Can terminology hide a multitude of sins.

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
dave3076
Roundtable Fellow
Roundtable Fellow
Posts: 87
Joined: Thu Dec 01, 2005 10:51 am
Location: London

Can terminology hide a multitude of sins.

Post by dave3076 »

If i created a supersystem that was all singing all dancing, ticked all the right boxes AND more importantly i was widely acknowledged as being a smart developer of systems, and i marketed something as groundbreaking and revolutionary and after lets say....2 years of live trading i decided, or it became apparent that the system needed its parameters "tweeked" or "shifted" to re-align to "current market conditions"....wouldn't it just be quicker for me to say i'd curve fit a system?. Wouldn't i be better putting my hands in the air and saying i was unaware i'd data mined or curve fitted and that i'm now going back to the drawing board. Or perhaps once you realise a status as a "smart" developer of systems this graduates you into a league where the terminology you use to basically admit you've curve fit or made a substantial oversight to the extent that parameters or rules needed to be changed, can smooth the heinous error over?.

If you have system A that has basically not changed much in performance for lets say 30 years. All the different conditions and changes the markets have made over those years, all the volatility spikes and contractions, all the black swans, and System A withstood the test of time. BUT it would be unmarketable to the masses because of the rough nature of the ride! The rules were extremely simple and more importantly LOOSE to account for all the changes the market had or may make, and that the system could still do it's job, but that you understand this is the nature of markets, to continually evolve and that if you want longevity in a specific system it needs to be loose enough for it to do its job BUT also to let the the market(evolve and re-evolve). And then you have System B, thats all singing all dancing but continually has to be tweeked or shifted. Doesn't the fact that system A exists inherently make system B curve fit?, to whatever degree!
Maybe some of you keep getting the same emails!.... (dunno where these people get my email from).
RedRock
Roundtable Knight
Roundtable Knight
Posts: 944
Joined: Fri Jan 30, 2004 3:54 pm
Location: Arizona

Re: Can terminology hide a multitude of sins.

Post by RedRock »

dave3076 wrote:If i created a supersystem
Maybe some of you keep getting the same emails!.... (dunno where these people get my email from).
Slippery brokers and system vendors purchase lists of "Opportunity seekers". Folks who are looking for the holy grail or have a demonstrated history of inquiring on offers for information relating to trading and other possible quick money making ventures. Most of these people are presumed to be a bit naive and open to the convincing sales pitch offering HOPE of a better way of living. Once you are on such a list... good luck. Best to ditch your email addy or keep it as your spam account when requesting info etc..

There will always be suckers, and salesmen to serve their needs... :roll: An optimist might consider these salespeople to be "educators"...
dave3076
Roundtable Fellow
Roundtable Fellow
Posts: 87
Joined: Thu Dec 01, 2005 10:51 am
Location: London

Post by dave3076 »

There does seem to be an uncanny correlation between me applying for mortgages and emails informing me i recently had a distant relative in nigeria pass away in a car crash and that they've left me $2,000,000,000. All i need do is send my details!. Maybe they traded system B for the 10 years preceeding their touching, tragic demise :)

My actual point was that it does seem that if you are widely accredited as being an "expert" or "guru", that what plain out seems to be some form of curve fitting, can be explained away and a quick change of terminology to "adjusting" the parameters, or portfolio etc etc etc because basically the system died or suffered an extenuated setback/drawdown can smooth it all over ready for the next holy grail that's about to be served up. Where as if a newbie or non accredited trader/system developer made the same mistake, you'd simply assume the worst. I still feel its a valid question.... if system A exists, does it inherently make any system that needs to be adjusted, (assuming adjustment wasn't actually built into the system at development stage), curve fit, to whatever degree? And shouldn't we focus more on our ability to understand the nature of system A'(s) and our own psychological capability to follow them. I always thought that a great saying was that good, well researched systems never die. All that happens is their risk profile fluctuates. Yes it may fluctuate so negatively, to a degree that it becomes un-tradeable, but it hasn't actually "died". Anyway, apologies for clearing my chest on the forum.... but after the 7th email in one day, i had to let off steam :evil:
RedRock
Roundtable Knight
Roundtable Knight
Posts: 944
Joined: Fri Jan 30, 2004 3:54 pm
Location: Arizona

Post by RedRock »

Damn..... I agree. Its a jungle out there. Glad we have a refuge like this ... And a platform like TBB to help.


Caveat emptor



Perhaps Bernie Madoff found the solution. Or not... rocks usually come out of trouser pockets in the wash.

I recall a forex system vender in the 80s. Full page color 'space age' ads promising 100+% returns with a 5k account. Funny. They had a new version each year and prior versions forward performance was unavailable. These outfits pop up every few years.
Paul King
Roundtable Knight
Roundtable Knight
Posts: 207
Joined: Mon Feb 23, 2004 9:13 am
Location: Vermont, USA
Contact:

Post by Paul King »

It is a sad fact in trading that the best looking equity curves are usually a contrary indicator of what will happen in the future, or what you should invest in or trade.

