Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Auto trading is only thing that conquered my darkside
I must caution you though, that if your backtesting is inherently flawed, all of this will be for nothing.
I see you went back to daily bars. Just bear in mind that if you're not backtesting tick/tick (or with market replay as the Ninjas call it) then again, if your transaction can ocurr all within the same bar, the engine has no way of knowing what ocurred inside that bar and what direction the bar moved first, in the middle, etc.
All of this optimization and what not has to assume there are no fatal flaws in your backtesting method and ability to translate forward.
At a bare minimum you should do a sanity check and forward test for some sample size, just to verify everything looks as good going forward as it did backward.
"A dumb man never learns. A smart man learns from his own failure and success. But a wise man learns from the failure and success of others."
Can you help answer these questions from other members on NexusFi?
Im using OpenQuant (trial) as it supports portfolio rebalancing and other features that I want to use. It doesn't unfortunately do sophisticated analysis of optimization parameters (although I am writing code to spit optimization out to file so I can view the results in Excel). It does do robustness as you can create a Monte Carlo instrument (simulated stock price) using the existing instrument drift (trend), volatility etc... Not sure how this works but I will certainly be investigating it.
I must second RM99 and let you know that your result might be flawed because your take profit of 28 pips seems really small compared to an average daily range of 110pips for EUR/USD.
If you consider your signal to be more reliable based on the daily time frame, take your signals based on that but manage and test your stratey on a lower time frame. You can do that quite easily in MultiCharts, don't know about Open Quant though.
I suggest you create your own journal thread on the forum to track everything. And more importantly, at the end of every week, go back and read all the journal posts for the prior two weeks. Hopefully it will help you avoid what I call the black hole phenomenon, where traders repeat the same mistakes over and over again without realizing it.
Here is an excellent example of a systems "journal". The author is publishing all of his work and code, but the important lessons are what happens over time and how he and the system are evolving:
With about 4 months of steep learning in the pocket I’m going to share my first Strategy. I called the Strategy ‘Myst’ after the adventure PC game from the nineties that I enjoyed so much. My goal was to make a Strategy …
In the lab for a few days, this is what I've come up with.
A trend/scalp hybrid strategy - basically trades the trend but also takes profit at 2 points (using ATR - volatillity estimate - for profit targets).
The entry is fixed but exit I've optimised based on ATR window and multiplier to take profit at 2 points. Thats it. I've tested on a basket of stocks, ETFs and currencies (need more data to test more instruments), including commissions and slippage. max drawdown is around 15% in the credit crunch. Its a long only strategy.
Next up, opening a futures account & paper trading
Eventually you want will want to move into strategies that are adaptable both horizontally and vertically on the charts. Do that and the backtesting becomes less important. ATR is crude and is a start.
A thread about techniques or methodologies of an automated strategy does not belong in this section, but rather in one of the Elite Automated Trading sections of the forum.
... So thanks to rigourous simulation & backtest and a nice pretty equity curve, I now feel confident to pull the trigger. Whereas If I'm left to make the decisions I'll bottle it and make a right pigs ear of it
Some stuff I was messing with yesterday. Pertaining to your comment above, I noticed a similar phenomenon. I was using an exotic bar type, and doing some crude backtesting with NT7. The fills were unlikely, VERIFIED UNREALISTIC with a replay.
I guess the cool thing was how consistent the system was. No matter the size of the bar, chart resolution or the period in which I tested (tested 3 months at a time for the whole year of 2011). All equity curves had almost exact same shape, winning % and everything remained consistent. Increasing the bar size would linearly increase the expectancy per trade.
I believe the pics below where of 6E.... oil was really insane. I deleted most of the stats, here where just two examples using a really small bar size with 1 contract.
