Topic: PSAR is the holy grail of indicators - page 2. (Read 6145 times)

It about using indicators to know what people think is going to happen and then find the spot where it will deviate.

UserName,

Thanks for bringing this indicator to light, had never heard of it before.

Noticed you're chart had it in the context of 5 Month / Daily display.

Did some research on it and found this helpful info on babypips web site:

--------------------------------------------------------
How to Trade Using Parabolic SAR
The nice thing about the Parabolic SAR is that it is really simple to use. We mean REALLY simple.
Basically, when the dots are below the candles, it is a buy signal; and when the dots are
above the candles, it is a sell signal.

Simple? Yes, we thought so. This is probably the easiest indicator to interpret because it assumes
that the price is either going up or down. With that said, this tool is best used in markets that
are trending, and that have long rallies and downturns. You DON'T want to use this tool in
a choppy market where the price movement is sideways.

Read more: http://www.babypips.com/school/parabolic-sar.html
--------------------------------------------------------

Combinations of moving averages of various types, sampling periods, and intervals can be used to fit a wide range of assertions to arbitrary data samples. I'm not dismissing Goomboo's results on that basis, I just know that to be true by virtue of having done it. In a small, mostly rigged game like bitcoin, I don't find technical analysis very meaningful at all. The best models for understanding bitcoin as it is can be found in the realm of pretty standard penny stock manipulations.

I guess "followed" would be a better way to put it.  If I write a bot that follows the 10/21 EMA rule today, I'll get clobbered, because now people are following it, and their act of following it will cause the indicator to not work any more.  Just because it is an old and well known indicator in the real world isn't really the problem, but now that it is a known indicator here it will no longer work here.

+1.  Exactly what I believe about indicators.

The market is an interesting place, because market activity never repeats itself. It does rhyme though, like the Mark Twain quote.

One can curve fit any indicator or methodology to the past and claim a holy grail, but the future is a different matter entirely.

The truth is, many of the well-known indicators do work in practice, even though they're "known."  This is a common myth about technical analysis.  In fact, the fact that many people follow them makes them self-perpetuating, which serves to reinforce their usefulness.

One need only look at Goomboo's backtesting of the 10/21 EMA rule in order to see that simple and well-known things can still work. 

I was speaking in much broader terms of the applicability of the PSAR. There are measures I have found helpful, particularly after adjusting the parameters, I find the usual fixed default values for things like MACD, Bollinger bands, etc on most canned chart packages ( e.g. Bitcoinica's charts ) are not very useful.

I base this on having done quite a bit of back testing on historical data with all manner of measures. For example I was using Bollinger bands years before John Bollinger put his name on them and I found them very useful for picking entry points for an enhanced dollar cost averaging savings approach for mutual funds/IRAs/401K plans/etc. OTOH I find them much less useful in application to individual stocks.

It's worth going through the exercise of analyzing these measures that get tossed around so frequently, for one thing it becomes evident that by adjusting the parameters one can construct quite different conclusions for both specific and general scenarios.

As far as reversals go, the simplest example of a useful indicator I can think of would be a triple top or bottom, and those don't always work out either.

I definitely believe there is a large aspect of self fulfilling prophecy resulting from the use of the same technical indicators by so many parties. That alone is reason enough to watch them, but I don't believe anything all the time.

Technical analysis is the practice of following trends.  If you want the PSAR or any other technical indicator to predict a reversal, good luck.

Though there were already downtrends before the PSAR flipped, it evidently served as a confirmation.

Agreed, I always found PSAR to be good for confirming what had already happened too late to act on it 

I would take issue with your claims that your post had a "purpose" or was "interesting." I'll let everything else stand, though.

The purpose of this post was to point out an interesting correlation between an indicator and market activity over the last few months.  "Holy Grail" was put there to draw people's attention, and nobody really believes that there is any "holy grail of indicators."

+1; but something else to add

Relationships between indicators break down once underlying market conditions which led to these relationships change. Conditions are constantly changing, particularly for something as inchoate as bitcoin. Since you have no idea why the indicators showed a relationship in the first place, you will have no idea when this relationship will break down, or what the new relationship will look like.

Holy Grail is an appropriate metaphor as it draws an implicit analogy between what you are doing and a hopeless, futile pursuit.

Note that PSAR is nothing magical. Most of time, it flips AFTER the change in sell/buy mood happens. At that time, it is obvious that rally/sellof started. But it looks nice when layed on historical data, though. ;-)

Indicators stop indicating once they become known.

Good find indeed