The Contents of a Bar

[ducati916]

Before I babble any further ...... the "partial post" you originally quoted ... was that yours or someone else's? Cheers.

No not mine. The partial post belongs to another.

jog on
duc

 

[ducati916]

Suppose that over time traders have experimented with trading rules, drawn from a very wide universe of trading rules, perhaps tens of thousands of different iterations. As time progresses the rules that happen to perform well historically, attract more attention and are considered serious rules by the trading community, while unsuccessful rules gradually fall by the wayside.

If enough trading rules are considered over time, some rules, by pure luck, even in a large sample, will produce superior performance, even if they do not genuinely possess predictive power.

Mechanical systems, through talking about 'probability' and 'expectation' are talking about predictive power.

The past, does not predict the future.

It would seem that I am denigrating the role of history. That is not my intention. After all "Those who cannot remember the past are condemned to repeat it."

jog on
duc

 

[Quant]

At first this thread ' appeared ' to be in good faith but now it seems something else , yeah all systems traders running statistical analyisis based on empirically significantly reliable rules are running on ' pure ' luck . There is another participant who will never get another second of my time and almost with as much doubt as the efficacy of systems trader will now in all likely hood stalk anyone who thinks differently . The only thing I can glean from all this is the OP "cant " systems trade .... Rock on

freaking fragile lawyers

 

[tech/a]

Analysis of data is commonplace in scientific research

Why is it that Data analysis in the Financial world is seen as less accurate.

Duc
You've been around a long time
All I've ever seen from you is arguement without
Any positive input to a trading plan or strategy

To be honest I think you enjoy circular meaningless
Exchange

 

[ducati916]

Analysis of data is commonplace in scientific research
Why is it that Data analysis in the Financial world is seen as less accurate.

Simply because of the number of uncontrolled variables. Some examples:

(a) each ordinal or cardinal buy/sell order of each market participant; and
(b) the information driving those ordinal/cardinal decisions.

Duc
You've been around a long time
All I've ever seen from you is arguement without
Any positive input to a trading plan or strategy

What is positive input?

Can it not be legitimate questions that create some thought/debate? My questions seek to challenge some of the assumptions inherent in mechanical trading based on historical time series data. If in point of fact these are not assumptions at all, then the challenge can easily be answered. If not, then the assumption that has been made, is potentially a uncontrolled risk or variable that may, under certain or specific conditions bite you.

jog on
duc

 

[ducati916]

At first this thread ' appeared ' to be in good faith but now it seems something else

, yeah all systems traders running statistical analyisis based on empirically significantly reliable rules are running on ' pure ' luck .

The thread is in 'good faith'. I engage with those that wish to, and leave in peace those that do not wish to.

You use 3 words: 'empirically, significant[ly], reliable.'

(a) Empiricism refers to historical, or the past; which implies that
(b) significant, referring to confidence levels calculated statistically, are high; which then
(c) gives reliable, or robust rules that have predictive power.

(a) + (b) does not = (c)

I simply provided an example, whereby luck alone could, account for the success of the rules selected.

jog on
duc

 

[ducati916]

A bit of fun for the systems traders out there.


There is a set of trading rules that only fails [true positive] 1/1000 times. However, when entering the trade, there is a 5% false positive rate in screening trades, which can take you out of the trade based on your risk management rules, thereby missing the return. You are only authorised by your partner/manager/firm to take a maximum of 10 trades each year. Trades are screened through this rule set randomly from all manner of securities, [stocks, futures, options]. A trade is screened and it is positive. What is the probability that the trade will fail? Will you take the trade?

jog on
duc

 

[ducati916]

So........a bit more fun:

Using my Bloomberg, I’d just entered in an index built from companies with “cat” in their names -- yes, the furry felines -- hit a button and watched it back-test to an 849,751 percent return.

I notified Andrew Ang, head of factor investing strategies at BlackRock Inc. Everything in my program was by the book, I assured him. It was rules-based, equal-weighted and premised on a simple story -- that people love cats.

So how, exactly, did I go about investing in cats? Factor funds rely on formulas, preset criteria that tell you which stocks to include and which to chuck out. It’s the idea behind things like value ETFs, which gather groups of shares that share the common characteristic of cheapness. The idea is that put together, they’ll beat the wider market.

