trend determination
- Thread starter robinxing
- Start date
Anyone can define trend any way they choose. There is no trend. Trend is a distraction.
You might experiment with Microsoft Excel spreadsheet and historical price data. If price value is greater than price 200 days earlier then buy. If price value is less than 200 days earlier then sell. You might experiment with different time intervals.
I find it constructive to think in terms of "price following" instead of "trend following".
<img src=http://www.semelab.com/mil/images/mil_grp_diode_intro.jpg \img>
A financial diode might consist of a set of rules that allow money to flow mostly into my account and occasionally out of my account.
You might experiment with Microsoft Excel spreadsheet and historical price data. If price value is greater than price 200 days earlier then buy. If price value is less than 200 days earlier then sell. You might experiment with different time intervals.
I find it constructive to think in terms of "price following" instead of "trend following".
<img src=http://www.semelab.com/mil/images/mil_grp_diode_intro.jpg \img>
A financial diode might consist of a set of rules that allow money to flow mostly into my account and occasionally out of my account.
You propose an all-or-nothing scheme: either this collection of "systems-as-indicators" tells you Yes (there is a trend) or No (there is not a trend). Another idea is to have many shades of gray between Trend and NotTrend.
Why not divide your account into eight equal pieces and use them to actually trade eight different conventional Trend Following systems? Now your account can vary between 100% long (all 8 systems are bullish) and 100% short (all 8 systems are bearish), with many shades of gray in between. When there is no trend, the systems' positions will generally cancel each other and your net position will be flat.
The system proposed by another poster goes by the name "Momentum" in software like TradeStation. It uses this definition:
Why not divide your account into eight equal pieces and use them to actually trade eight different conventional Trend Following systems? Now your account can vary between 100% long (all 8 systems are bullish) and 100% short (all 8 systems are bearish), with many shades of gray in between. When there is no trend, the systems' positions will generally cancel each other and your net position will be flat.
The system proposed by another poster goes by the name "Momentum" in software like TradeStation. It uses this definition:
- Momentum(Close,200) == Close(today) - Close(200 bars ago)
- If Momentum(Close,200) > 0 Then Reverse To Long
- If Momentum(Close,200) < 0 Then Reverse To Short
random trend notes
-- Define what a good entry means for you
-- Define your trading time frame
-- Define your trend
-- See if following your trend makes your entries "gooder" than not following your trend, or random entries
People are continually arguing about trends and their value. Many have investigated and concluded that trends have no value in their trading. I do not doubt them. Others (me included) have found great value.
The method of investigation has a lot to do with the conclusions. When people get into predictive mode, they of course find trends to be useless because markets are unpredictable. However, viewing the markets and trading from the proper (poker-gambling methodology) perspective, it is not difficult for me to see how trading with the trend tips the scales in my favor.
Don't be afraid to use your eyeballs. I once competed against a computer to identify entries in which trend recognition was a part of the methodology. The computer used a moving average. I looked for a higher low or a lower high as new bars came on the screen. Very simple. I was sure I was going to be embarrased. The results surprised me. The computer had almost three times more entries than I did, and my entries were of considerbly higher quality. I think anyone can do it.
Whether a budding trader can do anything with trends (or anything else, for that matter) goes back to what I first mentioned: what is a good entry? It has got to be clearly, and quantitatively, defined.
-- Define what a good entry means for you
-- Define your trading time frame
-- Define your trend
-- See if following your trend makes your entries "gooder" than not following your trend, or random entries
People are continually arguing about trends and their value. Many have investigated and concluded that trends have no value in their trading. I do not doubt them. Others (me included) have found great value.
The method of investigation has a lot to do with the conclusions. When people get into predictive mode, they of course find trends to be useless because markets are unpredictable. However, viewing the markets and trading from the proper (poker-gambling methodology) perspective, it is not difficult for me to see how trading with the trend tips the scales in my favor.
Don't be afraid to use your eyeballs. I once competed against a computer to identify entries in which trend recognition was a part of the methodology. The computer used a moving average. I looked for a higher low or a lower high as new bars came on the screen. Very simple. I was sure I was going to be embarrased. The results surprised me. The computer had almost three times more entries than I did, and my entries were of considerbly higher quality. I think anyone can do it.
Whether a budding trader can do anything with trends (or anything else, for that matter) goes back to what I first mentioned: what is a good entry? It has got to be clearly, and quantitatively, defined.
Now to get completely off topic, an important step for me was to stop testing by formulating a complete strategy and seeing the ending +/- $. A modular approach is better. (1) If you put a pro and newb in separate rooms, and give them the same entries, the pro will make money and the newb will lose. (2) Pro pool players take a lot of easy shots. (3) Pro poker players lose less then amateurs in bad situations.
Put that in your hat and carry it around for a while.
Put that in your hat and carry it around for a while.
bluedemon77
Guest
Please give an example of what you mean by modular.
=============
Yes RobinX;;
& 50dma,200dma is still a good starting point,
may want to add some to it.
Wisdom is profitable to direct
