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)
I had a few minutes to try a 'matrix optimization' in MultiCharts today, so I thought I would share here. There is a similar capability in TradeStation and MultiOpt (an excellent 3rd party tool for walk-forward analysis and optimizing in TradeStation). The TS version was, uh, difficult, though the MC one appears to be modeled after that one. The main utility of this tool is the ability to find an optimal out-of-sample (OoS) period for optimization. I optimized an instrument and one of my recent systems, just to see what would happen (on the systems I optimize, I usually have set periods based on timeframe for re-optimization.
Here is what the different periods look like:
The less obvious utility for me is to validate that a system will work over a wide variety of in-sample and out-of-sample periods. This follows Perry Kaufman's idea that 'loose pants fit all'. In this case, all periods seemed to work, though my parameters to 'pass' this particular system were pretty low.
Here is what Profit Factor looks like, along with the broader results of the matrix optimization:
The report also provides the parameters and dates (not seen above) for the optimization, so I can just build my code around those values as needed. The Robustness feature allows me to pick the fitness function(s) (e.g. Profit Factor) and the values that are considered passing. All-in-all a nice feature, though I personally try to develop systems that I do not need to optimize. This is an easy to use and fast little utility.
Here is a sneak preview of the Momentum Pinball system I'm working on. There are two variations of this system in the book, Street Smarts, but I am developing and testing the first variation:
The rules are as follows:
If the previous day's 3-period RSI of a one-period rate of change (LBR RSI) is > 70, set a sell short order at the low of the 60 minute opening bar
If the previous day's 3-period RSI of a one-period rate of change (LBR RSI) is < 30, set a buy order at the high of the 60 minute opening bar
Orders expire at the end of the trading session (GTD, expiration = today)
If short order is filled, set stop at High of 60 minute opening bar
If long order is filled, set stop at Low of 60 minute opening bar
If stop is hit either way, 1 additional entry is allowed for the day
Exit within 5 minutes of the close of the session
I deviate from the original rules, as this system was originally designed for ES (e-mini S&P) at a time when overnight trading was minimal or non-existent (1995), at least to my knowledge and understanding. The original rules state: get out if not at a profit; hold overnight if profitable. I observed, using SPY as a proxy for ES, that holding overnight in these times (27 years later) is generally not profitable. I'll test that variation another time, but for now, I'll be satisfied exiting right before the close.
The text also states that we should use a trailing stop, but Connors and Bradford-Raschke are not clear on the structure of that stop.
Anyhow, that's what I'm working on. To be continued...
Just a little interesting algo/systematic trader tidbit to share.... I get daily email updates from the 40in-20out trend-following project. Here is what their hypothetical trend-following portfolio (futures) looks like as of close of business today:
They have done a good year in 7 trading days.
Of course, I've been watching them for a while and even built a system based on those rules (not bad, actually). Some of this will go away when the market turns at some point (such is the nature of trend-following). But wow, talk about some fat tails.
Exactly. They had some spectacular results last year, but gave a lot of it back. Trend-following is really about catching fat tails when they occur.
Small wins turning to losses, big wins turning to losses, big wins turning to small wins – get used to that in Trend Following.
- Jerry Parker, one of the original turtles
@kevinkdog, from your keyboard to God's ears.... today's results:
You almost have to appreciate the symmetry.
I know I do this wrong, but for the hypothetical portfolio I maintain for my site, I never report marked-to-market profits, only closed position profits.
I am working towards the deadline of my next system, which should be published tomorrow. Here is one of the things I do for every system idea: a system process diagram. This system is a little more complex than others, but I figured I would share here: SAT-030 Momentum Pinball 1-Entry
SAT-030 Momentum Pinball 1-Exit
I will post results of my testing tonight or tomorrow.
