Can you reverse engineer a trading strategy by just looking at trades?

[giena32]

Hi Everyone,

I would like to ask a fundamental question. How many of you are afraid that someone by looking at your trades (broker statements) will be able to benefit in any way (reverse engineer your trading strategy)? Some people are afraid to use http://tradinganalyser.com/ (where you need to upload your broker statements) as they think broker statements contain sensitive information. By sensitive information I mean trades, as name and address can be easily removed from the statements.

And I am not talking here about real time front running or trading alongside. By reverse engineering I mean to understand the logic of your strategy and be able to replicate it in real-time.

I created tiny Google forms voting page here:

http://goo.gl/forms/AYEu3pmDHeSyjGzN2

If you would be so nice to express your own view it would be interesting for all of us to see voting statistics/results. Thanks.

 

[Baron]

Anybody who thinks that your strategy can be reversed engineered based solely on brokerage statements is completely delusional and paranoid. It's like thinking someone could build my house exactly as it stands based solely on my Home Depot receipts. Nonsense!

 

[Deuteronomy_24_7]

From a portfolio's daily P/Ls it's easy enough to regress against equity factors such as momentum, size or value to get a sense of the manager's style. Similarly with a futures or forex P/L stream, to see if the manager tends to be a trend-follower, breakout player, contrarian or whatever. Allocators do these analyses all the time.

 

[southall]

Some of the inputs to my systems could be found from my brokerage statements.

Lets say i was an FX day trader and i traded certain news events. I also trade other systems.

But you only looked at my biggest winning days and you quickly noticed they were on news driven days.

You could found out which pairs i traded, which events i liked to trade, and how long i waited before i took a trade, if i prefered breakouts or support/resistance trading.

Not saying anyone could replicate my success by doing this but you could find out a lot about my strategy.

High expectancy trading edges are like gold dust, you dont want to make them public.

 

[stevegee58]

Anybody who thinks that your strategy can be reversed engineered based solely on brokerage statements is completely delusional and paranoid. It's like thinking someone could build my house exactly as it stands based solely on my Home Depot receipts. Nonsense!

On top of this, there's the underlying presumption that you've somehow stumbled onto the Holy Grail. My experience in looking at other peoples' "amazing" strategies is that they're NOT.

There's nothing new under the sun, especially in the trading systems world. There's nothing worth reverse-engineering.

 

[J_Smith]

Hi Everyone,

I would like to ask a fundamental question. How many of you are afraid that someone by looking at your trades (broker statements) will be able to benefit in any way (reverse engineer your trading strategy)? Some people are afraid to use http://tradinganalyser.com/ (where you need to upload your broker statements) as they think broker statements contain sensitive information. By sensitive information I mean trades, as name and address can be easily removed from the statements.

And I am not talking here about real time front running or trading alongside. By reverse engineering I mean to understand the logic of your strategy and be able to replicate it in real-time.

I created tiny Google forms voting page here:

http://goo.gl/forms/AYEu3pmDHeSyjGzN2

If you would be so nice to express your own view it would be interesting for all of us to see voting statistics/results. Thanks.

Of course it is possible if you give the time stamp of the trades - all you have to do is mark the trades on a few different timeframes, and you can see then where the trader got in and got out.

How do you think I cracked it

Most don't have to worry though, as if you are talking about standard text book methods, then you can post as many of those trades as you like!

Those in the know will never post details of their trades anywhere - any trades I have posted here are just messing around trades - I will never ever post the price levels and time stamps of serious trades - too much at risk if the details got in the wrong hands, like H3, who could automate it and make a fortune, to add to his current millions!

Never ever let sensitive information into the public domain, especially on silly trading websites!

J_S

 

[KDASFTG]

Greetings G,

You asked: “How many of you are afraid that someone by looking at your trades (broker statements) will be able to benefit in any way (reverse engineer your trading strategy)?”

Response: In your first question, it depends on what you meant by benefited. Personally, I would not be too concerned if someone was possibly able to learn something from my PnL because, the PnL just reflects the results of the trading. It does not reflect or expose the “Process” by which these trades were generated. In answering your second question, my trading “Process” is a function of essentially two components:

Component #1.Technical Analysis (i.e. the TA Tools and Techniques used to create my Edge).

Component #2. Mental Analysis (i.e. how I think, and what I am thinking while I trade, which allow me to properly execute my TA edge).

