Dump it Here

[ducati916]

1. Can the indicator be expected to perform favourably in all conceivable market conditions/contexts (i.e. universally)?

2. If not, which conditions/contexts require alternative tools, or avoidance?

3. How might one, identify the unfavourable conditions/contexts early, so that an alternative tool, or evacuation, may be quickly effected?

1. Probably not. Indicators that signal 'oversold' conditions, in 2008, would likely have got you in trouble.

2. Any condition where the rise in volatility exceeds the model that you have built, its capacity to withstand.

3. Where a measurement of volatility, say the VIX [as an example] rises above the VIX level that predominated in your backtesting.

jog on
duc

 

[ducati916]

1. It seems easier to go back to what constitutes an argument.
People are making statements which lead to a reasoned conclusion.

2. If their statement (premise in this case) is not supported by their evidence then they cannot arrive at a sound conclusion.

3. The construction of the argument has nothing to do with belief.

4. You seem to be creating some mystique over the nature if the evidence. As I see it, that would be an entirely separate argument and dilute something relatively straightforward with epistemology.
Whenever minutiae come into play philosophy falls into the hole of infinite regress.

1. In general, I agree with you. However the nature of this thread is abstract and attempting to universalise the concepts. Thus, an 'argument', which is very specific, while it could be made [with some difficulty] would possibly bog down even more than this discussion already has.

An argument in which we move from one idea that is known [proven] to be true to the inferential move, whereby we move from that first fact to a second idea, that can also be judged a fact, based on the strength and force of the first fact [idea] is how I prefer to argue, however, not all do or can.

When we deal with faith, we enter a fuzzy territory where as Cynic states, the absence of proof, is not proof of absence.

2. Agreed re. [1] but, as I said, it gets fuzzy.

3. I suppose it depends on what that 'belief' is founded on. A proven fact...then belief is acceptable. On a statement of faith...in the context of an argument, not a strong place to be, but on a faith basis, belief will [likely] occur.

4. The strength of the proving evidence that established in the first instance a fact as a fact, is important. Obviously evidence can take many forms.

jog on
duc

 

[rederob]

1. In general, I agree with you. However the nature of this thread is abstract and attempting to universalise the concepts. Thus, an 'argument', which is very specific, while it could be made [with some difficulty] would possibly bog down even more than this discussion already has.

An argument in which we move from one idea that is known [proven] to be true to the inferential move, whereby we move from that first fact to a second idea, that can also be judged a fact, based on the strength and force of the first fact [idea] is how I prefer to argue, however, not all do or can.

When we deal with faith, we enter a fuzzy territory where as Cynic states, the absence of proof, is not proof of absence.

2. Agreed re. [1] but, as I said, it gets fuzzy.

3. I suppose it depends on what that 'belief' is founded on. A proven fact...then belief is acceptable. On a statement of faith...in the context of an argument, not a strong place to be, but on a faith basis, belief will [likely] occur.

4. The strength of the proving evidence that established in the first instance a fact as a fact, is important. Obviously evidence can take many forms.

jog on
duc

Arguments will be either deductive or inductive.
If you are dealing with information unique to an individual - ie information which is beyond reasonable measure - it holds value to the individual and whoever wishes to accept it. It may at best be a weak inductive argument.

 

[cynic]

I would argue that it does.

Evidence is information which assists the decision maker in determining facts. To do so, the evidence must fulfil two criteria when directed at information.

(a) what is the ultimate fact to which the evidence is intended to make more/less likely; and
(b) how does the evidence make the ultimate fact more/less likely.

(i) facts in issue; and
(ii) facts going to credibility.

Facts going to credibility are those that bear upon the reliability of witnesses or a particular element of the witnesses evidence.

The evidence does not have to make a fact certain [stronger if it does] but the evidence must make the fact more/less probable.

Weight and probative value are words used to describe probability. Numbers [again] inform probability, viz. it can be measured by numbers, which gives weight or probative values.

(a) Classical Approach Frequency: 4 Jacks in a pack of cards 4/52 = relative frequency.
(b) Relative Frequencies: is where you have X observations with Y from your variable, which gives you Y/X expressed as a %.
(c) Likelihood ratios: to test the significance of the evidence [footprints found were made by a male adult shoe, low likelihood. Footprints made by an adult male shoe with a crack in the sole, higher likelihood.]

