Wednesday, July 28, 2010

The System...

Quick update on the trading system I have been developing. I have now tested all the way through the first half of the year on a 5 minute timeframe.

I have the following statistics to share.

Based on 1000 shares of IWM
Trade costs

Monthly total Weekly avg Net Total trades TD IB
Jan $1,320.00 $330.00 $550.00 220  $  2,197.80  $    770.00
Feb $970.00 $242.50 $200.00 220  $  2,197.80  $    770.00
Mar $2,080.00 $622.50 $1,107.00 278  $  2,777.22  $    973.00
Apr $80.00 $98.00 $1,053.00 278  $  2,777.22  $    973.00
May $1,290.00 $322.50 $408.00 252  $  2,517.48  $    882.00
Jun $160.00 $40.00 $1,210.00 300 $2,997.00 $1,050.00
Totals $5,420.00 $243.25 $2.00 1548  $15,464.52  $ 5,418.00

Max Profit $1,710.00 Best Day $2,130.00 Average Profit $0.28
Max Loss $720.00 Worse Day $1,910.00 Average Loss $0.13
Ratio 237% Ratio 112% Ratio 215%
% Winners 32% Best Week $1,810.00 Total Wins 251
% Losers 66% Worse Week $1,450.00 Total Losers 511
Ratio 49% Ratio 125%

I have added emphasis on the key statistics I feel have the most bearing on the validity of the system.

Prior to that...How about those trading costs? I think traders who try to develop their own systems eventually look at this column and say....hmmm...I think its time to start my own brokerage...what a matter if you are winning or loosing THEY ARE WINNING. Always. Anyway that can be overcome with appropriate trading scale...and the power of compounding.

First the monthly is clear that winning months outpace loosing months by a wide margin. This is very positive and it means that if you take even a crude exit strategy which does not handle chop very well, and does not take profits at any point along the route...this is very good indeed.

Second and more concerning is the percent of the total trades that are profitable vs trades taht are loosers. In this case there are nearly twice as many losers for every winner.

The consolation prize is that the average winner is more than twice the size of the average loser..again speaks loudly to the validity of the system. This gives you enough edge to get those monthly totals...its a razor's edge though.

The system's details shall remain proprietary until I bear out all the testing and see if it works on different timescales and instruments. It is incredibly simple and I am goshdarn proud that something so stupidly simple can work decently well.

My next step in the process of backtesting is to tweak the exit strategy. I have at least 2 ideas for this one involves profit taking the other doesn't. I will probably eventually mash them both together to take profits along the way.

I want to bring the ratio of % profitable vs % losers UP while not sacrificing the average size of the winners vs losers...

If I am able to accomplish that with any of the next 2 ideas...I think I can start a hedge fund...

Especially if I can then take it to different assets and different timeframes...which is also part of the next steps in the process...but basically once I find the proper exit strategy...then computers will take over and help me figure out if it works accross time frames and asset classes.

The backtesting is arduous and I sacrifice many hours of sleep doing it. But I am confident now that the hard work is going to pay off.

The key to testing the first half was to see how well the strategy would do in periods of low volatility and high volatility. It doesn't do as great in high volatility because of the exit strategy. I saw lots of chop eating away profitable trades and taking me out of trades too soon.

So I am confident this strategy works exceptionally well. I just now have to come up with an exit strategy that allows this thing to hang in with the trends it so precisely identifies.

Like I said...don't be surprised if in a few months you read about some guy running around Prospect Park naked Archimedes style. I feel I am on the verge of something incredibly awesome.

No comments:

Post a Comment