In order to successfully reverse engineer my particular style of trading, one would have to first fully understand what I do and why I do it, then 100% replicate my TA, and then also share my “Trader Mindset” which I exercise while I am executing. So therefore, in my opinion, and in consideration of all these things, I believe that the probability of capturing and executing all of this from just looking at my PnL is in the “noise level", statistically speaking.

Thanks for the exercise in thinking.

KDASFTG

 

[J_Smith]

Also, only use these websites to identify people worth talking to - act the fool, like I do, but never be a fool, and never be afraid to tell anyone to "fuk off and don't be annoying me"


J_S

 

[J_Smith]

Greetings G,

You asked: “How many of you are afraid that someone by looking at your trades (broker statements) will be able to benefit in any way (reverse engineer your trading strategy)?”

Response: In your first question, it depends on what you meant by benefited. Personally, I would not be too concerned if someone was possibly able to learn something from my PnL because, the PnL just reflects the results of the trading. It does not reflect or expose the “Process” by which these trades were generated. In answering your second question, my trading “Process” is a function of essentially two components:

Component #1.Technical Analysis (i.e. the TA Tools and Techniques used to create my Edge).

Component #2. Mental Analysis (i.e. how I think, and what I am thinking while I trade, which allow me to properly execute my TA edge).

In order to successfully reverse engineer my particular style of trading, one would have to first fully understand what I do and why I do it, then 100% replicate my TA, and then also share my “Trader Mindset” which I exercise while I am executing. So therefore, in my opinion, and in consideration of all these things, I believe that the probability of capturing and executing all of this from just looking at my PnL is in the “noise level", statistically speaking.

Thanks for the exercise in thinking.

KDASFTG

I agree entirely if you are talking about manual order entry - some people automate to get rid of the emotional issues that can ruin your trading.

The downside of automation is drawdowns - unless you are up there with the super duper multi million $ algo traders, which I would imagine no retail trader is!

So, if you have plenty money, automating will make you money, providing your method works over a set number of trades.

The beauty of short term manual trading, is, you only need do bout 2 hrs per day, then shut down and go bout your business - you are out of the matket with no risk to your capital.

2 hours is not much, so, once you have the best tools, your head in order, there is absolutely no reason whatsoever why you can not make money every day you trade - as you are in full control and following the market, not trying to predict market direction!

J_S

 

[userque]

Hi Everyone,

I would like to ask a fundamental question. How many of you are afraid that someone by looking at your trades (broker statements) will be able to benefit in any way (reverse engineer your trading strategy)? Some people are afraid to use http://tradinganalyser.com/ (where you need to upload your broker statements) as they think broker statements contain sensitive information. By sensitive information I mean trades, as name and address can be easily removed from the statements.

And I am not talking here about real time front running or trading alongside. By reverse engineering I mean to understand the logic of your strategy and be able to replicate it in real-time.

I created tiny Google forms voting page here:

http://goo.gl/forms/AYEu3pmDHeSyjGzN2

If you would be so nice to express your own view it would be interesting for all of us to see voting statistics/results. Thanks.

It would be better for one to use local highs and lows of an instrument as signals to decode as you get much more history to work with.

That said, simple systems can likely be decoded. (MA cross-overs, etc.)

(The data points can feed into an algo designed to brute force over all known indicators and likely parameters using genetic programming. They can also be fed into Eureqa, and/or any NN/machine learning software package. Machine Learning will try to build a model just like it would if you fed it stock data, or any other data.)

Slightly more complex systems can likely be emulated (not decoded) if given enough data.

Extremely complex systems (machine learning) can be emulated if given enough data, but as I've stated, it would make more sense to emulate the actual 'market' than someone's system.

Simply saying a system can be emulated doesn't mean a lot without knowing the "how well" part. How well a system can be emulated will depend upon the skill of the hacker, and the amount of data the hacker has to work with, and of course, the complexity of the system.

Seeing that most systems can't emulate the local highs/lows of a market instrument very well, I suspect most attempts at emulation/decoding will be worthless at worst; as good as the system being emulated as best (still worthless?).

Even so, I suspect there exists talent that can emulate very well given enough data to work with; but such individuals are smart enough not to waste their time doing so since they realize it makes more sense to emulate the markets, rather than someone else's attempt at doing the same.

And the above doesn't even take into account that the brokerage data may include a mixture of different 'systems' that where used by the target over the years.

In short, my answer is, as usual: it depends.