IB Trading Olympiad inspired me

[Maksim79]

Hi, all.

IB's trading Olympiad inspired me. I have opened an account and coded a few strategies using the Java API. I found it to be extremely difficult to create a winning trading strategy.

I understand how financial instruments work and have programming experience with perl/java. I have spent countless hours parsing and disecting tick-by-tick market data for forex and some index futures. The trading stratigies that I have come up can be divided into following categories:
a) loosing strategies
b) strategies that barely make money (Russell 2000 emini - about 100$/day) - since the edge is so small, I just attribite it to "luck", as blindly going long in a bullish market.
c) strategies that are profitable virtually everyday, but extremely high frequency. In that case commissions (even IB's super low commissions) negate the profits.
d) profitable strategies that could be profitable, if you can predict price range.


My current work flow is as follows:
1. come up with a theory
2. write a perl simulator to test on previously recorded market data
3. code up in Java, if necessary.

Trading Olympiad results clearly show that fully automated trading is very real. My question to all is this - do you think it's possible to create a decent trading strategy, without in-depth understanding of all aspects of FE? I would be very interested in critique and exchange of ideas on the subject.

Feel free to e-mail or post a reply!

 

[RussianMike]

If it was easy everyone would win . How much finance experience do you have? I dont know what the F/X commissions are but perhaps you can arb that(maybe not b/c I dont know what the portfolio size they give you to play with)

 

[Andy Nguyen]

Trading Olympiad results clearly show that fully automated trading is very real.

Not that clearly
If by fully automated trading system used in these IB Olympiad, you mean a turn-key system where the contestants turn it on the first day and never touched/twisted it during the entire time, then I'm very doubtful.

My feeling was that to win big in these kind of contest, you either have to bet ALL-IN in a direction or constantly twisting it, hence defeats the definition of a fully automated trading system. To gain some profit is a whole different story.

In either cases, the top winners are either extremely lucky or very knowledgable of the market and how the system works. Gus, one of our members here and the second place winner of 07 IB Olympiad is a testament of the later.

I would rather be lucky :D

I'd love to hear what you theory you come up with and which are the components of your current system.

 

[Maksim79]

If it was easy everyone would win . How much finance experience do you have? I dont know what the F/X commissions are but perhaps you can arb that(maybe not b/c I dont know what the portfolio size they give you to play with)

Mike, I'm not saying it was easy - I'm saying it's doable. As for my experience, I have supported electronic trading for for about 5 years now. All trades had to be entered through an API and the decissions about trading had to be made by an algorithm. I love the concept of arbitrage, unfortunately, arbitrage opportunities are hard to find.
I have created an algorithm doing triangular arbitrage on 10 currency pairs, but the risk/reward ratio is too great for it to make any decent money.

 

[Maksim79]

I'd love to hear what you theory you come up with and which are the components of your current system.

Where do i start ...

I had multiple theories. Here's one:

if you place limit order on both sides of the market (like market makers), and have stops on both sides, X * spread, the probability of both limits executing (thus pocketing the spread) is higher than executing a limit order on one side and getting stopped out.

The problem is the following: If the spread between limit orders is wide enough, there's almost a 50/50 chance of making and loosing money. On the other hand, if you reduce the spread to just one tick, the probability gets much higher, you would make tons of money if it wasn't for commissions.

Feel free to critique this one

 

[Daniel wang]

hi

what kind of algorithm do you use? do you mind if i take look at your algorithm? thx

 

[Adam G]

I also started up my own IB account two years ago after the contest. I primarily trade options and leveraged ETFs. FAS, FAZ, and DXO have been good to trade. What you might want to look at doing is intermarket analysis. There are a few books out there on this to start you out. I'm working on some intermarket strategies because I have a day job and can't really trade all the time. I worked on developing my own arbitrage strategies two years ago, but they really didn't make that much money. I stopped working on the arbitrage once I was making consistent and larger returns with options.

If you look at doing intermarket analysis you might want to check out Ninjatrader to back test your algorithms. If you feel real lazy you can even have it implement the trading strategy via the TWS API.