RISK

[ducati916]

Don Corleone

I came to Aussie Stock Forums because I believed in the opportunity and strong legal system. I pay my taxes, contribute, and am a law abiding citizen, regrettably, my posts have been tampered with, deleted and I suffered personal abuse.

Don Corleone, I come to your home, on your daughters wedding day, and I ask for justice.

bullmarket must be hung, drawn and quartered, his entrails taken to the swamp and fed to the crows in his own soupbowl...........
His very name must be stricken from the record, no record of his existence can remain for posterity.

But seriously, I doubt MrBull is unduely concerned, after all he seems to revel within the environment, giving as well as receiving, even after promising me personally, that he was in point of fact not appearing until the end of the week, here he was .......back again, catching me redhanded in the jar, with my fingers wrapped around his largest and finest chocolate chip cookie.
If it had all been one way traffic, then fair enough, I don't approve of bull ying posts either.

The actual topic of risk, is an important one, and real-life examples always make better reading than purely theoretical, as whoever designs the theoretical is open to bias in building the model........this way strengths and weaknesses are highlighted..............its nothing personal, just business!

jog on
d998

 

[tech/a]

So then to quantify RISK is in every single case subjective so to over come the problem allocate RISK.

This then places a bottom on the RISK that we take on any one trade.
We have control of that risk.
Its called an INITIAL STOP

In itself this is not sufficient enough to Manage Risk

What we dont have control of is the number of times our stop will be taken out in any period.---So to equate the effectiveness of our setting of a stop and its effectiveness as a money management tool we need 3 more pieces of information.

(1) Over X period (The longer the better) what was the greatest string of consecutive losses?
(2) Average consecutive losing trade.
(3) Whats is our average win to our average loss.

Without these three important pieces of information setting of a stop could be simply like filling a bucket with a hole in it.

Take it away Duc--i --san

 

[ducati916]

My portfolio, which appears under Fundamental II is assessed as follows;

Analytical model utilized
Fundamental analysis (Value & Cash-Flow)

Quantification of risk model
Deterministic. (My testing extends from 1901 to present day, and includes 3,432 securities analyzed) *note...these are not 3,432 INDIVIDUAL securities, they are often the same security, at different points in its listed history, there have been many that have been delisted, merged, etc. This survivorship bias is overcome, as they have been done by hand from individual records, and are not extrapolated forward in time.
When I add a further 6000 examples, then it will just qualify as a STATISTICAL model

Risk management model
Diversification (micro, macro, geographic, strategies, asset class allocation)
*proprietary*
VAR

Components of risk assumed
Slippage
Market risk

Systemic risk assumed
Exchange rate
Political

Systemic risk allowed for
Inflation
Interest rates
Information flow (in regards to earnings surprises that are negative)

Non-systemic risk allowed for
Psychological risk

jog on
d998

 

[tech/a]

Now I know why I allocate RISK.

Deterministic.
What has it found. What figures have you been able to extrapolate?
How are they used?
What then do you do with this information and how does it effect your trading?

 

[tech/a]

Duc.
Im also interested in what it is you actually derive that then gives you a measure of RISK.What are you determining?

How many individual securities make up the list which you have studied?

Where do you get Fundamental Data back over 100 yrs?
Where do you store it?

 

[ducati916]

tech/a

Determinism, is a philosophical argument, that has been applied to the market
Very briefly, past events and or conditions are predictive of future events.

Therefore you would test your set of conditions, historically, and going into the future.

If you had a large enough sample, long enough timeframe, robust statistical design, your deterministic sample may show statistical significance.
If too small, your results, assuming a positive outcome, are deterministic, rather than statistically significant.

What has it found. What figures have you been able to extrapolate?

It has found, on aggregate, a 30% compounding, unleveraged return.

How are they used?
What then do you do with this information and how does it effect your trading?

They are used in a similar way to TT, in that you apply risk management to your *numbers* and trade them.

Im also interested in what it is you actually derive that then gives you a measure of RISK.What are you determining?

The number of winning trades (Win%)
Average % return per trade

How many individual securities make up the list which you have studied?

About 900 odd now.