The above are merely statistical hypotheses.

(d) Circumstantial evidence: where like strands of a rope, no individual strand is sufficient, but when all together, they can bear the burden [of proof].

When assessing the evidence in combination, the rule is that the probative value of multiple items of evidence supporting the same factual allegation is greater in combination than the sum of the parts.

An example: there is a bank robbery and Ms Brown is charged with the offence.

Witness A states: the bank robber was female; and
Witness B states: heard the robber say her birthday was in July; and
Witness C states: saw the robber drive away in a red car.

These strands cumulatively add to: 1/2 x 1/12 x 1/10 = 1/240

The Crown is entitled to say that Ms Brown is 240 times more likely to have these characteristics than a random person [assuming of course that she is in point of fact female, has a red car and her birthday is in July].

Note that there does exist a logical flaw in the methodology.

Unless 1/240 of the adult populace, happen to be, July born female owners of red cars, then the Crown's assessment would be unreasonable.

However, I grant that said flaw, does not detract from the presented argument.

Therefore the outcome, which in this case is 'belief', I would argue that this can be very dependent on the [quantity and or quality] of evidence given.

If I have understood correctly, I counter argue, that the level of confidence in 'belief' is not the critical issue here.

The critical issue is, not so much the validity (or lack thereof) of the philosophy being contested, but the question of which tools might be appropriately employed for validity testing.

The arguments presented thus far, are things for which belief, whether past, or present, is associated, and therefore qualify as, examples of, correlations.
Such correlations are compatible with validity, because these would be expected in the event the philosophy happens to be true.

The very concepts of argument structure, scientific method, probabilistic assessment etc.are all things that either, are, or were believed. As such those too serve as correlations adding to the body evidence.

Witness A: "Witness B is about to speak the truth"
Witness B: "Witness A just told a lie"
Witness C: "I agree with both Witnesses A and B"

Could there be some similarity between these paradoxical statements and the matters under discussion?
=====================================================================================
So let's consider what this might have to tell about financial markets.

The Funny Mentalist says hold!
(Because the alphabet of acronyms, when viewed in its entirety, says all is well!)

The Geek says sell!
(Because the milk shakes, i.e. mooing beverages, bollocks quartets, mendacious corn chips, i.e. fibbing nachos, and her surfie dude mate, named Elliot, are all in probabilistic agreement, this one's going doowwwnnnn!)

The Quantum Lord says:"Doesn't matter, just order one with the lot!
(Because it simply doesn't matter!)

The Darts Champ says: "Whoa! Wait a minute! I think it best, to sit this one out!"
(Because some of the darts fell out of the board, before he could determine what it all meant.)

The Lunatic says: "Sell now, and then in two weeks time, buy back in!"
Because the moon is full, and the new moon is due, a fortnight hence.)

The cynic says: "You have all gotten it wrong, but that's alright!"
(Because cynics profit from traders getting it wrong.)

 

[rederob]

The critical issue is, not so much the validity (or lack thereof) of the philosophy being contested, but the question of which tools might be appropriately employed for validity testing.

Some unusual language has been used in these examples as philosophy is the realm which includes logic.
WRT to testing a premise, what matters is that it can be determined true.
If a premise is not true, the conclusion must be unsound as the argument is not valid.
The idea that "choice of tools" per se would be a determinant is unlikely. However, separate arguments would ensue if the means for testing truth were disputed.
In strict terms, and using duc's bank robbery example where Ms Brown is charged with an offence, and witness statements from A, B & C (regarded as "facts" for the purpose of simplicity here), the prosecution might conclude his/her case by saying "and that person is standing in the dock". The reasonable deduction, however, was only in relation to a person with those characteristics as standing in the dock. Thus, a deductive argument in this example provides a probabilistic outcome for whom the robber might include.
So let's instead turn this into a inductive argument from the prosecution. Here are some real world example where statistics lead to initial prosecution, and later the verdicts being overturned.
Theoretical probability of guilt on grounds of gender/DOB/colour of car could be 240:1 or 10000:1, but if Ms Brown owns the only red car in town then the prosecution's case begins to look pretty solid.
In logic it would not however be considered "sound" as the statements supporting it remain probabilistic.

