Tuesday, May 4, 2010

Meta-Trader - Candle Weight indicator

Welcome back Meta-Traders.

This morning I was just staring at this daily chart of EUR/USD trying to figure out what it was telling me - aside from the fact that its going down of course!

Then it occured to me the candles themselves have a weight. Big dark candles are heavy and big white candles are light. Put another way, black candles have a negative value, and white candles have a postive value. The bigger the candle, the bigger the value.

Then it occured to me that if you just add up these values over the past say 14 bars, you come up with the candle weight indicator. The result represents the total amount of movement over the past 14 bars as measured by the sum of the difference between the Close and the Open for the lookback period.

Note this value is closely related to the total range of the pair for that period - except it does not include candle highs and lows and also does not include the gaps.

There are a few things that I like about this as a trading indicator:

- Its always negative when the market is going down, and postive when its going up (at least for the selected timeframe and lookback)

- It has no fixed range - on the contrary, its range depends on range of the underlying pair over the lookback period.

- It switches from postive to negative and back again smoothly by taking the vote of all the candles equally.

A few minutes later, I had the Candleweight indicator - which is now uploaded to my Yahoo group under the name CandleWeight.mq4.

This indicator is really the height of simplicity so it appeals to me at a fundamental level, particularly because its completely price based and prices never lie - not for an extended period of time anyway.

A few minutes after that, I threw some logic into an EA (called the CandleWeightEA of course) that goes long once the CandleWeight value crosses up through some constant value, and goes short when the indicator crosses down through some opposite condition.

Then I ran a quick optimization on the Hourly charts from January 2009 and I found some broad areas of profitability. Before I get too excited, I realize this system will probably perform like a lot of other trend followers, never miss the big move, but suffer some hefty drawdown while the market sloshes around sideways.

Anyway, stay tuned for the Candle Weight EA and results of back testing and optimization.








4 comments:

  1. Hello Chris,

    Thank you for the post :o) It is actually an interesting idea and this is exactly the initial basis of the RSI indicator only that the RSI is a normalization of that addition over the total change for that given period so the RSI tells you what percentage of the movements for the past X bars equaled loses or profits while your indicator tells us the unormalized change.

    Obviously it would be useful for trend following but probably a cross is not the best idea. You can use a tactic similar to that of Watukushay FE. For example wait for the indicator to cross X value below and then enter in the direction of the trend when the indicator shows a retracement. This allows you to maximize profits on trends. Just some simple ideas for you to try :o)

    And how about coding the derivative of the indicator ? this would signal turning points in direction which could be useful to design strategies. What do you think ?

    Nice post Chris, hope to see more of your indicator and EA developments :o)

    Best Regards,

    Daniel

    ReplyDelete
  2. Daniel-

    I appreciate your comments - particularly because I ran into a brick wall on back-testing. I found profitable parameters for periods of time, but nothing that was profitable across a multi-year period.

    Also, understood what you mean about normalizing the indicator based on ATR or something like that. For some of the larger parameters (50 pips) the EA had good tests for 2010, but didn't trade at all in 2007 due to the absolute-value nature of the indicator.

    I'll give you other ideas a try and post the results. Thanks again,

    Chris

    ReplyDelete
  3. Hi. I am forex-cat.

    Your article is always useful.
    Thanks.

    Good luck trading today.

    ...my blog:
    http://forex-chart-analysis-and-a-cat.blogspot.com/

    ReplyDelete
  4. I have not got into creating any indicators yet, but it would definitely be a fun exercise.
    I suppose all my time is spent on playing with EA's and such.
    Tiger Tail "tried" to take a second live trade, but I got the re-quote error 138. I have been researching it and found some ideas to recode how I enter a trade. It just goes to show you that there always seems to be something to learn. (and definitely learn on how different a demo account behaves from a live account)
    I hate missing trades, especially ones that would have been profitable, as it messes with the "predicted" data for the year.

    JT

    ReplyDelete