If it's a simulation, then it's curve fit or includes "postdictive" errors, if it's actual performance then it's fraudulent (i.e. made up), illegal (e.g. insider trading, front-running, etc.) or what I call "deferred risk" like a ticking timebomb just waiting to explode (e.g. LTCM)

Fortunately the first case (curve-fit historical simulation) is relatively easy to avoid if you have a sound system development process and reasonable expectations for what "good" looks like.

Paul
alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Post by alp »

dave3076 wrote:And shouldn't we focus more on our ability to understand the nature of system A'(s) and our own psychological capability to follow them. I always thought that a great saying was that good, well researched systems never die. All that happens is their risk profile fluctuates. Yes it may fluctuate so negatively, to a degree that it becomes un-tradeable, but it hasn't actually "died".
I think so. It's how I work, although, through my few experiments and research I have found that the methods we 'rely' on, are actually very imprecise and have very little to tell us about the unknown future, even though they might indicate us what we might probably expect.

But, of course, if we crave for certainty and for stability. And if deep down we all crave for stability, then I argue that we are all conservationists. If we are all conservationists, then we are more likely to feel comfortable with the traditional. And if we have realistic expectations (maturity), we're more likely to stick with tradition. But sometimes, we'll try something new, in the hope there might be a different way of doing things, something better, something new. And, perhaps, we might end up understanding better the meaning of tradition.

Life goes on. As the saying goes: "Everything is changing, but nothing changes. Still, there are changes."
dave3076
Roundtable Fellow
Roundtable Fellow
Posts: 87
Joined: Thu Dec 01, 2005 10:51 am
Location: London

Post by dave3076 »

Im sure it was dwight Eisenhower who said "Things are more like they are today than they have ever been before".

My thinking is there are certainly certainties in the marketplace. As long as there are prices on freely traded indices, stocks, commodities, bonds, indexes and banana's, there will always be point A in a price series, and there will always be point B in a price series. Your point A and my point A and your point B and my point B will almost certainly differ. The nature of how price gets from point A to point B now, how it did it 10 years ago, and how it will do it in 10 years time is certainly uncertain, but it WILL. I've found that if i find myself looking at anything that largely accumulates it's profit(s) from HOW price travels from point A to B, then i should expect the volatility in terms of the risk involved in trading it to fluctuate much more so than something that profits from the certainty that markets WILL always travel from point A to point B, irrelevant of how good "HOW" systems/results look from a historical perspective. Ive also found that the more i try to make "HOW" irrelevant to results, the more robust my system and thinking gets. And the more i think about what is "HOW" and examine "HOW" how can be thought of in terms of it's employment in portfolio selection, entries and exits etc, the more i see that i should stay as far away from "HOW" as possible, and concentrate solely on "WILL" and tolerate all the negatives that trading "WILL" brings, and prepare myself psychologically for that task.
RedRock
Roundtable Knight
Roundtable Knight
Posts: 944
Joined: Fri Jan 30, 2004 3:54 pm
Location: Arizona

Post by RedRock »

J. Simmons
T. Crabel

"HOW" may have some small role in their performance?

They are exceptional exceptions of course.
dave3076
Roundtable Fellow
Roundtable Fellow
Posts: 87
Joined: Thu Dec 01, 2005 10:51 am
Location: London

Post by dave3076 »

Certainly exceptional people! Which i am certainly not!. Well my mum thinks i am but she's biased :D

That aside i've tried to comprehend this game within my own limited capabilities, and understand what i can get my head around and what i can't. There will always be those that can see things that others just can't see, those that have natural affinity, those that are just plain out superb at what they do, many of which voice opinions and give invaluable insights on here, but from my limited experience, natural affinity for trading, computer programming and mathematical skills, i can find no logic in looking for any form of constants within variables! I'm not skilled enough and fear i never will be! But i know there is more variance in "HOW" than there is in "WILL". There's variance in both, but much more so in the former. So i'll stay away from any parameter, dynamic portfolio selector, entry, exit, whistles and bells based on "HOW". And if my limited abilities at arriving at decisions based on hard data steer me from profit, i'll revert to instinct on this one and suffer the consequences of having the courage of conviction.
RedRock
Roundtable Knight
Roundtable Knight
Posts: 944
Joined: Fri Jan 30, 2004 3:54 pm
Location: Arizona

Post by RedRock »

Many say the single most important aspect in this game, is finding your own game. And all the things you mention, will come into focus in the goodness of time. Its a journey of discovery and learning about what is within as well as outside. With persistance and a bit of luck, anything is possible.
Post Reply