 

[ducati916]

1. Arguments will be either deductive or inductive.

2. If you are dealing with information unique to an individual - ie information which is beyond reasonable measure - it holds value to the individual and whoever wishes to accept it. It may at best be a weak inductive argument.

A deductive argument moves from the universal to the particular: therefore the universal must be a true fact for any inferences to be true. This is the [synthetic] a priori.

The inductive moves from a particular to the universal. The premise in these arguments must both be true and sufficiently large [wide] in scope to make it a sound foundation for an argument. This is empiricism. This has been predominantly employed as the 'scientific method', viz. experimental. This is the 'effect' looking for a 'cause'.

Arguments of faith cannot be deductive or synthetic a priori. Arguments of faith can be inductive. There is of course a major issue: which is that to become a true cause of the effect, evidence sufficiently strong/convincing must be adduced to raise the cause above a mere probability to a truth.

The argument [in logic] put forward currently is: the absence of proof, is not proof of absence. The burden of proof is therefore shifted from him making the assertion, to those refuting the assertion, if any final answer is possible. Generally, in questions of faith, they never are as a definitive proof in logic is very difficult find.

Hence in my evidence example we have a crime, which is the particular [effect] that must be assigned [in the highest probability approaching or finding truth] a cause. A probability, while not 'the' truth, can be sufficient as a proof.

jog on
duc

 

[rederob]

A deductive argument moves from the universal to the particular: therefore the universal must be a true fact for any inferences to be true. This is the [synthetic] a priori.

The nature of an argument depends on the premises.
The premises can be determinate of either universals or particulars.
The premises may well be true, but they may only provide grounds for the conclusion, ie a more likely outcome (probability), as was demonstrated in the examples above relating to Ms Brown.
I won't tackle universals as there are literally too many schools of thought and they merely confuse. For example, Rex is a dog. Rex is particular (ie one dog). The concept of "dog" is of many, therefore universal. "Dog" now becomes one kind, therefore particular. If "dog" is a set, or a class of things then it is not material (ie of substance). Rex is a real living dog .

A probability, while not 'the' truth, can be sufficient as a proof.

I can only say that it would not be a sound conclusion to draw. In a legal sense, would you say it was beyond reasonable doubt?

 

[ducati916]

1. The nature of an argument depends on the premises.

2. The premises may well be true, but they may only provide grounds for the conclusion, ie a more likely outcome (probability), as was demonstrated in the examples above relating to Ms Brown.

3. I can only say that it would not be a sound conclusion to draw. In a legal sense, would you say it was beyond reasonable doubt?

1. I would argue that the nature of an argument is whether it is a deductive or inductive argument.

2. True, where the argument is inductive. A deductive argument [assuming a true premise that is being reasoned from] that proceeds in valid logical manner, must be true.

3. Correct. And legal proof is based upon an inductive argument. You have the crime, which is the effect or outcome, you are reasoning backwards to elicit the cause [whodunnit].

jog on
duc

 

[cynic]

Thanks for taking the time to read and respond to my post/s. It may seem immodest of me saying this, but, despite the diversity and depth of our respective convictions, our willingness to continue engagement in discussion, is meritorious.
This is intended as a a compliment to both you and I.
I am patting myself on the back,(not many canines know how to do this), and I invite you to do likewise.

Some unusual language has been used in these examples as philosophy is the realm which includes logic.

Knowing that the English language has some nuances, additional consideration of word choice can sometimes prove beneficial.

Needless to say, words with multiple, distinctly different (albeit often relatable) definitions, often require additional care in their usage and interpretation. Sometimes contextual considerations are sufficient for discernment of intended meaning, othertimes clarification may need to be sought, lest misunderstandings arise.

WRT to testing a premise, what matters is that it can be determined true.

Many might agree, however, I do not share your confidence.
Scientists often devise experiments in their efforts to disprove hypotheses.
Some might say that falsification is an equal, or perhaps even paramount consideration.
Others might argue that falsification equates to proof of a hypothesis that is an inverted form of the hypothesis to which falsification is applied.

If a premise is not true, the conclusion must be unsound as the argument is not valid.

I can agree that the argument is not valid when one or more of the premise/s are untrue.

A faulty argument's conclusion, may or may not happen to be true.

