Aaron chat

[PhoenixTrader]

Aaron,

one focus question, was the market chat meant for purposes of doing like you do?, using your product offering or having you manage trading accounts?

 

[Aaron]

Quote from PhoenixTrader:

Aaron,

one focus question, was the market chat meant for purposes of doing like you do?, using your product offering or having you manage trading accounts?

Not sure what product you are referring to. I don't sell anything that would help someone create or manage a hedge fund and I don't consult on hedge fund creation. (Hmmm. I would consider it if ever approached, though.) The only thing Schindler Trading "sells" is money management.

FFastTrade sponsored the chat and asked me, as a satisfied customer, to speak on a topic that might be of interest to the Elite Trader community. I am always happy to recommend FFastTrade to high volume traders of electronically traded futures. The chat also gave Schindler Trading some visibility and I hope you'll consider us as a potential addition to your portfolio.

 

[DisciplinedHedg]

I was reading your website:

>>>One way to help protect against that is to have an in-sample data set and an out-of-sample data set. You backtest your strategies always on your in-sample data. You backtest it, you optimize it, and then when you’re all set and you’re ready to trade it, you test your system or signal on your out-of-sample data set. My final go or no-go decision is based on the out-of-sample test. If my test strategy is profitable on the out-of-sample data with realistic slippage and commission applied, then I’ll start trading it on a small level. <<<

I thought about this myself in the past. But realized, or so I think, that this is STILL backtesting. Aren't you still attempting to optimize your system based on historical data. Whether it is in-sample or out-of-sample, once the final system is in place, all the data becomes in-sample since your system becomes optimized for those historical periods of time. Granted the system was modified only on the original in-sample data, but once you have the final system, you've still created a system where it "fits" your historical data, both in and out samples.

And as a result, have you really escaped the problem to begin with? I've pondered this quite often. What are your thoughs Aaron?

 

[Aaron]

We optimize our portfolio on the insample data. Once the strategy is all ready to implement -- and it is rare that a new strategy is created and looks to be statistically profitable -- we do a final check on the out of sample data. The check is primarily to get non-curve-fit estimates of the profitability and drawdowns of the system.

But if the system were unprofitable out-of-sample, then we wouldn't trade it (probably). Now you could say we are curve-fitting a binary "trade/don't trade" parameter. And you would be correct.

Yes, this is still backtesting and the system created still "fits" the historical data. But, unfortunately, historical data is the only kind there is.

 

[DisciplinedHedg]

Quote from Aaron:

Yes, this is still backtesting and the system created still "fits" the historical data. But, unfortunately, historical data is the only kind there is.

Thanks for the reply...I guess that is the problem. In creating a system, what more can you do than optimize it and fine tune it as "real time" data comes in. Otherwise, someone by now would have found the holy grail...

 

[praetorian2]

Aaron- Thankyou again for doing the chat and answering all my questions about hedge fund setup and trading.


Is there any reason why you chose the 4.6% standard deviation as opposed to some other number?

Have you ever thought about adding equities to your program?

I bet that if you did a program that looked for stocks hitting new all time highs, and used a simple stop loss and trailing stop loss after a certain point you'd have a very profitable system that could be very mechanical for you.

 

[Aaron]

Aaron- Thankyou again for doing the chat and answering all my questions about hedge fund setup and trading.

My pleasure.

Is there any reason why you chose the 4.6% standard deviation as opposed to some other number?

4.6% daily std is big enough to be aggressive and to try to get as much juice out of our strategies as possible, but it is small enough such that we are unlikely to go bankrupt.

Have you ever thought about adding equities to your program?

Yes. And then decided against it due to the complexity of allocating realized and unrealized capital gains to the partners in the Schindler Fund. Fortunately, futures are marked-to-market nightly so there are no unrealized capital gain issues.

 

[EliteThink]

Aaron, thanks for all the great input. If I'm understanding correctly you have been able to become part of an emerging cta fund and have institutional money under management. With a relatively young fund, how did this come about?

 

[Aaron]

Quote from EliteThink:

Aaron, thanks for all the great input. If I'm understanding correctly you have been able to become part of an emerging cta fund and have institutional money under management. With a relatively young fund, how did this come about?

My pleasure! I hope some ET members will put together funds and be successful with them.

The CPO heard about Schindler Trading from one of the database websites and called us. We visited them and a few days later they hired us.

 

[OPTIONAL777]

Quote from Aaron:



My pleasure! I hope some ET members will put together funds and be successful with them.

The CPO heard about Schindler Trading from one of the database websites and called us. We visited them and a few days later they hired us.

Looking at your fund results, you are down 30.9% for the year!!!

Man, that's a bad drawdown, especially for those who went into the fund at the end of last year.

Three losing months in a row, that sucks even worse.

You realize that you have to make over 44% for investors who came in at the end of last year just to BREAK EVEN????

Now, if you have a 20% month next month, do you take your 20% profit from that 20% month? And are your also including fees and commissions in your performance table?

Dude, you had better turn this ship around quickly.