My model buys any U.S. company with “cat” in it, like CATerpillar, or when “communiCATion” is in the name. It rebalances quarterly to keep trading costs low. That’s important for when Vanguard or BlackRock license it and charge a competitively low fee.

Because of my stubborn desire to produce claw-some returns, I took my thesis and ran with it. Fine, so my first few trials didn’t spit out exactly what I wanted. No biggie, I’ve got the statistical resources of Bloomberg LP at my fingertips -- so I tinkered with the data until it did.

At first, I only invested in companies beginning with C - A - T to capture the essence of my investment thesis.

Not great. But expand the data-set a little, CAT anywhere, and the returns look stellar, making my hypothesis look better. In the scientific community, this is called p-hacking

Contains some of the pitfalls inherent in backtesting. They can all be fixed of course, but you need to be aware of the issues.

jog on
duc

 

[rb250660]

Anyone with 1/10 of a brain know this is never going to work. So what's your point? His back testing methodology was probably perfect, it's the underlying theory as to why profits are generated that is flawed. I fail to see your point about bad back testing. So tell me, how do you fix it? Change the letters to DOG? Or might it be better to say that after X bar pattern, the probability of Y happening, is Z.

I say that if done correctly
X + Y does = Z

 

[ducati916]

it's the underlying theory as to why profits are generated that is flawed.

Correct.

People can convince themselves of all manner of fundamental/economic/other reasons as to why the profits are being generated.

Some may rely on the p-test to validate their hypothesis as to why/how profits are being generated.

All science entails human judgement, and using statistical models doesn’t relieve us of that necessity. Working with misspecified models, the scientific value of significance testing is actually zero – even though you’re making valid statistical inferences! Statistical models and concomitant significance tests are no substitutes for doing real science. Or as a noted German philosopher once famously wrote:

There is no royal road to science, and only those who do not dread the fatiguing climb of its steep paths have a chance of gaining its luminous summits.

In its standard form, a significance test is not the kind of “severe test” that we are looking for in our search for being able to confirm or disconfirm empirical scientific hypothesis. This is problematic for many reasons, one being that there is a strong tendency to accept the null hypothesis since they can’t be rejected at the standard 5% significance level. In their standard form, significance tests bias against new hypotheses by making it hard to disconfirm the null hypothesis.

jog on
duc

 

[tech/a]

Duc
What babble are you going on about

More circular waffle with absolutely no value.

 

[ducati916]

More circular waffle with absolutely no value.

It is abstract, certainly. You seemingly find no value in the abstract, which is fine. Some might.

jog on
duc

 

[ducati916]

Statistical analysis can and will produce an edge .. I think that's beyond debate

Very far from being beyond debate.

The statistical outliers, which give the 'probability' of an outlier event have no place in the financial markets.

August 31 1998 - 1/20,000,000, expectation 1/100,000 years
3 days in August - 1/500,000,000,000

In 1997 Dow falls 7.7% - 1/50,000,000,000
The 1987 crash - less than 1 in 10 raised to 50 power. Essentially impossible.

So the point of all this?

When the backtesting crunches all of the data, available to be analysed, we have the following possibilities:

(a) the back-tester excludes these points in constructing their trading plan; or
(b) the back-tester includes these points and adds a risk management strategy to the trading plan.

If (a), well good luck with that.

If (b), then this is more interesting.

Day-trading stocks should manage the risk as about as well as it can be managed. Flat every market close and a new entry the following day, unless a trading halt is called. Then at the next open, well who knows. Futures day trading, if limit lock days still exist, is not as risk free as stocks on any given trading day. Day trading options, pretty much a waste of time, although with the weeklies, not impossible. As long as you are long volatility, your risk is managed to entry price.

Many [most] will position trade over an indeterminate time period, but longer than 1 trading day. The previous declines and odds were from an index. An individual stock will likely be a multiple of any index move.