Years ago a piece of machine code I wrote, performed exceptionally well on the majority of my tests, by producing correct answers to the problems it was presented, it wasn't until the later stages of testing a few incorrect results alerted me to the presence of a fault in the coding.

Perhaps its not the ideal analogy, given there can be dispute about whether the written software represents a faulty logical process or a faulty premise.

Whichever way it is sliced, as an analogy for a faulty argument, it does serve to demonstrate the potential for serendipitously correct conclusions.

The idea that "choice of tools" per se would be a determinant is unlikely. However, separate arguments would ensue if the means for testing truth were disputed.

Methods chosen for testing truth, can be unsuited to the testing of some claims.
Eg. Consider the following claim:

"The only things that exist are those that are believed to exist."

Out of the many popular methods of contesting such a claim, what argument can be presented to challenge it?

Surely, any argument will by the time it is presented, be already known, else it could never be presented!

And if the presenter thinks the argument presented is dubious, the presenter is still presenting something that is believed to exist (ie. a dubious argument).

Every argument presented, is itself, at minimum, a correlation adding to the body of evidence supporting the claim it purports to challenge.

In strict terms, and using duc's bank robbery example where Ms Brown is charged with an offence, and witness statements from A, B & C (regarded as "facts" for the purpose of simplicity here), the prosecution might conclude his/her case by saying "and that person is standing in the dock". The reasonable deduction, however, was only in relation to a person with those characteristics as standing in the dock. Thus, a deductive argument in this example provides a probabilistic outcome for whom the robber might include.
So let's instead turn this into a inductive argument from the prosecution. Here are some real world example where statistics lead to initial prosecution, and later the verdicts being overturned.
Theoretical probability of guilt on grounds of gender/DOB/colour of car could be 240:1 or 10000:1, but if Ms Brown owns the only red car in town then the prosecution's case begins to look pretty solid.
In logic it would not however be considered "sound" as the statements supporting it remain probabilistic.

I largely agree with your reasoning here as it appears to be along similar lines to mine.
===========================================================================================
"Balance of probabilities" and "beyond reasonable doubt", are to my understanding, representative of compromises that have been made in favour of practicability, due to the impracticability of the known alternatives.

Care needs to be taken when choosing/devising methodology for probabilisic assessments. Some theoretical applications have been conveniently founded upon the premise of an even distribution of distinct qualities across the population under consideration.

This convenient foundation can ofttimes prove inconvenient to those applying certain methodologies to uneven distributions.

Consider the following example:

The four outermost dots on the 5 dot face of a standard six sided die have worn away leaving only the middle dot.
Effectively that face has become identical to the 1 dot face, meaning that 5 is no longer a possible outcome.
There now exist 5 distinct outcomes, from each throw of the die.
None of those outcomes holds the probability of, 1 out of 5, because the outcome distribution is uneven.

1, holds a 1 out of 3 (i.e. 2 out of 6) probability

2,3,4 and 6 each hold a 1 out of 6 probability

Now let's consider how this knowledge might prove beneficial in analysis of financial markets.

Whilst a market for a chosen instrument is open, the price has the potential to move up, move down, or remain motionless.

So three qualities of price behaviour have been defined with respect to movement (or absence thereof).

Based upon the history of a financial instrument under consideration, is attribution of equal probability, to each of the above qualities justified?

Are larger moves to be treated as one move, or multiple consecutive moves, in a direction?

And what of immobility, is a period of same to be accounted as one, or multiple consecutive exhibitions of price behaviour?

Does time enter into consideration? If so, how is it to be apportioned in respect to each of the price behaviours?

Is the distribution of the price behaviours even?

How might this information be useful to a market strategist?

 

[ducati916]

Methods chosen for testing truth, can be unsuited to the testing of some claims.
Eg. Consider the following claim:

"The only things that exist are those that are believed to exist."

Out of the many popular methods of contesting such a claim, what argument can be presented to challenge it?

Surely, any argument will by the time it is presented, be already known, else it could never be presented!

And if the presenter thinks the argument presented is dubious, the presenter is still presenting something that is believed to exist (ie. a dubious argument).

Every argument presented, is itself, at minimum, a correlation adding to the body of evidence supporting the claim it purports to challenge.

You are simply repeating the same argument as previously: and the answer to that is also still the same: that something believed [in faith] that cannot be proven and cannot be disproven, simply remains an item of faith.

