While I agree on getting rid of what people would describe as 'typical indicators' (price derived indicators of the same source instrument), I disagree on it being impossible to have a strategy that works in trend and "sideways". Simply extend the horizon and holding period to trade multiple days at a time, with a scaling in and scale out strategy.
Due to time constraints, please do not PM me if your question can be resolved or answered on the forum.
Need help? 1) Stop changing things. No new indicators, charts, or methods. Be consistent with what is in front of you first. 2) Start a journal and post to it daily with the trades you made to show your strengths and weaknesses. 3) Set goals for yourself to reach daily. Make them about how you trade, not how much money you make. 4) Accept responsibility for your actions. Stop looking elsewhere to explain away poor performance. 5) Where to start as a trader? Watch this webinar and read this thread for hundreds of questions and answers. 6) Help using the forum? Watch this video to learn general tips on using the site.
If you want to support our community, become an Elite Member.
Last edited by Big Mike; February 11th, 2015 at 04:01 PM.
The following user says Thank You to Big Mike for this post:
I am not sure I follow on the horizon and holding period. My "bots" that perform well only do on daily charts and on certain instruments. For instance, reversals are good on GC and bad on YM/ES. MA's are good on YM/ES and bad on GC. I guess mixing the two works but it is no holy grail.
What I mean (and in agreement with previous posts here) is that I haven't been able to find an automated strategy that performs on every instrument, and if I find something close I take a trade every leap year. LOL.
I read a couple of online trading forums and found for daytrading index futures traders are actually scalping.
All the so called"trend following" systems are not really about trend. The systems based on intraday EMA are actually following price moves, not following trend.
A person walk along a direction and his dog run around him,sometimes at same direction sometimes at opposite direction.
The person's way is called trend.
The dog's way is called price moves.
You follow person's way you have high probability.
You follow dog's way you have low probability because dog's direction is random.
So a market trend is mostly decided before the day begins on daily chart.
Last edited by wmwmw; February 15th, 2015 at 10:00 PM.
Nice comparison, but I am focusing on the dog now because the owner gets whipsawed on a consistent basis.
@wmwmw what are you using to follow the trend on daily charts? Are you using filters as proxy? How does it perform on an instrument such as CL? I have not found a trend following strategy that I can rely on, not yet...
I know this is a type of trading that I do not do... but to say all indicators are useless and that all are created equal is not true. I am glad to see you say this.
It takes a long time to understand how indicators work. I find that most people who use indicators have no clue how to apply them. By that I mean taking conventional indicator buy/sell signals at face value and shrugging when they does seem to work.
Take ADX by itself... I see many people using this as a trending indicator....which it is not!! This indicator shows the STRENGTH of the current trend...not the trend itself. So if the ADX(30) is rising above 25, whatever the current trend is Bullish or Bearish, it is getting stronger... personally I don't like this indicator as it is severely smoothed and I don't find it particularly useful...However the DMI+/- from which the ADX is calculated is excellent
I use a lot of indicators in my equity trading. I don't focus on one particular indicator at the exclusion of others. All indicators give false information at one time or another but if you develop a suite of indicators that do not perform alike (eg. MACD and TRIX though calculated differently essentially plot the same) you can reach a consensus from the TRENDING of each indicator (so many people focus on the number itself) and arrive at a decent conclusion.
I really wonder how you can apply indicators that were really designed for continuous data charts, then use them on Tick charts and expect the conclusions to be the same. I am certainly not an expert on Tick charts but that does not seem reasonable to me.
I don't subscribe to every indicator under the sun...many are crap and many of the good ones out there are not described properly. It takes time to understand how they really work... also some that work fine in daily trading where there is usually good continuous data are crap in intraday trading in illiquid situations.
The following user says Thank You to Underexposed for this post:
What you're talking about here is intraday market-typing.
Firstly, you can't know anything for certain, so you're going to have to work with probabilistic outcomes and accept that you'll often be wrong.
There are two ways we can approach this . . .
Your underlying concept here is to assume that because past futures have resembled past past, future futures will. In other words, we anticipate auto-correlation. We expect that the future state of affairs will mirror the current state of affairs.
Track the range expansion as the session develops and extrapolate this to predict the range for the full session. So, if you're using one minute bars and you know that there are 360 of them in a session (there aren't!), then at any time you could calculate:
Predicted Daily Range = ( Current Range / Bar Count ) * 360
You're basically calculating how much each new bar is contributing to the daily range expansion and then extrapolating.
Keep in mind that even though this might predict the daily range with absolute accuracy, it isn't telling you anything about how that range will form. Maybe the market drifts steadily lower for the entire session, with the high at the open and the low at the close, or maybe the daily high and low are in place within the first 30 mins of trading.
Want to get more advanced?
Then you need to discount (i.e. normalize) for time of day. This means collecting a lot of data. Each 1 minute range is then normalized against the average range for that particular one minute bar each day. If you don't do this, then your predicted daily range will wildly overshoot at the start of each session - the additional volatility that is usual in the first few minutes of trading will have an overwhelming weight on predicted range well into the session.
Your underlying concept here is to assume that all things are mean reverting. The market's choppy now? Then a trend is sure to follow. There's a nice directional move unfolding? Then things are sure to become range-bound sometime soon.
You need a measure of "trendiness", and I'm not talking about some over-precessed indicator such as the ADX - something as simple as a percentage of consecutive up/down closes will do - and then you assume that this measure will revert to it's mean.
One way to do this is to take a measure of the range, make it directional, and then apply bollinger bands to this. Whenever your measure is below the average and heading towards the lower band then you anticipate a reversion to the mean.
WHICH METHOD TO USE?
That's up to you! My belief is that you need to take a "philosophical" stance on market behavior and then use your strategy to back this up.
Maybe neither approach is sufficiently nuanced?
Perhaps you need to use Neural Networks to recognize the patterns that characterize different market types?
I hope that my answer helps, and at least gets you thinking about objective ways to approach this problem. Most of all, successful trading is about problem solving, and you have gone far enough to define a problem - now you just need to find a solution for your strategy.
The following 3 users say Thank You to BestBrokerDeals for this post:
The most popular method is to use a few moving averages equidistant from each other like a 100 ema, 200 ema, and 300 ema. When they they tangle the market is chopping; when they run free of tangle the market is trending. Then you'll need to set a variable of how much time transpires after the last intersection before declaring it trending again. You'll need to fiddle around with these parameters to get it right.
The following user says Thank You to Itchymoku for this post: