- Joined
- 28 December 2013
- Posts
- 6,392
- Reactions
- 24,319
It pays to read what others think
Came across below video - Is there any merit to primarily using CCI Indicator for determining trend along with entry/exit ?
Is there any merit to primarily using CCI Indicator for determining trend along with entry/exit ?
Hi Skate,Howard Bandy back in Dec 2016 made these comments
“Stocks have a tendency to revert to the mean and breakouts usually do not last very long. I estimate the risk of drawdown based on recent performance, then trade in such a way that I can manage drawdown. That is, holding a few days at most whether long or short. Again and again. The sweet spot is high accuracy and short holding period. Holding longer than about three days increases the risk considerably, as does accuracy below about 65%”
A short holding period is not for me
Trading over a period of a day or three is easier said than done. I prefer to trade weekly as the joy of trading a daily strategy has eluded me. It's not for the lack of trying it's more to do with the constant workload which is not for me.
Howard Bandy trades short term
Howard Bandy has confirmed that he trades this method successfully. Howard knows what he talking about, he's smarter than the average bunny. Does any member currently trade this way who would like to share their experience?
Skate.
Just want to say it is a VERY interesting post.Hi Skate,
I trade in this exact fashion, it is most definately not for the feint of heart. I would love to say it has been a raging success, but I've managed to turn on pretty much all my strategies at the precise moment the broader market takes a dump and so I cannot report any major success. In fact I think people could make a fortune if they just short the markets when I decide to get in...
For this type of strategy I use Howard's book Quantitative Technical Analysis as a guide. I've played with Machine Learning as well as Traditional systems and TBH the best one I've found to date is a simple RSI strategy.
I use python to search all tickers, for each ticker run a walk-forward analysis of the strategy, 2 years inSample, 1yr outOfSample. Generate a report for that analysis. Report contains an image of the equity curve as well as basic metrics Accuracy, W/L but most importantly CAR25 and safeF for my given risk statement.
For each ticker the python script also spits out the parameters for each outOfSample run in amibroker AFL, tradestation easyLanguage and metatrader MQL4 and 5.
I search through the reports manually (a report is only generated if it is at least profitable - reduces number of reports). One of the best I have found was MSFT, I've been trading it for a year or two now with this basic RSI strategy and it has worked well to date.
In order to run a strategy like this I think automation is a must and thus I use either Metatrader or more recently TradeStation.
View attachment 143486
I have a system which trades specific US stocks such as MSFT or VISA etc. The system trades LONG only, holds positions for 1-10 days and is mean reverting in nature. It will trade between 150-250 positions per year. Positions must be taken at the close of the market when the signal for entry or exit is calculated. Currently I trade the system live using Admiral Markets Metatrader 5 platform trading CFD's on the stocks. Automation is important because of my irregular day job. Like all mean reversion strategies the system is highly susceptible to commission drag. Any suggestions or recommendations or past experence with IB or TradeStation would be greatly appreciated as this particular problem is doing my head in...
TBH the best one I've found to date is a simple RSI strategy
Just want to say it is a VERY interesting post.
Thank you
Just want to say it is a VERY interesting post.
Thank you
I gave weekly RSI trading away when I backtested it over dot com and 2008, it needs an exit mechanism to get you out in those massive dips, but I couldnt find one... stoplosses do not work with mean reversion, at least in my experience.@Willzy thank you for a great thought-provoking post. I remember you making a post back in December 2020 referencing this type of trading. At the time I had no input as you mentioned software that I was not familiar with, using Amibroker exclusively in my trading.
Howard Bandy's comments
I spent more time than I can remember trying to code a system similar to what Howard was suggesting. I have traded a "Mean Reversion Strategy" with limited success & I found the same as you that "the commission drag" was enormous. Admittedly, stocks do have a tendency to revert to the mean. When Howard implied that breakouts usually don't last very long. I personally had trouble reconciling this "as fact" because it was at odds with my research & experience.
Howard is very experienced
I don't doubt Howard when he said holding only a few days, again and again, is profitable for him but that's his "Modus Operandi" (M.O) or in English "his way of doing things". His statement is at odds with my research of holding for about three days, he even goes on to remark that holding positions beyond three days increases the risk considerably, & accuracy goes out the window.
A short holding period is not for me
Trading over a period of a day or three is easier said than done. I prefer to trade weekly as the "joy of trading" a daily trend strategy or a mean reversion strategy has eluded me. It's not for the lack of trying it's more to do with the constant workload which is not for me.
Now you are talking my language
I have traded a "RSI Weekly Strategy" with success. It was around (2016 / 2018) that I was trading every profitable strategy I coded up. The last quarter of 2018 brought the house of cards crashing down, taking a complete day to exit all my open positions, it was an exhausting exercise that I never want to repeat. At the moment I trade a fist full of strategies that I find manageable.
When you mentioned a "RSI Strategy" it got me thinking
I was wondering if I was still trading this strategy these last two years (the last 730days) would it be still profitable? So there is no cherry-picking the results, I decided to do a backtest from 30/6/2020 to 30/6/2020 inclusive. The last two years haven't displayed consistent returns but have been consistently unkind at times.
I thought it would be a good exercise to post the results
Well on face value the results for this period aren't too shabby. But that period missed the COVID-19 Flash Crash, so I extend the backtest an extra 6 months to take in this period. Taking the COVID period into consideration the overall results become average at best.
View attachment 143495
Skate.
What a great quote
I have to agree with @BlindSquirrel "giving back profits - sucks"
Most new traders probably also suffer from a dose of over-optimism and Dunning-Kruger effect.....
This post will not be relevant to traders that trade on fundamentals alone. Traders that use fundamentals for stock selection and then use technicals to trade may have encountered some of the problems that I’ve encountered in the past.
I’m basically a technical trader, but I like to know about world and domestic issues that may affect the environment in which I’m trading. So these are the fundamental issues that help give me context to my trading, I don’t look at company fundamentals to make decisions. I will look at some basic information like the sector a company is in and the volume of shares traded.
The point of this post is that I believe that the KISS principle is most important to success in trading, this has been the case for me in my personal journey.
In the past I’ve experimented with building many trading methods using indicators, and a lot of these experiments go something like this;
I’d start with the basic idea using one or two indicators, test and find a problem then add another indicator to fix the problem, test and uncover yet another problem, try to counter this problem with another indicator, etc. The chart ends up looking so congested that I can hardly see the price action but it seems to work until I test it on another segment of data or another market.
This is the type of thing I did years ago when I traded futures, I was 100% technical then. I even created a number of my own indicators and managed to finally create a method that worked specializing in one market, but it worked for two years and then the third year it barely turned a small profit for the year. A year’s workload and just a few dollars for my effort. I knew that I had much more to learn if I wanted to make an acceptable amount of money every year.
If you are doing what I did then I urge you to try a method that is as simple as you can make it, the KISS principle, and make it less specialized by testing to see if each step in your development works for all markets.
Learn from my mistakes if you’re making them and take the shorter road to success.
"lipstick on the pig"
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?