There is no getting out of the rabbit hole.

* I would qualify that by saying on present knowledge. As knowledge increases, some conundrums are resolved.

jog on
duc

 

[ducati916]

Care needs to be taken when choosing/devising methodology for probabilisic assessments. Some theoretical applications have been conveniently founded upon the premise of an even distribution of distinct qualities across the population under consideration.

The writing of insurance premiums being one example.

jog on
duc

 

[ducati916]

Now let's consider how this knowledge might prove beneficial in analysis of financial markets.

Whilst a market for a chosen instrument is open, the price has the potential to move up, move down, or remain motionless.

So three qualities of price behaviour have been defined with respect to movement (or absence thereof).

1. Based upon the history of a financial instrument under consideration, is attribution of equal probability, to each of the above qualities justified?

2. Are larger moves to be treated as one move, or multiple consecutive moves, in a direction?

3. And what of immobility, is a period of same to be accounted as one, or multiple consecutive exhibitions of price behaviour?

4. Does time enter into consideration? If so, how is it to be apportioned in respect to each of the price behaviours?

5. Is the distribution of the price behaviours even?

6. How might this information be useful to a market strategist?

1. Clearly not: as a die has a fixed probability. A market moves in response to future events. It is not currently possible to foretell the future, thus the event, thus the response. While a knowledge of the past is useful [Santayana], history rhymes [Twain].

2. Multiple moves. The market is not a true stochastic process.

3. Multiple.

4. Yes. On market open and market closed. Future events are stochastic, the market is not.

5. No.

6. Mainly because you can devise methodologies that trade and profit from these inefficiencies.

jog on
duc

 

[barney]

@rederob @cynic @ducati916

Interesting dialogue from you three guys (I am making the premise that you are all guys … could be totally errant in that assumption)

Anyway, great conversation, and superb articulation ….. I sense potential Law, Philosophy and Language degrees all round

ps I have none of the above but still appreciate the debate ….
Carry on forthwith and hitherto gentlemen

 

[rederob]

1. I would argue that the nature of an argument is whether it is a deductive or inductive argument.

A premise which is clearly probabilistic can be true, eg. most dogs are not called Rex.
As will all premises of a deductive argument if it is sound.
It's a guess to assume from only the conclusion that an argument is of one type rather than the other.
Put another way, an argument could be constructed such that new evidence presented (ie. not as in the original example) proved that only Ms Brown could have been the robber, as this was deduced (see comments at #3).

2. True, where the argument is inductive.
A deductive argument, [assuming a true premise, that is being reasoned from] that proceeds in aalid logical manner, must be true sound.

3. Correct. And legal proof is based upon an inductive argument. You have the crime, which is the effect or outcome, you are reasoning backwards to elicit the cause [whodunnit].

Not necessarily. For example, video footage of the robbery shows Ms Brown's face clearly, plus the gun she fired into the air. The stolen money and a gun was found in Ms Brown's red car. It was in fact found by police who had raced to the scene immediately an alarm was set off. Serial numbers on the stolen money proved it had only arrived that morning and was not for release that day. Rifling on the round dug from the ceiling proved the gun belonged to Ms Brown. Gun powder residue found on Ms Brown's gloves as she entered the car matched a subsequent round fired from the same gun by ballistic and forensic experts. Capping all this off was a signed confession which went into copious detail regarding her motives, planning, and execution.
Although some aspects of these events are probabilistic, others defy logical explanation unless Ms Brown was the actual perpetrator.

 

[cynic]

You are simply repeating the same argument as previously: and the answer to that is also still the same: that something believed [in faith] that cannot be proven and cannot be disproven, simply remains an item of faith.

There may exist a very subtle distinction, between that and my earlier wording of the argument.
But for the purposes of highlighting the importance of appropriate tool selection, it is near enough, to be good enough.

There is no getting out of the rabbit hole.

* I would qualify that by saying on present knowledge. As knowledge increases, some conundrums are resolved.

Whilst we persist with our current choice of tools, I am tempted to agree.

But what if there wasn't only one tool available?
What if other tools existed that could be implemented in a practical way for harvesting of evidence?

A man spends a lifetime eating meals of solid food, served in a bowl, via use of a fork.

One day the waiter serves him a bowl of soup.