Of the risk management strategies that I have seen on this site, which is only I'm sure a small sample, all would essentially be crushed in these type of outlier events. For those who traded through the market collapse of 2008, which I did, Bear and Lehman vaporised and Citi etc trading at fractions. Looking at the historical charts [bars] gives a very false impression, the contents of the candle is telling fibs. The bars imply that trades took place pursuant to a stochastic pricing process intra-day. Nothing of the like actually took place: liquidity just evaporated, in the US markets which claim the deepest most liquid markets in the world. One consistency is that in times of [extreme] stress, liquidity does not exist.

Most traditional risk management strategies that I have seen on this site, require and rely on there being liquidity. If there is not, you do not have a any risk management at all in place.

Mathematics and statistics in particular, are seriously flawed tools, at specific points in market [dis]function. They are fine at most other times. That is the problem.

jog on
duc

 

[tech/a]

August 31 1998 - 1/20,000,000, expectation 1/100,000 years
3 days in August - 1/500,000,000,000

In 1997 Dow falls 7.7% - 1/50,000,000,000
The 1987 crash - less than 1 in 10 raised to 50 power. Essentially impossible.


Where and how do you determine these probabilities??
Your discussing statistical analysis in a scientific manner and using (From what I see)
Made up statistics.

Again meaningless argument.

 

[ducati916]

Where and how do you determine these probabilities??
Your discussing statistical analysis in a scientific manner and using (From what I see)
Made up statistics.

Again meaningless argument.

No they [the statistical calculations] are accurate to the probabilities.

As to meaningless, clearly they are meaningless to you, which is fine. Maybe you don't use statistics or probabilities in your trading. For those that do however, the problem with using statistical analysis and probabilities in their trading [mechanical systems] should be obvious.

jog on
duc

 

[tech/a]

Haha

Duc much has passed since we first met.( Well nearly )

Real Scientific Statistical analysis is in the Family.

But
Jog on

 

[ducati916]

Haha

Duc much has passed since we first met.( Well nearly )

Real Scientific Statistical analysis is in the Family.

But
Jog on

tech/a,

Not all uses of statistical analysis are inappropriate, for example looking at physical characteristics such as height, weight. These lend themselves to statistical analysis and the results will be reliable.

Statistical analysis applied to the financial markets is not [as] reliable as its proponents advocate. It is useful until it is not. The problem being that you never know when it stops being useful.

So really this thread is more about how to manage the risk than anything else. Which by implication is critical of theory touted as risk management, that is no such beast.

jog on
duc

 

[ducati916]

So in that vein:

(a) You are a swing trader. You trade a mechanical system, or a discretionary system or even a hybrid of the two.

(b) You enter a new long position on day 1, or possibly open a new 'portfolio' of long positions.

(c) On day 2 that position or portfolio gaps 15% lower and is trading lower fast.

Statistical analysis of this scenario would tell you that the probabilities of this happening are remote to impossible, particularly if you have used some form of (a) backtesting and (b) particularly in the portfolio scenario.

What would, or should, the 'average' individual trader, trading his own money do in this circumstance?

jog on
duc

 

[Triathlete]

So in that vein:

(a) You are a swing trader. You trade a mechanical system, or a discretionary system or even a hybrid of the two.

(b) You enter a new long position on day 1, or possibly open a new 'portfolio' of long positions.

(c) On day 2 that position or portfolio gaps 15% lower and is trading lower fast.

Statistical analysis of this scenario would tell you that the probabilities of this happening are remote to impossible, particularly if you have used some form of (a) backtesting and (b) particularly in the portfolio scenario.

What would, or should, the 'average' individual trader, trading his own money do in this circumstance?

jog on
duc

In my situation 15% is my maximum stop loss position. So I would close the position immediately.

 

[ducati916]

So I would close the position immediately.

Which is the response that I would expect to see, particularly on a single position basis. Would the same response apply if your entire portfolio contemporaneously, followed suit?

Also, I'm interested, with regard to an alternate thread that discusses 'diversification': is diversification, or concentration, your preferred methodology?

For tech/a

“But everything was not fine. Long-Term, which had calculated with such mathematical certainty that it was unlikely to lose more than $35 million on any single day, had just dropped $553 million — 15 percent of its capital — on that one Friday in August 1998. It had started the year with $4.67 billion. Suddenly, it was down to $2.9 billion. Since April, it had lost more than a third of its equity.”

This where the calculation came from.

jog on
duc