Kiwi. Please quit giving away secrets. I agree completely with the power of such empirical fitting. But I use first order infinite impulse response filters to do that, and add to your fitting tuning of their crossover. Some minor advantages IMO. Mike.
you missed the famous trend follower ( profiled in the book by the same name, which by the way, is just out in paperback) who places a chart on the other side of the room to see if its trending. he doesn't use first order impulse response filters. its much simpler than those complex tools.
1st Order IIR filter? ... an EMA, by any other name... right Mike? Then again I'm not a signal theorist. i.e. , Alphafilter = (1 - alpha) * Alphafilter(-1) + (alpha * Price) == XAverage = XAverage[1] + SmoothingFactor * ( Price - XAverage[1] );
sorry i have to laugh - i was beginnig to wonder when someone would shoot this down. now had it been an auto adaptive moving average with browian motion theory filter - now then you would have something. mark brown ps it would also have to have session auto centering reset for the beginnig price input.
Sheesh, don't give it all away ! I'm sure Monsieur Hypostomus Plecostomus is using one of those filters you mention but he ain't giving it away foir nutin' ...
Before you can judge on trend following, you have to be able to find the trend, because you cannot judge on something you never knew or experienced. If you canât find the trend there are two possibilities: First: you are not smart enough Second: there is no trend because trend does not exist I assume that those who do not believe in trend following, were not smart enough to find the trend. This doesnât mean they are stupid. They are probably much smarter than I am. But in one thing I might be smarter: in finding a trend. I donât know if trends do exist, but I trade what I think is the trend. There are a few ETmembers who know what the results of this trading is. I think, based on what they saw and still see, that they are convinced that there is a trend most of the time. But I donât feel the need anymore to prove it. If all markets are in a transition to surprising the person who thinks he has the direction figured out, than the markets should revise their tactics as I make on average over 80% profitable trades in trend following. I seldom get stopped out. Another explanation could be that I have a lot of luck each day. But luck isnât luck anymore after +500 trades and +80% profitable trades. Most people who think that the market is after them to get their money, in fact use this explanation to protect their own ego. For your ego it is always better to blame anything or anyone except yourself in case of a failure.
the coinflip analogy: http://archives.math.utk.edu/ICTCM/VOL13/P002/paper.pdf runs happen in coin flips, runs happen in the market---can these be reversed engineered to profit on the next run/trend. NO. regards, surf
Equalizer. I can't fool anybody, it seems. If it were worth the trouble I would try to prove that the IIR preceded the EMA. Mike.
The very large capital people have a dilemma and they are smart enough to recognize it. They take trend analysis to extreme limits and just do as well as they can within their limitations. That will not change. You can see how they get their confidence limits in the literature and you will note the answer to your Q is there on the top of the pile. QED. If you want to use what you know have as a resourse personally, the coding can be done most easily by exporting THE differential ratio out of the MLR instantaneous result for any market display you chose. Then, with that in hand you may chose any specific price limit or value you wish and apply the differential ratio as the m of the line passing through you limiting point of choice. Repeat as desired. Here is an example of how this all walks forward. MLR of any duration gives you the direction of the market and its m gives you a time rate of change of price. You willl chose the trough preceding the market turn for your point of choice since you are rational. The line y=mx+b is the case that you asked for that continues to work as time passes and passes your test where the number of times is really flexible since you can choose any number and be right. For beginners this is the traditional (See dbphoenix level of thinking) condition for holding and finally exiting. For elliiot wave this line marks the trend limit of all the waves from incidence to termination. For fib folks they will be working off the peak following the trough you picked (perhaps). Their S limits (for example) will horizontally intersect your line when the period of the retrace is completed. You will get to see that the line pivots on the point you chose so that over time the next trough becomes incident with the line and the price value then leaves the line but remains within the trend side of the line. To see the end effects that any such trend follows by the inherent adjustment of the MLR as it is repeatedly reevaluated is a specific tangible event in time. This construct does not end but, in effect, it just takes out the surprise that markbrown is still experiencing at his stated level of perception. As in any mathematical/scientific effort, limiting effects are determined within the model that is in use. In this model a person reaches a fork in the road. 1. He continues to work on the same trend fractal. 2. He moves to an internal fractal of the initial fractal. This occurs a consequence of the study of the market. We all know how to measure trends. The m was the measure and a non trending situation has but one value of m. That is easy to recognise and make note of. The price succession of peaks and troughs in trends is well known. At some point a person can recognize that transitions between extremes represent money making opportunities and they go to another level of effectiveness and efficiency. People who are not able to see trends in markets are simply people who are so remote from the market that they cannot see price movement in markets. They are not traders as a consequence. Many people hold instruments for periods long enough to bridge trend after trend. They are not using techniques to trade trends; it is possible that their investment strategy is not a trend following strategy. Large capital pools exhibit a characterisitc that looks like the pool is always in the market and the transition form one instrument has a yield that approaches the performance of the collective market indexes. If a peron is going to become more and more effective and efficient he is going to use mathematics to be able to become more and more effective and efficient. If you are able to: 1. pick any starting point, 2. Take a measure of the slope of a trend going on (in any way you chose). MLR is a fun one. 3. Draw trend lines 4. Make periodic adjustments to stay current, 5. And recognize when the slope comes to zero occasionally. Then you will be able to make money continually with your capital by trend following. Not being able to see the current trends (there are many concurrent ones) and being in a state where you are getting surprised continually is just a conditional mental state. It does not have to be a permenant dissabling choice up to a certain point.