The man endeavours to consume the soup, with his fork.
(This fork has served him so faithfully for so many years and for that reason, he has the utmost confidence in its universal utility.)

After repeatedly failing in his concerted effort to extract some soup from the bowl, he calls the waiter over and complains:

"There must be some mistake, this cannot possibly be edible food! It keeps dribbling through the tines of my fork!"

To which the waiter replies:

"I am so very sorry to hear that sir!

Perhaps sir might like to consider using a spoon!

Many of our patrons have given assurance they find the soup far more nutritious and enjoyable when consumed that way."

=======================================================================================================

Now to relate this to financial markets.

Know thy tools, know their strengths and limitations!

So a trader's toolkit serves well for some time and then suddenly the market decides it no longer wishes to cooperate.

Perhaps this trader is a Quantum Lord presiding over an army of open positions as they are flattened by a formula one steam roller.

Or perhaps a Geek suffering the death by a thousand stop outs.

Or perhaps a Funny Mentalist puzzling over how Enron managed to pass audits.

Or a Lunatic wondering how the Draghiton managed to subvert the power of the moon.

Or a Dartist Trumped by a Don.

Or a cynic wondering what the FTSE is going on with the EUR/USD.

What insights might a trader glean from these happenings?

 

[rederob]

"WRT to testing a premise, what matters is that it can be determined true."

Many might agree, however, I do not share your confidence.
Scientists often devise experiments in their efforts to disprove hypotheses.
Some might say that falsification is an equal, or perhaps even paramount consideration.
Others might argue that falsification equates to proof of a hypothesis that is an inverted form of the hypothesis to which falsification is applied.

Logic does not have an alternative here, so if you were to maintain that viewpoint it becomes impossible for you to be treated coherently.
All science is probabilistic.
However, the outcome of a completed experiment can be known with certainty.
This is a different sense to proving, for example, if the scattered ashes from a cremated body can be resurrected. Without evidence to the contrary, most people would think it a highly unlikely event. From an epistemological perspective many aspects of science give us knowledge which can be relied upon in all known circumstances. For all intents and purposes, these would be called facts. As would outcomes of completed experiments.
For a scientist to falsify an hypothesis, he/she would need to present a counter argument, ideally deductive, which contained premises which could be proven true. If the scientist has no basis for testing truth then no reasoned conclusion can be sound, let alone most probable.

 

[Skate]

Information overload, but this is same stock universe, same time period, using different system code (my Amibroker interpretation of Nick Radge's "Weekend Trend Trader").

Again, also remember trading the "optimal" = "most aggressive" number of positions may give you max theoretical returns but also DRASTICALLY increase your Risk of Ruin (i.e. chance of blowing your account, particularly in the first 2 years).

Suggested reading:
http://bettersystemtrader.com/riskofruin/
View attachment 92083
View attachment 92084

First off, I would like to echo @barney words about the interesting topic being discussed by @rederob @cynic & @ducati916, well done guys.

https://www.aussiestockforums.com/posts/1014049/

There is a great Thread: Hypothetical Monthly Momentum Portfolio vs. Index

Comments
I'll make the comments in here in the 'Dump it here' thread, it will save clogging the Hypothetical Monthly Momentum Portfolio thread.

Back-grounding on Metrics relating to Position Sizing
This is going to be a large post to explain System Metrics used in strategy development & I will relate that to Position sizing. The interpretation is sometimes ambiguous in its understanding & in isolation will tend to mislead. Using the CAR/MDD metric for position sizing is weak at best & can lead to faulty thinking.

Metrics
Metrics are one of the most academic topics about trading systems, they are guideline to let you know how your system has performs over time. There are so many metrics, I find it difficult to know which one is more appropriate to evaluate a system which is really quite frustrating for me.

A good metric is your trading account balance.

Parameters
Choosing parameter values can be tricky. It depends on system under test & its complexity of code. It is important to understand that all smart optimization methods work best in continuous parameter spaces as some algorithms may experience trouble finding optimum value.

Moreover
We will use CAR/MDD for consistency of graphics to add to the confusion. The graphs are from the Hybrid Strategy. (only the dates have changes) The CAR/MDD ratio is the metric used in relationship to optimisation for optimal position sizing & it’s a standard Amibroker metric that is very useful to measure system performance as it analyses percentage annual growth divided by the maximum drawdown.

