July 17, 2007 - 02:19 pm:

David,

Have you tried shortening your lookback period in your backtesting? You may be able to use your stat experience to find tradable 'current' correlations.

Cheers Lafee

Don't ask an academic if what he does is relevant

david_louisson

Member

Post Number: 284
Registered: 02-2004

Rating: N/A
Votes: 0


Wednesday, July 18, 2007 - 06:45 am:

Lafee

Thanks for your suggestion.

I back-tested a lot of systems when I was (rather unsuccessfully!) trading equity-based CFDs, but I'm more into spot forex now, and at this early point my approach is sufficiently discretionary as to make any kind of (automated) back-testing unworkable. Hence I'm really only "forward testing" using both demo, and small live, accounts. Of course all of the major forex pairs involve USD, so I expect that finding opportunities for non-correlated trading is much more difficult.

Best wishes
David

colin_twiggs

Member

Post Number: 2983
Registered: 09-2002

Rating: N/A
Votes: 0


Thursday, July 19, 2007 - 03:05 pm:

Resillent1,

Apologies it has taken me a few days to find my Thomas Dorsey book.

Dorsey equates risk to volatility. He then measures volatility for the S&P 500 index, a sector and a component stock.
~ Divide index volatility by stock volatility to get the % of risk attributable to the market.
~ Divide sector volatility by stock volatility for the % of risk attributable to both market & sector.
~ Subtract market % to leave sector risk.
~ The balance (100% minus market & sector %) is risk attributable to the stock.

Regards,
Colin

(Message edited by colin_twiggs on July 19, 2007)

resillent1

Member

Post Number: 99
Registered: 10-2006

Rating: N/A
Votes: 0


Friday, July 20, 2007 - 02:59 pm:

Thanks Colin

I recently updated my library to include most of the books you list as most influential on the incredible charts trader store. However I skipped over the Thomas Dorsey’s Point & Figure Charting book. Looks like I will have to rectify that.

I have some issues with solely equating risk to volatility but his writings should be an interesting read.

Regards

Resillent.

p.s. Incredible Charts Traders Store (Moneybags) service and delivery was spot on.