Some ways to define a Trend.
- Thread starter dartmus
- Start date
Just settling in for a quiet night tonight with the US holiday and majority of my written FTSE100 options having expired OTM last friday. Still, it doesn't pay to rest on my laurels whilst I also have futures positions on, no point giving back profits. Even though the indexes are still moving up, the trend is my friend, very often the reversals start at option expiry or public holidays, so I need to be ready to unwind my positions in the blink of an eye, moving averages just don't cut it. I rely more on market action/behaviour to define the trend as often the trend has changed/in transition well before the price move has been exhausted, often this exhaustion thrust will appear most clearly as a tail on candlestick charts. Although to be honest I don't use charts per se, just making the point. It doesn't automatically follow that I don't have a sound working knowlege of TA and Charting, it is amazing the amount of wrong places we look, only limited by our determination, and the things we learn before finding something repeatable and successful. Generally, the market doing a flip without warning will normally only occure on the back of market moving news. These and many other hints regularly occure, you just have to have an open mind and take the time to find your own answers, independent thought is king in this arena, remember if it were so simple as a few lines and squiggles on a chart everybody would succeed. Oh, it also usually pays to hedge your bets.
Good luck on the journey.
Cheers John
Good luck on the journey.
Cheers John
Mathematically, you can define a trend as "auto-correlation" in the "returns" time series. For example, consider day-to-day time series of a stock. The difference in daily prices (today's price minus yesterday's price) is defined as "return" on this stock. If there exists correlation among successive returns (auto-correlation), then you can say that the the price is trending.
But this is not enough for trading success. You also need to measure the "strength" of the trend, often also called "momentum". This is a measure of how long the trend is likely to continue. The most important clue for this comes from trading volume.
Note that all measure of trend are lagging. If your strategy requires more confirmation for the trend, then you will have less risk but lower profit (you are jumping on the wave later than most). If you require less confirmation, then you take more risk with potential greater rewards. Managing the two is essence of risk management (optimization).
Volume is not the only way to calculate strength. The significance of volume is limited. If you have big volumes all day like in S&P futures, then volume has no importance. I calculate strength without the use of any volume at all. And with fairly good rate of success. Behavioral finance is more important than volume to me.
I agree: trend and strength combined should give good results.
I agree: trend and strength combined should give good results.
Ultimately, if you are a trend follower, you need to have multiple strategies in place, because it is just too difficult to determine when a trend has begun, and when it will end. You also won't know once it begins whether it will be a nice smooth trend or one of those ugly choppy trends.
Sure, there are some clues in the case of some blow off tops such as gold in 2011 and super volatile bottoms such as in the stock market in 2009. But, in many cases, trends just end quietly. All you can do is try to capture as much of it as you can.
Sure, there are some clues in the case of some blow off tops such as gold in 2011 and super volatile bottoms such as in the stock market in 2009. But, in many cases, trends just end quietly. All you can do is try to capture as much of it as you can.
What indicators of strength do you use?
Exactly. The trend should be monitored all the time to check the strength. Strength can change over time so trading should be adapted.
Home made recipe that should stay in the house. Study behavioral finance, there you can find answers to many questions.
I gauge the strength of a trend like this: If buyers (or sellers in a downtrend) step in around a 9-period EMA and price makes another higher high (or lower low), the trend is strong and exceptional profits are likely if you add to a winner. If buyers/sellers step in around a 20-period EMA and price makes another higher high or lower low, profits can be extracted from the trend in both directions (with-trend or counter-trend), but larger swings will occur in the direction of the trend. Sometimes there's a trend (HL/HH or LH/LL) but the channel is so wide that ample profits can be extracted in either direction and moving averages less than a 50-period EMA are fairly useless.
For me, the key to trading in a trending environment was to identify how to enter on pullbacks. In a strong trend pullbacks are so shallow that a smaller bar interval is often useful for seeing the continuation patterns.
In case of an uptrend, the faster EMA (9-period in your example) will always be higher than the slower EMA (20-period). So you are saying that, in case of uptrend, if price exceeds the faster EMA, then the uptrend is strong?
In this case, both EMA are indicating an uptrend (or downtrend), and if the price is between the two EMAs then the trend is no so strong, but still large swings will occur in the direction of the trend, and hence smaller profits can be extracted?
In this case, it seems you are describing a more volatile scenario wherein trends on a smaller time frames must be identified.
Perhaps the key to trend trading should be a 3-step process. First a measure of volatility must be devised to ascertain the time frame for the trend. If volatility is high, then trends will be occurring on smaller time frames. Once the time frame is identified, then the direction of trend should be determined (EMA, trend lines etc.). Once that is done strength should be determine (volume, price action as NoDoji described, macro-economic factors for longer time frames, mysterious homemade recipes based on psychology etc.).