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)
Kevin Davey is the Founder and CEO of KJ Trading Systems and will be monitoring this thread so that he may answer any questions that you post here relating to his products or services, primarily focused on algorithmic trading systems.
Please keep in mind that some customer service/technical support issues are best handled through proper channels at KJ Trading Systems.
Kevin Davey is a full-time professional trader and a multiple time winner of the World Cup of Futures Trading Championships (traded live with cash). Kevin Davey also contributes articles to SFO Magazine (Stocks, Futures and Options) as well as Active Trader Magazine.
Kevin Davey was profiled as a "Market Master" in the book "The Universal Principles of Successful Trading" by Brent Penfold. Kevin has also presented multiple webinars on futures.io (formerly BMT) that focus on algorithmic trading, which you can find in our Webinars section.
In addition to this thread, I will also be asking Kevin Davey to stop by on occasion for a casual webinar where he can answer questions via audio while also sharing his screen to visually demonstrate any points as needed. The date/time of those sessions will be announced here in this thread. These sessions will be limited to questions only, there is no prepared presentation. After the session ends, the recording will be posted in this thread.
Feel free to ask any questions below and we'll do our best to get them answered.
The futures.io (formerly BMT) "AMA" (Ask Me Anything) series is by invitation only. It is part of a new program we are launching shortly called "Certified Trustworthy", something that has been months in the making. I will provide all the details of this new program as soon as it is ready for launch.
I'd like to start by asking you how important you think Portfolio Trading is? Do you tend to have a single strategy that trades multiple products, or multiple strategies that trade multiple products?
How much importance do you place on the Portfolio Backtesting aspect of systems development?
Most traders seem to put all of their effort behind building just one trading system, and then they trade just a single product with that system. I prefer to look at multiple systems in multiple markets, and I do my best to ensure the trades they are taking are as uncorrelated as possible. This is not an easy task as the average tools available don't adequately do this job in my experience.
I think Portfolio Trading is the closest thing out there to the "Holy Grail." For people starting out with small accounts, it is really hard to do, so most stick with one strategy and one market. Maybe that is partly why so many small traders lose. They put themselves at the mercy of one market, and one strategy.
When I had success in a trading contest, I was trading the same exact strategy (with different parameters), in multiple markets. Nowadays, I tend to trade different strategies in different markets, although I still have some single strat / multiple markets, too. I think both ways, if done properly, have merit.
So I tend to put a lot of faith in diversifying. Different markets, different timeframes, different trading styles, different macro approaches(futures, options, spread trading, etc), different strategies. Diversifying helps smooth things quite a bit, and takes away the reliance on any one strategy. With multiple strats, it also makes it easier to kill a bad strategy, and it makes it harder to cheat on a strategy.
For analysis, I tend to (surprise!) keep it simple. That is my bias, since in my old career, I used to have statisticians working for me. I remember one asking me "what conclusions do you want me to justify with this data?" Well, I like the data to tell me the answer, not for the data to be manipulated to what I think it should tell me. I think more complicated analysis is easier to manipulate to say what you want it to say.
One example: If you want to see correlation, just plot a simple X vs Y graph, and stand a few feet away from it. If you can see a relationship, then there is likely correlation. If it looks like a shotgun pellet pattern, there probably is no correlation.
Of course, many times you need numbers to tell you things, so I use Excel (and Excel macros) for many things. I also have some other software I have written or found over the years, but I use those only a little. I usually take what I want right from the performance reports, and (lightly) manipulate it in Excel.
Can you give us an idea as to some of the markets you are trading? Do you routinely trade equities, or is it primarily futures?
One of the problems I have is getting data out of NinjaTrader and into Excel, and then doing portfolio analysis. What I mean is that I want to combine say 3 strategies trading 5 markets, and have one big equity curve and one big correlation report. I've not been able to do this.
All my active trading is in futures. Leverage and tax reasons are the biggest reasons. I see my stock trader friends struggle with buy/sells wash sales, etc. I just input 1 number from my future brokerage into my tax software...
My trick to doing portfolio analysis with Tradestation output isn't really even necessary nowadays, since Tradestation now has a portfolio tool called "Portfolio Maestro." I've never used it.
My trick: I output the daily P/L results for all the strategies into Excel. I then add in missing dates (when there were no open trades), and end up with a table of date in one column, and each individual strategy in its own column.
Once it is in that format, it becomes pretty easy to create an overall equity curve, drawdown curve, do daily correlation analysis, and run Monte Carlo sims.
What is the criteria from moving from the incubation to small size live money trading? What if the incubation period has been in draw down? It seems easy to decide to trade real money if the incubation has been profitable.
For example you have a system the you know that the max draw down based on your monte carlo and back tests is 5,000 but your incubation (of 3 months & 25 trades) has had a draw down of 2,000? Is that still safe to take to the real money arena, we still are within statistical parameters? Or are we looking for only profitable incubation periods? Where is the cut off line?
At the start of incubation, I am assuming I have a profitable strategy - I just want to ensure that I made no drastic mistakes. If there are mistakes, I expect them to show up in the 3-6 month incubation period. Note that this is different than if you were trying to use incubation to prove the strategy is good.
So, I am looking for performance during incubation to be "similar" to the walkforward test. There are a few ways to define "similar:"
1. You could use Student T test to see if walkforward and incubation are from the same distribution (meaning are they similar or not)
2. You could use Statistical Process Control to determine if the trade process is in "control." If it is, then the strategy is functioning in incubation as it did in walkforward.
3. Print an equity curve of walkforward and incubation together. If you stand a few feet away, and can tell when incubation started, then that is a problem. If it is tough to distinguish where one ended and the other started, that's good.
So, for your exact question, without more info, I'd have to assume that $2000 incubation dd could be expected in a $5000 dd walkforward case.
BUT, let's say every walkforward drawdown, regardless of size, was 1 week or less in duration. Then, in incubation, you've been in drawdown for 5 months. Even if the drawdown amount was within historical bounds, I'd still be scared by the duration being "abnormal."
In cases where I am not sure if incubation "passes" I'll typically let it run a few more months, maybe 9-12 months.
I was wondering if you found it easier to build systems on larger time frames, say daily? or do you have a time frame you like to focus on. I know a lot of books and developers out there use the daily time frame, as well as the system you gave out in the webinar. What are your thoughts on shorter time frames say 1 hour?
Great question. There are really 2 questions here: 1) what is a good timeframe to use for the bar charts, and 2) what is a good holding period for trades.
As far as timeframes go, I am all over the place with various timeframes, both common and uncommon. Anywhere I feel I can get an edge. I do some things that many people would consider odd. For example, I use a 1 minute chart for one strategy, yet hold the trade for 480 bars (minutes) typically. I use 60 minute bars for another strat, yet I hold the trade for days or weeks.
The key, at least for me, is in the holding time. I would absolutely love a strategy that held for very short periods of time. Intraday would be perfect. And that is usually where I start when I create a new system.
Unfortunately, the "best" system (the way I define it) usually ends up requiring a longer holding time, days to weeks many times. This seems to get above the noise level in the price signal.
All that being said, lately I have started to focus more of my effort towards intraday systems, short holding times and 60 minute or under bars. It is harder than creating a daily bar system, but I suspect it will get easier with time.