Where do you get Fundamental Data back over 100 yrs?
Where do you store it?

I am a collector (investor) of First Editions, particularly finance first editions, and I have publications dating back to 1888.
A lot of the older data was collected via the Cowles Commission.

On my bookshelf baby!

jog on
d998

 

[It's Snake Pliskin]

Duc,

..no problem We`ll have to agree to agree and I only call it as I see it.

Consult your bookshelf to let us know about hedging a portfolio. This is important.

Snake

 

[tech/a]

Hmm

Then Duc I see we have nothing to discuss.
Your response is evasive of direct questions.
I dont believe you have mis interpreted the questions--you are far too astute for that.

There are no specifics.
At least I can be definitive in determination of risk almost instantly.(Once results have been attained)

I'll continue later.

 

[ducati916]

tech/a

Im also interested in what it is you actually derive that then gives you a measure of RISK.What are you determining?

My *risk* is (amongst others) business, or credit risk

Market risk, or volatility is irrelevant as a metric of risk to my methodology.
Therefore there is no requirement for a stoploss as a risk management tool.
A change in the fundamentals is a trigger for an exit, and a time based exit, at 3yrs. Either one will trigger an exit. With regards to a %measure of my risk, there isn't one in the same context that you advocate, my exit may be triggered at a 10% loss, or a 90% loss, which requires a very high success rate to mitigate risk, and fairly broad diversification.

My backtested results suggest a 95% success rate or winning percentage.
The average return per winning trade is 87.5%
The average loss is 25%
Once you play with the figures over a variety of time based exits, and any losses, with the winning trades, you would return 30% compounded.

A stoploss (excluding the CFD guaranteed stoploss) is not definitive of your risk.........a successful exit at the stop price is definitive of your risk

The two methodologies are just night and day.
This is why when people talk about mixing and matching Fundamentals with Technicals to create some form of superior hybrid, they are just asking for problems as the definition of risk, the quantification of risk, and the management of risk are just so different as to be for all intents and purposes incompatible.

jog on
d998

 

[ducati916]

Snake

Consult your bookshelf to let us know about hedging a portfolio. This is important.

Ok, nice basic one.
This will not go into detail, but will outline the methodology.

Find an overvalued common stock.
Find one that also has a convertible bond or preferred stock as part of the capitalization.

Buy the bond long.
Sell the common short at parity, or +ve spread.
Convertible needs to be relatively close to investment value.

Wait for the common to fall.
When it falls, it will usually fall further and faster than the convertible.
If this happens, buy back the common, and sell the bond.

If nothing happens, convert the convertible into common to close the trade at breakeven, less brokerage, + any dividends or interest you accrue.

There are others, but see how that one sits with you first.
jog on
d998

 

[tech/a]

My *risk* is (amongst others) business, or credit risk

This could alter 6 mthly if reporting is 6 mthly.What your doing is then evaluating the risk to the company and once evaluated,determining its value and I presume potential for a trade.There is a number of ratio's you could use.

Fine nothing wrong with that but I would put forward to you that that would be an excellent fundamental analysis culling of a potential universe of stocks to trade.

By trading then in a systematic/mechanical way by allocating the risk within your model of trading, I would contend that your results would improve.Infact without testing Currently this idea would have your portfolio looking a little healthier and you may well be in other more profitable trades from within your universe. Anyway perhaps you would like to combine ideas on your universe (I have the US data) and trade a parallel portfolio based upon Mechanical methodology---would have to be designed and tested.I can do that if I know your universe.

Market risk, or volatility is irrelevant as a metric of risk to my methodology.
Therefore there is no requirement for a stoploss as a risk management tool.
A change in the fundamentals is a trigger for an exit, and a time based exit, at 3yrs. Either one will trigger an exit. With regards to a %measure of my risk, there isn't one in the same context that you advocate, my exit may be triggered at a 10% loss, or a 90% loss, which requires a very high success rate to mitigate risk, and fairly broad diversification.

Understand,but leaves you open to a possible 100% loss in multiple stocks.

My backtested results suggest a 95% success rate or winning percentage.

