Ask HN: Has anyone used machine learning to trade? • Hacker News Board

Re: Ask HN: Has anyone used machine learning to trade?

#61

Post by dpapathanasiou » Mon, Nov 23, 2009, 6:04 PM UTC

schemer wrote: post #59

dpapathanasiou wrote: post #58

Earlier quoted context omitted.

Half your momentum trades are profitable? Is that before or after trading costs? How often and at what volumes are you trading?

All I can say is that 8/10 are losers is highly improbable. Given no other information, if you enter a trade there is a 50/50 chance that the next tick is going to be up or down. But notice I said given no other information, there is much more information from the behavior of the charts. If you are a small player the charts will help you be right more than 50% of the time. The rest is how your algorithms and trading…

That is a gross oversimplification.

It also seems you've backed off your claim that 50% of your trades are profitable (if you actually trade).

Re: Ask HN: Has anyone used machine learning to trade?

#62

Post by solutionyogi » Mon, Nov 23, 2009, 10:40 PM UTC

If you know .NET, Do have a look at NinjaTrader and their forums. You will find quite a few folks who are in to automated trading.

I ventured into retail trading (manual) and lost around 5$. Email me if you are interested, I can give you more details (email address available in my profile).

Re: Ask HN: Has anyone used machine learning to trade?

#63

Post by schemer » Tue, Nov 24, 2009, 3:19 AM UTC

dpapathanasiou wrote: post #61

schemer wrote: post #59

Earlier quoted context omitted.

All I can say is that 8/10 are losers is highly improbable. Given no other information, if you enter a trade there is a 50/50 chance that the next tick is going to be up or down. But notice I said given no other information, there is much more information from the behavior of the charts. If you are a small player the charts will help you be right more than 50% of the time. The rest is how your algorithms and trading…

That is a gross oversimplification. It also seems you've backed off your claim that 50% of your trades are profitable (if you actually trade).

Here is something a little bit more complicated. 70-80 % win ratio. Please stop with the 8/10 losses misinformation.

http://www.tradingtheodds.com/strategy-performance/

Re: Ask HN: Has anyone used machine learning to trade?

#64

Post by crpatino » Tue, Nov 24, 2009, 10:34 PM UTC

steveplace wrote: post #49

smanek wrote: post #42

Don't bother. D.E. Shaw, Renaissance Tech, Goldman Sachs, etc. have already hired hundreds (if not thousands) of people smarter than you to do exactly this. They have more Math, Physics, Econ, CS PhDs than you can count/

They have more Math, Physics, Econ, CS PhDs than you can count/ So did BSC, AIG, LEH, and MER

More to the point! The big disadvantage ML methods is that they are just fancy statistics. The main faulty assumption here is that they believe the future looks a lot like the past, so they try to find even more complex patterns in historical data to extrapolate future trends slightly faster than everybody else.

In order for this approach to work, you need very precise market conditions. When markets are stable, your edge is zero and you would be better off investing in some sort hedged index (minimum effort, decent result). When markets are not that stable, it is when you can anticipate trends and get a decent profit for your effort. However, when markets suffer big unexpected turns is when fortunes change hands... and it is when your algorithms give you negative advantage! While you are warrantied to keep following the numbers wherever they take you,a savvy finance human agent has a chance to smell the blood and discretely head for the exit before everyone else. So, it is your call guys. In my opinion, this will not be a viable approach until we get some type of semantic-web-ly mine-able news feeder.