CAR/MDD
The compound annual return divided by maximum % drawdown & the higher the CAR/MDD score, the smoother the equity curve and the better the system, so they say.

Profit factor
The Profit factor can be a good measure, since it divides the profit of winners by the loss of losers. It’s a quick way to look at the chances of your system being profitable.

Risk-reward ratio
Risk-reward ratio can be measured by dividing the slope of the equity line by the standard error of the equity line. It’s an important metric to be able to calculate the optimum position sizing for a system.

Sharpe ratio
The Sharpe ratio describes how much return you receive from the added volatility for holding a trade. Basically, the higher the sharpe ratio the better the system. However, Sharpe ratio has come under criticism for not recognising that upward volatility is more desirable than downward volatility.

Sortino ratio
The Sortino ratio improves upon the Sharpe ratio by isolating downside volatility from total volatility.

K-Ratio
K-Ratio is another popular measure and examines the consistency of an asset’s return over time also the K-Ratio detects inconsistency in returns.

Poker (rehash)
The game of Poker has lots of rules but consider if there were 'NO rules' in the game of Poker. Imagine if you could have 40 or even 52 cards. Tell the dealer "Just keep them coming" till you have a great hand, a winner hand (a winning Portfolio)

Why use a handbrake
As traders, why do we limit our Portfolio (our hand of cards) to 5, 10, 15 or even 20 positions in our trading portfolio or trading system (I personally use 40 & sometimes 53 positions)

"QUESTION"
Look at the optimisation results for position sizing & tell me using the CAR/MDD metric how many positions should my Hybrid Portfolio have ? (for clarity the MaxPos are on the side of the graphic)

The question is directed at
@Wyatt
@Newt
@debtfree
@peter2
@qldfrog
@Lone Wolf
@willy1111
@Habakkuk
@jjbinks











1. 1990-2019 position size Optimisation XAO Capture

2. 2015 Calendar year position size Optimisation Capture

3. 2015-2016 Financial year position Optimisation Capture

4. 2016 Calendar year position size Optimisation Capture

5. 2016-2017 Financial year position Optimisation Capture

6. 2017 Calendar year position size Optimisation Capture

7. 2017-2018 Financial year position Optimisation Capture
More if required..

Skate.

 

[willy1111]

The beauty of being self employed is that you can do it which ever way you like

Personally I wouldn't/don't do it off CAR/MDD alone.

I like to know what the MDD is in percentage terms to see if I can handle it. I like to see what it is/was through the GFC period, starting in 2006 so the system has a chance to get fully invested and then going through 08 and out the other side.

 

[Newt]

On position sizing Skate, totally agree you have to be extremely careful with any optimising, aiming for generic but hopefully significant results across many market regimes and time periods.
One other important factor you've hinted before is position size itself. For a large portfolio, it may not be wise to trade the theoretical optimum for fear of moving the market, as you've mentioned here before.

It sounds like you spent many years working up your systems. If you're willing to share, how did you adjust parameters? (e.g. manual tweaking of parameters page, or careful optimising, or both?)

 

[Skate]

The beauty of being self employed is that you can do it which ever way you like

Personally I wouldn't/don't do it off CAR/MDD alone.

I like to know what the MDD is in percentage terms to see if I can handle it. I like to see what it is/was through the GFC period, starting in 2006 so the system has a chance to get fully invested and then going through 08 and out the other side.

"Personally I wouldn't/don't do it off CAR/MDD alone"
Neither would I, that was the point of the post, I wanted to get this point across. Using the CAR/MDD metric for position sizing is weak at best & can lead to faulty thinking. Of all the metrics you can use for system development, position size is well down on wish list.

"I like to see what it is/was through the GFC period, starting in 2006 so the system has a chance to get fully invested and then going through 08 and out the other side"
No can do, I only have data back to June 2009

Data back to June 2006
My development days are over, I have 3 Norgate licences (Silver package only). When I was in the Strategy development phase I had the advantage of the Platinum package with 25 years of historical Data that included access to delisted securities and historical index constituents.

F.Y.I (Silver Package)
Norgate Data - Stock Market Silver Packages: Has access to currently-listed equities (and all other listed security types), indices & current index constituents with 10 years of historical data. (from June 1992 on)

Skate.