the basic flaws in TA

[kiwi_trader]

Quote from John Kuran:

So, what does this mean? We can't use DSP, Fast Fourier Transform or anything based on an assumption of a continuous price/time series (markets are discontinuous due to gap openings, price shock, etc.). Same thing for Arima, garch, VaR, CAPM, etc. Moving Averages are nothing more than low bandpass filters (DSP - Digital Signal Processing).

What we can do is seek out the point of temporary stability (trading ranges), monitor it for the onset of instability (breakouts / breakdowns) and trade accordingly.

What we can't do is predict prices. What we can do is look at current market conditions and look for changes in those conditions.

Markets are like weather systems. We can't predict with certainty what the weather will be like 3 years or 3 weeks out but we can do it for the next 3 days or so. We can't be precise in this prediction (ie, "the temp will be 84 degrees Fahrenheit at 2:54 pm EST on Friday") but we can be accurate if we keep it rough (ie, " a high pressure system is moving in, driving out the low pressure system that was responsible for the rain of the last 4 days. Expect sunshine with some clouds, temperatures in the 80s during the day, 60s at night, etc."). Same thing with the markets.

Thats a good post regarding the "physical" properties of the markets. And I largely agree with it at that level. But to some extent it ignores the "psychological" properties of the markets - that they are basically made up of large numbers of people interacting about their instantaneous perceptions of value.

That interaction and the fact that many use a mix of tools like trendlines, channels, and moving averages to build those perceptions mean that these "flawed" tools can be used to build effective technical trading systems. In my case, those tools add to the techniques you suggest to give a clearer perception of the likelihood of support, the likely direction of the next move and how I would detect that the probability had changed for the better or worse (from my perception if I had a trade on).

 

[Cheese]

Quote from John Kuran: ..
Markets are like weather systems. We can't predict with certainty what the weather will be like 3 years or 3 weeks out but we can do it for the next 3 days or so. We can't be precise in this prediction (ie, "the temp will be 84 degrees Fahrenheit at 2:54 pm EST on Friday") but we can be accurate if we keep it rough (ie, " a high pressure system is moving in, driving out the low pressure system that was responsible for the rain of the last 4 days. Expect sunshine with some clouds, temperatures in the 80s during the day, 60s at night, etc."). Same thing with the markets.

You may not appreciate what a good analogy the weather is for market conditions.

The importance of the near term prospectus has been flagged up by a few at ET. And if you move prediction from 3 days to 1 day ahead then powerful accuracy can really be harnessed by the player.

That said, for a predictive trading model you have to do an awful lot of homework, study and problem-solving before you even have a chance to discover much or most of the framework solutions that finalize such a model.

 

[Lights]

how convenient. you use the media well.

Quote from marketsurfer:

hi wolfe, hi nick------


What some figure out immediately, and others never seem to get acting like i am the antichrist---- IS.......

i ask these questions and start these controversial threads to deepen my understanding of the subject, to get a handle on what traders are thinking NOW, not last week but NOW, and to stimulate the overall progression of trading knowledge. Being able to understand exactly how other traders view the market in its present state is key to understanding the market itself. i use all tools, yet remain flexible and am not sold on any one.

thanks for the most excellent posts!

best,

surfer

 

[OddTrader]

The basic flaws in TA?

Perhaps the basic flaws in TA would be probably we don't know what are the basic flaws in TA, if any.

 

[cnms2]

We have to be careful with analogies. For markets it may not be true that you can always predict better smaller time frames, because the signal / noise ratio is poorer. On the other hand 3 years from now it will still be November (God willing ... ), so you kind of know the weather range. This is not the case with the markets.

Quote from Cheese:
You may not appreciate what a good analogy the weather is for market conditions.

The importance of the near term prospectus has been flagged up by a few at ET. And if you move prediction from 3 days to 1 day ahead then powerful accuracy can really be harnessed by the player.

That said, for a predictive trading model you have to do an awful lot of homework, study and problem-solving before you even have a chance to discover much or most of the framework solutions that finalize such a model.

 

[Cheese]

Quote from cnms2:

We have to be careful with analogies. For markets it may not be true that you can always predict better smaller time frames, because the signal / noise ratio is poorer. On the other hand 3 years from now it will still be November (God willing ... ), so you kind of know the weather range. This is not the case with the markets.

Weather seasonality was not the analogy. It was the act of predicting weather, short term compared to longer term.

Curiously and as no doubt others have noticed there is seasonality in price behavior. Where I have compiled a given market such as in an index futures market the effect makes itself noticed. I exclude crop or such futures which are expected to have seasonality influences. Sometimes it is strikingly consistent if first one or two other filters are applied. However I have not modelled or studied it sufficiently and offer no tested expertise on seasonality patterns as an accurate predictor.

 

[nononsense]

Quote from Cheese:

...
That said, for a predictive trading model you have to do an awful lot of homework, study and problem-solving ...

Right Cheese.
Let me add, only "discretionary artists" can make money without it.

 

[OddTrader]

Quote from Cheese:

However I have not modelled or studied it sufficiently and offer no tested expertise on seasonality patterns as an accurate predictor.

Thanks for telling us the truth, finally!

 

[Cheese]

Quote from OddTrader:
Thanks for telling us the truth, finally!

Are you being odd Oddtrader? I have never set out to make a study of seasonality!!

But you can tell this thread what your views are on the role, if any, of seasonality in price behavior.

 

[OddTrader]

Quote from Cheese:

But you can tell this thread what your views are on the role, if any, of seasonality in price behavior.

Perhaps we should keep studying about it, then one day hopefully we could start testing some "truly tradable and profitable" signals and systems, before quitting trading completely. My 2 cents.