Over what time---?? Must be over a year as reporting will only be available in detail once a year.The highest success rate I have seen with any method.Short term methods normally report high win rates and the highest I have seen variified is 72% Longterm is seen as 35-45% and this is a figure derived by systems analysts both fundamental and technical.Your current rate in your example reflects this--early days.

The average return per winning trade is 87.5%

Again the highest I have seen,truely amazing.

The average loss is 25%

With a 5% loss rate this is insignificant.

Once you play with the figures over a variety of time based exits, and any losses, with the winning trades, you would return 30% compounded.

I would like to know your average trade length winners and losers.

A stoploss (excluding the CFD guaranteed stoploss) is not definitive of your risk.........a successful exit at the stop price is definitive of your risk

Its definative of INITIAL Risk.It will limit the maximum loss on initial capital to the value assigned.Loss from highest high to exit is drawdown from maximum profit within anyone trade,thats nothing to do with initial stop losses.

The two methodologies are just night and day.

Yes but each have benifits that maybe helpful.

This is why when people talk about mixing and matching Fundamentals with Technicals to create some form of superior hybrid, they are just asking for problems as the definition of risk, the quantification of risk, and the management of risk are just so different as to be for all intents and purposes incompatible.

I dont think so,Its been done before and will be done again.To a degree I have done something like this (Ok loosley) with T/T.
The Fundamental analysis has been done by the Margin lender on the stocks he allows in his Margin list.That margin list is my universe of stocks that I use for trading.

 

[tech/a]

Its definative of INITIAL Risk.It will limit the maximum loss on initial capital to the value assigned.Loss from highest high to exit is drawdown from maximum profit within anyone trade,thats nothing to do with initial stop losses.

Sorry miss read your statement.
Yes true but this is true of ANY stop or exit,if its not taken then the result will be different. This doesnt detract from the reasons for defining your risk in the first place and the benefits of doing so.

 

[It's Snake Pliskin]

Snake

Ok, nice basic one.
This will not go into detail, but will outline the methodology.

Find an overvalued common stock.
Find one that also has a convertible bond or preferred stock as part of the capitalization.

Buy the bond long.
Sell the common short at parity, or +ve spread.
Convertible needs to be relatively close to investment value.

Wait for the common to fall.
When it falls, it will usually fall further and faster than the convertible.
If this happens, buy back the common, and sell the bond.

If nothing happens, convert the convertible into common to close the trade at breakeven, less brokerage, + any dividends or interest you accrue.

There are others, but see how that one sits with you first.
jog on
d998

Duc,

Thanks for your quick reply and exemplary forum style.
I see where hedging a portfolio is quite hard work. So for say 8 stocks being hedged using the same method would acrue a lot of brokerage and be time consuming.

Is it worth it?

Feeel free to join in all.

 

[tech/a]

my exit may be triggered at a 10% loss, or a 90% loss, which requires a very high success rate to mitigate risk, and fairly broad diversification

Infact it also requires a very high reward to risk .

If a $10 stock loses $9 then the remaining stock valued at $1 needs to increase by 1000% to just return to break even. People lose sight of this when they hold a stock which continues to decline. Infact it becomes rediculous when you consider that your expecting a stock which is underperforming (To get itself in negative) to then outperform by massive amounts.

Quantification of risk model
Deterministic.

This is no different to ranking,you have not determined risk rather you have ranked it.Risk being uncertainty--then the uncertainty remains.
Allocating a risk then becomes a "Next Level" and its possible use has been suggested above.

 

[ducati916]

Snake

As to being worth it, that is really an individual choice that should be governed by your risk exposure calculation.

Total Risk = Controlled risk - Uncontrolled risk
With the reward component following a similar balancing calculation.
Viz. a fully controlled reward will fall at the lower band, with highly speculative reward falling in the upper bands.

An alternative hedging strategy is the Pairs trade
Here we take two common stocks exposed to the same business (an example) could be Coca-Cola & Pepsi Cola.

If (and lets assume this is the case) one is over valued, and one under valued, then we sell the over valued short, and buy the under valued long.
You are (sort of) market neutral.

