How do professional quantitative traders view traditional technical analysis?

Personally, I believe that indicators derived from volume and price are strictly technical indicators. Many people simply equate KDJ, MACD, moving?average, Williams and other indicators narrowly to technical analysis. But if I ask these questions here, we may think differently:

Is VAR a technical indicator?

Does BS count as a technical indicator?

Does GARCH, ARMA, SV count as technical indicators?

Does the classic Lee-Ready algorithm count as a technical indicator?

In fact, all of the above can be technical indicators, but some are simpler, some are more complex. Even if you introduce a variety of stochastic processes and non-linear models, as long as it is based on the volume of price, it is hard to believe that they are not technical indicators? The world's largest CTA?Winton and the second largest AHL development to date, so far their model inside the technical indicators is still a very important alpha source.

Whether technical indicators are useful or not, it is true that for different areas of quantitative work, they can be very useful. This is something that really does get looked at very differently for different areas of quantization.

First, for statistical arbitrage, there is a specialized class of alpha from volume price analysis, which, according to the practical experience of a friend of mine who worked at world?quant, still has a place in the highly mature U.S. market;

Secondly, for a variety of arbitrage strategies, isn't covariance regression a technical indicator?

The most classic spread SD is not a technical indicator?

Finally, back to the CTA, it can be said that most of these strategies are technical indicators, of course, the development of the CTA?fund, the data source has become more and more diversified, including the spot data, macro data and so on. But even so, the quantity and price strategy still accounts for a large proportion;

Why do people so disgusted with the technical indicators, 2 reasons.

1, is that it is too simple, and there are too many parameters, there is no way to verify the validity of it, to recognize the validity of such a simple thing is a seemingly low thing. This is how I think about this issue, any public model can be used to earn money? The market itself is adaptive, if a model has the market has predicted the effect of the use of more people, has long been 50-50.

2, is with the development of the market and the source code of the public, simple linear technical indicators do indeed have long been highly ineffective. You can't see that the first invention of the Williams indicator that who is not also full of money. Of course, in the end he failed, but this is another issue. Instead, there are all sorts of non-linear indicators that introduce stochastic processes and so on. The market is evolving and progressing, and the technical indicators used to derive and predict the market are of course also rising.