Whats Stopping Retail Dealers or Banks From Brute Forcing a Clients Profitable Stgy?

[murray t turtle]

Quote from illiquid:

Open a second account at a different brokerage.
--------------------------------------


Randomly implement large offsetting positions, especially before big new reports. A couple of 50% blow-ups is always good for shaking them off your trail.

=================
illiquid;
FUNny but true in some markets.

[2]arch-28;
really dont know if this will help you with your bank,
but hopes it does.
Frankly, wish every banker was as honest as my banker Dad,
but some bankers are not!!!!!!!!!!

[2.28]Sounds like , that is in liquid markets i trade;
that could be a concern only on a countertrend,
not a trend system. Your call whether it aplies to your markets.

[3] Some may miss illiquid truth, cause humor is there.

 

[9th Gate]

Quote from Businessman:



Also if i had a rule about when i
wasnt going to trade, this would
be almost impossible to reverse engineer
anyway.

This is a good point too. A good system would include some good entry filter rules that would almost impossible to figure out. The big banks and hedge funds wouldn't bother because most smaller traders are directional traders and this goes against the trade strategy of some of the bigger players anyway. If you were trading through a big investment bank and found a way to create a short term statistical advantage on a scaleable trade plan I have no doubt they would jump all over your concept. In the "When Genius failed" book trading desks were said to have publicly boasted of exploiting knowledge of customer flow. Once they pre-screen you as a successful trader they could simple make a decision to follow you into the market trade by trade without ever having decoded your system.

At the end of the day we no choice anyway, we need the liquidity and margin from somebody. I would never give any strategy to a company that claims it can use it through their API and "watch" your trade system operate.

Good post.

 

[rufus_4000]

Quote from ptunic:


That said, I'm curious if employees at retail brokers are able to look at which retail traders are making money and try to reverse engineer those trades.. does that happen at all? I'm talking about not with the companies' permission, but say a tech who illegally uses their information access (a DBA, Database Administrator, for example) and says they are "debugging" a possible bug, and in the meantime dumps a profitable retail traders' trades to their own laptop for some evening analysis.

On a related topic to this thread, I read another post a while ago saying some Prime Brokers go so far as to not only monitor who the best traders are, but even approach them to see if they want funding to create a hedge fund. The poster was pissed off since he thought his trades were somewhat private and moved his account to another Prime Broker.

Anyhow.. a fascinating subject, thanks for the contributions everyone.

-Taric

I was the person who noted the situation in the second paragraph. The scenario was not a normal one, it was a professional trader (not retail) that made approximately 700% annualized return on 2-3M of capital, trading frequently (a black box), and with a sharpe ratio of > 2.5 (unheard of). Hence attracted the unwanted interest from the prime broker / asset management arm.

This type of situation is generally welcomed by the professional traders (more money never killed anyone). Most asset management firms manage "wrap" or separate accounts, in which they would hire individual traders to "advise" (since the funds are "managed" by the asset manager), and the wrap manager would be paid a percentage of the management fee (typically 10-15% of the wrap fee). It is not un-usual at all. Usually the "wrap" manager would graduate to be a portfolio manager in some cases.

As an aside, I classify professional traders as individuals with their own memberships, registrations, lines to the exchanges, only use the broker / dealer for back-office processing (clearing, financing and settlement), not brokerage (as these professional traders are considered brokers themselves by the regulators).

I know little of "true retail" world, I was never on that side of the ibank. But I would imagine that given that there are literally "millions" of retail accounts, it would be rare for an account that only has <500k in size to attract any type of attention. It is possible that a tech could take some data and do analysis, however, it is rare that a single tech or two have access to all the necessary data (trades, positions, financing, etc), since they are usually held in different data sources.

Rufus

 

[ptunic]

Thanks again rufus.. very interesting.

-Taric

 

[IV_Trader]

Quote from rufus_4000:

I was the person who noted the situation in the second paragraph. The scenario was not a normal one, it was a professional trader (not retail) that made approximately 700% annualized return on 2-3M of capital, trading frequently (a black box), and with a sharpe ratio of > 2.5 (unheard of). Hence attracted the unwanted interest from the prime broker / asset management arm.

This type of situation is generally welcomed by the professional traders (more money never killed anyone). Most asset management firms manage "wrap" or separate accounts, in which they would hire individual traders to "advise" (since the funds are "managed" by the asset manager), and the wrap manager would be paid a percentage of the management fee (typically 10-15% of the wrap fee). It is not un-usual at all. Usually the "wrap" manager would graduate to be a portfolio manager in some cases.