Another strategy is the Fair value hedge
Here we use the futures market trading price against the cash market to establish *fair value* (there will be either a premium, or discount)
Lets assume a discount, on the opening we assume a return to fair value, so sell the cash market, buy the futures...........you can work the same trade using common stocks that you follow, assuming you believe that you know how they trade.

jog on
d998

 

[ducati916]

tech/a

This is no different to ranking,you have not determined risk rather you have ranked it.Risk being uncertainty--then the uncertainty remains.

In a way you could call it that. You are ranking low risk, against high risk, and loading your portfolio with low risk selections.
Through analysis you have determined that your stock selection is undervalued, and represents a bargain, thus it ranks as low risk.

Through *backtesting* the undervalued stock selections (or your ranking methodology) as a strategy have been tested to see how successful it has been. This generates the first number.........%winning trades and %losing trades.

From there, we look at the aggregate %return of the winning trades, and the aggregate %loss of the losing trades.

Lastly I look at the timeframe that these results have been generated in.
If the annual compounding rate is high enough to compensate for the total risk, where total risk = controlled risk - uncontrolled risk then I would assume the strategy.

Infact it also requires a very high reward to risk .

If a $10 stock loses $9 then the remaining stock valued at $1 needs to increase by 1000% to just return to break even. People lose sight of this when they hold a stock which continues to decline. Infact it becomes rediculous when you consider that your expecting a stock which is underperforming (To get itself in negative) to then outperform by massive amounts.

It could do, but it doesn't have to.
This is where diversification as risk management is so important to my overall strategy.

Lets take my starting capital at $1M
I allocate say 25% to a common stock strategy, 25% to Bonds, 25% to arbitrage, 25% to real estate.

$250K to common stocks at $12.5K (or 5% per stock selection) in a portfolio of 20 stocks. Absolute worst case scenario, you lose 100%, or 25% of your capital. But equally you could have 250 securities, each with $1K in them.
The law of large numbers starts to work in your favour, the wider the diversification, but at the expense of return, and is the basis of the insurance underwriting industry.

However, with a *determined* 95% success rate @ 85%+ return against a 5% loss rate @ 25% you are really asking how thoroughly have you researched, and how strong is the correlation. Does the correlation make sense from your understanding of finance theory (or whatever field you study)

Assuming reality follows history, then you will realize substantial profits.
That is the risk you assume, and attempt to control via your understanding of, and management of risk.

Business, or operating risk, is very different to market risk.
Technical analysis, which trades market risk, is an extremely lagging form of analysis and therefore must follow the herd, and assume the herd knows what it is doing, this is the underlying theory that underpins Efficient Market theory thus the risk management tool utilized must fully control risk, viz. if risk = uncertainity, then uncertainity must be converted to certainity. This is the purpose of a stoploss,......a stoploss = 100% loss, therefore, there is no uncertainity, there is no risk.

jog on
d998

 

[tech/a]

Total Risk = Controlled risk - Uncontrolled risk
With the reward component following a similar balancing calculation.
Viz. a fully controlled reward will fall at the lower band, with highly speculative reward falling in the upper bands.

If I set an allocation of a 5% ( of purchase price) As an initial stop where is and how would you quantify uncontrolled risk?

In a way you could call it that. You are ranking low risk, against high risk, and loading your portfolio with low risk selections.
Through analysis you have determined that your stock selection is undervalued, and represents a bargain, thus it ranks as low risk.

Through *backtesting* the undervalued stock selections (or your ranking methodology) as a strategy have been tested to see how successful it has been. This generates the first number.........%winning trades and %losing trades.

From there, we look at the aggregate %return of the winning trades, and the aggregate %loss of the losing trades.

Lastly I look at the timeframe that these results have been generated in.
If the annual compounding rate is high enough to compensate for the total risk, where total risk = controlled risk - uncontrolled risk then I would assume the strategy.

Not withstanding the above you still have uncertainty in any one event you are exposed to unlimited risk,with an average of 25% loss a string of even 2 losses at the high end of the "Averaging" could render your initial capital in effective if you were to liquidate the loss and highly unlikely to recover as my example above shows if you hold it.
Purchasing 10 stock with this sort of uncertainty could be catastrophic.

