Those that poopood on moving averages, thoughts on this?

Simple MA based trading systems stopped working around 2010. I base this statement on a dozen or so research papers I've read regarding profitability of simple technical indicators. The papers can be divided into 3 categories. Those that used data pre 2010, post 2010 and some that included 2010 data. You can guess that that the papers using pre 2010 (or there about) show that simple TA indicators can be profitable and vice versa for post 2010. I'm using 2010 as a rough estimate of when it seems that simple TA systems seemed to stop working. One of the papers specifically mention this conclusion as well(yes the data period also included 2010 data). That same paper also speculated that simple TA systems might still work on weekly prices. I don't remember the names of those papers, but a poke around google scholar should yield the relevant info if you want to read further.
 
traderGOD said:
while moving averages are support and resistance, the steeper they make them less strong and more susceptible to a break.
More...

Please explain how MA are support and resistance. Unclear to me referencing the literal definition of MA. A particular MA can under certain long term market conditions be a major frequently cited technical support or resistance level but that's not the real intent or purpose of an MA per se - it's not like, for example, a pivot point which is an on-the-run calculation derived from H,L,C.
 
Last edited:
bitstream_ryder said:
Simple MA based trading systems stopped working around 2010. I base this statement on a dozen or so research papers I've read regarding profitability of simple technical indicators. The papers can be divided into 3 categories. Those that used data pre 2010, post 2010 and some that included 2010 data. You can guess that that the papers using pre 2010 (or there about) show that simple TA indicators can be profitable and vice versa for post 2010. I'm using 2010 as a rough estimate of when it seems that simple TA systems seemed to stop working. One of the papers specifically mention this conclusion as well(yes the data period also included 2010 data). That same paper also speculated that simple TA systems might still work on weekly prices. I don't remember the names of those papers, but a poke around google scholar should yield the relevant info if you want to read further.
More...

This post mentions one of the factors which I previously noted in an earlier post which requires tuning and adjustment. Note that the "success" was time frame dependent - you mentioned 'Weekly'.

Also, I have read literally hundreds of academic white papers on trading since the 1990's - including graduate level theses. I've always found them to be less than useful for the most part. The better, more applicable white papers IMO are published by the trading exchanges.
 
The reality is that the way moving average systems have traditionally been used have either always worked, and will continue to work in the future, or have never worked, and never will work. Most of us know it is the later. Same for most other well known TA indicators.

The reason MAs as traditionally used don't work is because of lag. The technical term is group delay, as well as phase distortion for IIR style MAs. Some newer designed MAs are truly near-zero lag. However I've found that they need to have fairly high frequency cutoffs to be responsive, and generally have inadequate attenuation in their frequency stop bands.

True zero-lag (zero group delay) MAs are available with excellent attenuation, but are non-causal. You can look that term up. That doesn't mean they are useless for trading, just that you will need to find creative ways to utilize them.
 
panzerman said:
The reality is that the way moving average systems have traditionally been used have either always worked, and will continue to work in the future, or have never worked, and never will work. Most of us know it is the later. Same for most other well known TA indicators.

The reason MAs as traditionally used don't work is because of lag. The technical term is group delay, as well as phase distortion for IIR style MAs. Some newer designed MAs are truly near-zero lag. However I've found that they need to have fairly high frequency cutoffs to be responsive, and generally have inadequate attenuation in their frequency stop bands.

True zero-lag (zero group delay) MAs are available with excellent attenuation, but are non-causal. You can look that term up. That doesn't mean they are useless for trading, just that you will need to find creative ways to utilize them.
More...
what do they lag? crossovers may lag but that is not the only way of using ma: there are many others and i will explain them here using charts as a examples
a 5 period ma will show what the price has been doing over 5 periods while a 100 period will show what it has been doing over 100 periods.
if you say they lag, that is when they are used as a cross over; if 5 crosses 20 from top giving a sell signal,how valid is that signal.it is only valid as short term sell IF both 5 and 20 are crossing above a upward sloping 100 ma. which means it may be a buy signal or opportunity for a longer term player
i have studied price action for the last 10 years and i have realised price action or indicators are both based on price. Both show the same thing one cannot show anything more than the other. Both depend on interpretation
If price moves above a 10 period ma it shows that it is high,but if price goes above another 10 minute bar it tells you it is above the last bar only : which is the more significant event?
 
Last edited:
panzerman said:
The reality is that the way moving average systems have traditionally been used have either always worked, and will continue to work in the future, or have never worked, and never will work. Most of us know it is the later. Same for most other well known TA indicators.

The reason MAs as traditionally used don't work is because of lag. The technical term is group delay, as well as phase distortion for IIR style MAs. Some newer designed MAs are truly near-zero lag. However I've found that they need to have fairly high frequency cutoffs to be responsive, and generally have inadequate attenuation in their frequency stop bands.

True zero-lag (zero group delay) MAs are available with excellent attenuation, but are non-causal. You can look that term up. That doesn't mean they are useless for trading, just that you will need to find creative ways to utilize them.
More...
see my post no 20 in this thread
 
this is a trade on peperstone training simulator see chart below

marked with red arrow: there was a crossover of 5 ma[brown] with 20 ma [red] which led to a down move which touched the up sloping 100 ma [green].
simple buy signal.
sim.png
 
In the chart below even if you had entered along well after the cross at the red arrow you would have got 1:1 RR.
the 5 ma [brown] was well above the 20 ma [red] showing good momentum.
sim 2.png
 
You must log in or register to reply here.
Back
Top