As an aside, I classify professional traders as individuals with their own memberships, registrations, lines to the exchanges, only use the broker / dealer for back-office processing (clearing, financing and settlement), not brokerage (as these professional traders are considered brokers themselves by the regulators).

I know little of "true retail" world, I was never on that side of the ibank. But I would imagine that given that there are literally "millions" of retail accounts, it would be rare for an account that only has <500k in size to attract any type of attention. It is possible that a tech could take some data and do analysis, however, it is rare that a single tech or two have access to all the necessary data (trades, positions, financing, etc), since they are usually held in different data sources.

Rufus

exelent posts , Rufus. I assuming that you referring to traders that trading the same entity over and over againg , like certain FX pair or SPX. But what about retail trader that constanty keep on changing his universe ? Does it make more difficult to run "reverse enginiring" analysys in this case? Here is more details :
1. Trader A is a retail with account with> 2 mil and trades stock's options
2. Trader A had an excellent returns year after year without negative month
3. Trader A takes positions on more that 150 different stocks an year and the universe keep on changing from year to year.
4. Trader A is actually very welcome others to take position in the stocks AFTER he open his.
Can his broker/ bank figure out what he is doing (if they want to) ?
Thanks a lot

 

[rufus_4000]

Quote from IV_Trader:

exelent posts , Rufus. I assuming that you referring to traders that trading the same entity over and over againg , like certain FX pair or SPX. But what about retail trader that constanty keep on changing his universe ? Does it make more difficult to run "reverse enginiring" analysys in this case? Here is more details :
1. Trader A is a retail with account with> 2 mil and trades stock's options
2. Trader A had an excellent returns year after year without negative month
3. Trader A takes positions on more that 150 different stocks an year and the universe keep on changing from year to year.
4. Trader A is actually very welcome others to take position in the stocks AFTER he open his.
Can his broker/ bank figure out what he is doing (if they want to) ?
Thanks a lot

In a sense, the scenario you described is no different than an option oriented small hedge fund. Would the Trader A do option complex? Butterflies? Condors? Complex strategies tend to be a little more problematic since it would require more possible combinations. Let me think as a pure intellectual exercise.

From a pure brute force stand point, I can see some a first iteration algorithm pretty easily. The fact that the 150 stocks changes year to year is not a problem, since there are only about 100,000 listed options (including flex) in any given time, and most of the large enterprise tick historical DB would have all of them in archive. So it is just a matter of re-computing relative variables of each option chain against others, so a possible volatility arb play can be detected, a sector / baskeet rotation play can be detected, a relative skew play can also be detected.

I am not saying that after the first iteration (of 100k x 100k x trades / positions) would yield a correlation model. In fact I don't think it will on the first pass, but then it is the Quants job to cull through the matricies, and setup the parameters for the next iteration.

Keep in mind that the end result may have weird statistical properties, so it is a simulation of the original model, not the model itself. Especially given the Trader A's only changing position year to year, there might not be sufficient data points to run simulation against the real model to further tune correlation.

 

[prt_systems]

Quote from rufus_4000:

....
For instance, NYSE average trade size was around 600-700 shares in '98. Last time I looked up, the average trade size is now 250-300 shares. You might think it is because more retail customers, but from the fact that NYSE volume is 50%+ done by "program trading", almost everybody is employing some type of "slicing" algorithm these days.

Rufus

..and this is all too obvious to anyone that simply watches the market reaction to trades under various market conditions.

Adapting to this fact has been interesting but not difficult - perform enough (low cost) experiments and you really can infer the algorithms and program against them .....

 

[OneHipCat]

There is also the problem of your broker coat-tailing your trades if you're consistantly picking winners. I read it happens frequently that brokers may share a profitable clients move with their prop traders and other clients.

 

[IV_Trader]

Quote from rufus_4000:

Keep in mind that the end result may have weird statistical properties, so it is a simulation of the original model, not the model itself. Especially given the Trader A's only changing position year to year, there might not be sufficient data points to run simulation against the real model to further tune correlation. [/B]

thanks a LOT for your reply , RUFUS. especially helpful is your last paragraph about "end results" , that reinforce my initial thoughts that its not so easy to do. Let me know if I can PM to you sometimes with other questions.

 

[mhashe]

Quote from rufus_4000:

I was the person who noted the situation in the second paragraph. The scenario was not a normal one, it was a professional trader (not retail) that made approximately 700% annualized return on 2-3M of capital, trading frequently (a black box), and with a sharpe ratio of > 2.5 (unheard of). Hence attracted the unwanted interest from the prime broker / asset management arm. .......................................

Rufus

Goes to show that the real market wizards are those we'll never hear about. Wondering if he was a currency trader. Thanks for your informative posts rufus.