Take your example that your trading now. Using $50K as a trade size each trade.
To sell all stocks now that are over 10% in loss would mean massive destruction to your capital base.So your really stuck with them and hope they make up the low end of "Averages",unless ofcourse you sell those in profit as well to offset losses. So your holdng both winners and losers.

In contrast T/T has 10 stocks and everyone of them is a winner.
Simply I cut losers and hold winners.

Diversification is wise but still desnt negate risk mitigation in any one investment.To argue that sound diversification balances poor risk management in anyone asset is avoiding the problem not solving it.

Business, or operating risk, is very different to market risk.

True so I simply eliminate the effect of market risk on my business and deal purely with my business risk.I dont have to quantify market risk,my trading model doesnt care less about actual or percieved market risk.

Technical analysis, which trades market risk, is an extremely lagging form of analysis and therefore must follow the herd, and assume the herd knows what it is doing, this is the underlying theory that underpins Efficient Market theory thus the risk management tool utilized must fully control risk, viz. if risk = uncertainity, then uncertainity must be converted to certainity.

The herd can do whatever it likes.If it follows my trades positively then I'll stick around---negatively then I'll go away. There are no assumptions,I'm either profiting and maximising return by staying with it or its losing in which case I wont stick around very long.

 

[lesm]

Business, or operating risk, is very different to market risk.
Technical analysis, which trades market risk, is an extremely lagging form of analysis and therefore must follow the herd, and assume the herd knows what it is doing, this is the underlying theory that underpins Efficient Market theory thus the risk management tool utilized must fully control risk, viz. if risk = uncertainity, then uncertainity must be converted to certainity. This is the purpose of a stoploss,......a stoploss = 100% loss, therefore, there is no uncertainity, there is no risk.

Duc is spot on here.

Regardless of whether you use a 10% initial stop loss and/or a 2% trailing stop loss, it is a certainty that you are going to lose that percentage. Especially, if you are using an approach where you are waiting for the market to prove you wrong (i.e. your stop loss is hit before you exit the market/trade). This is what you see in a lot of trading systems.

The only variance is the actual realised $ value that the % stop loss equates to.

If you exit the market when a target profit (whether it is $ or % terms) is hit then the loss will not be realised, as the stop loss has not been triggered, hence no loss incurred.

cheers.

 

[tech/a]

So then you would prefer not a known 2% capital loss (Allocated) but a possible 100% capital loss (Unallocated) in anyone trade.

Why?

Duc hasnt answered string of losses and average number of consecutive losses V average consecutive winners.
I can bring these important aspects into the discussion when we have 2 benchmarks--duc's and mine.

Your post aludes to this.

On lagging.
The point of a buy trigger is anything but lagging.
The make up of the analysis maybe lagging but the trigger itself is a point in time---infact its a NOW point.


On Business risk.

Yes I agree it is very different to market risk.
However it is anything but certain.
Its analysis is subjective as I doubt duc has 900 benchmark studies for all aspects of business risk studied.Operating risk can alter very quickly so again subjective.
it can also be argued that fundamental analysis in itself is also lagging badly as information cannot be found for 12 mths at time of reporting and benchmarked back over say 3 yrs.but like tech analysis at the time of completion of analysis it its on the day of reporting then there is no seen lag.

Tech analysis is no different---past to generate analysis---now for a result to supply a buy trigger.

 

[Knobby22]

I disagree a little with a few statements.

There is a famous proof that some economists got a noble prize for that shows diversification reduces all risks except market risks i.e. market goes up and down. This can be protected from using options.

Secondly, though risk can't be fully quantified and analysis can be somewhat subjective, it can be estimated.
Stock picking relies on estimating competitor and environmental risks and their effect on future profits and estimating whether the market has under or over compensated for these risks. A good stockpicker who uses technical analysis as a check can do very well.

There are many ways to obviate risks and many of those ways result in giving some of the gains up. e.g. take diversification, if you follow this route to the nth degree then you will only get market returns.
Stop losses cost in transaction fees.
Options cost and can limit returns.