Wednesday, 19 June 2013

Automated Trading

We have been actively trading Flo for a long time now have a lot of experience and statistics available.

It is my strongest belief that as the markets have changed and continue to evolve with technology, the way for traders to now succeed (succeed = to make consistent profits) is to use rule based, back and forward tested, methodology.

Having interacted and/or mentored with hundreds of traders, I believe that traders fail because of a lack of a properly tested trading plan or a failure to follow a trading plan.

The medicine for this, I believe, is to trade using an algo. The very act of properly creating, testing and using an algo means that there is a properly created and tested trading plan which is adhered to.

Trading is a business of probabilities. When we see "A" there is a predominance of probability that "B" happens. We then wrap money management around this and we have consistent probability.

The least important part of all this is what "A" is. There are millions of "A"s that result in high probability "B"s. It is NOT difficult to create a consistently profitable algo.

What is more difficult is to create a consistently profitable algo with the average trade size and drawdowns that do not cause the trader to turn the algo off. That means the trader needs to find an "A" that fits his trading DNA.

The basis of finding an answer to this question is determining a couple of basics:
  • Are you an inside out or outside in trader or both
  • What size of swings/oscillations do you want to trade taking into account that the markets moves in thrusts and pullbacks during the trade
Algo trading and HFT trading is increasing exponentially, not just because of the speed required but also because a single trader can trade a"portfolio" of markets thereby smoothing out his equity curve.

While HFT implementing HFT is beyond the pockets of most retail traders, MFT (Medium Frequency Trading) is definitely not. MFT involves scalping the market with automation. Why MFT?
  1. Very high win rate due to small profit targets
  2. Reduced risk as the trader is in a trade for a short time albeit many times a day
  3. Triggers on easily identified bursts of order flow and momentum
  4. Using an algo means the trader needs to be smart once - creating the algo -not smart for every entry and exit.
  5. Your trading plan can be both back tested and forward walked so that you can be as certain as possible that Consistent Profitability (CP) is there, all before any money is risked
The methodology is still the same: inside out trades identify the trend and trade the pullbacks in the direction of the trend. Outside in trades identify extremes and fade moves to revert to the mean or better.

My tool of choice has been MultiCharts but there is now something new for someone who cannot or does not want to get involved with programming. This something new is BLOODHOUND, an add-in to Ninjatrader. BLOODHOUND is a tool box of logic and solver modules that can be picked and clicked and put together like a Lego model to create both signals and strategies that can be fully or hybrid autotraded. I've been working on porting Flo over to BLOODHOUND since the end of May and the results are looking pretty good. I'll post more on this in the future. In the mean time, click the BLOODHOUND link and have a look at their videos.


3 comments:

  1. Good to hear your voice EL. Ta for nudge re: Bloodhound...will check it out.

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
  3. hi El, glad to hear from you. will surely take a look at bloodhound. how is your book on MP progressing? rgds from Switzerland

    ReplyDelete