33 replies to this topic

[dcengr]

Looking at the weekly, same constraints...code did trigger a KR.



Results of all 7 tests....top row shows how many bars (in this case weeks) held. Maybe more here than originally thought and very supportive of my view that better buying opps lie ahead

This is the kind of stuff I like, things that can be shown. But to be statistically relevant, more samples should be present. A test going back to 1900s may reveal something completely different.

I'll run some stuff on my end since part of my code is already done.

[J.Bilkins]

Hey TE, Tzu here. I like the way the guys are explaining in detail why they think that way. I really like how they show why in detail and are open to suggestion. I wish I could have that type.....attitude(?)....but as you know I simply and bad with explanations and I automatically expect that people should figure things out correctly the way I have. This has been helpful to me.

[arbman]

dcengr, here's a little more than a candlestick system you can trust if you want to do historical scans;

1. Calculate the width of the bollinger bands at multiple time frames, let's say 2.5, 5, 10, 20, 40, 80 wks for 2.5, 5, 10, 20, 40, 80 wk periods.
2. Calculate the CCIs of the price at these multiple time frames. So, you will have 6 time series with different lengths in each.
3. Find the closest matching patterns historically starting from the largest to the smallest, so first find the highest correlating 80 wk, then 40 within it etc. Iterate until the correlation falls below 60% in each of the time frames.
5. Repeat these until the 80wk correlations in the historical data falls below the 60% at least. Now, you have separated all of the possible time frames that might be relevant to the current time frame.
6. Calculate the BB widths and the CCIs for N/2 time periods in the separated time frames, so the BBs and CCIs in the following 40wk for the 80wk time periods, the following 20 wk for the 40 wk time periods etc...
7. Calculate the average expected bollinger band widths and the CCIs by weighting with the correlations in each of these N/2 time periods previously calculated...
8. Inverse transform the current prices according to your expected BB and CCI widths and make a forecast...

This is a pure statistical method by using the historical data. The momentum and cyclical data will allow you to mold the existing data and make a forecast without these since you evolve the determinant characteristics of the current prices, but you have to have a very good model. In fact, the better your model is the more accurate your historical scans will be. BTW, to be even more accurate, you can start with the currency correlations, then yields and then stock indices of the same time periods, or simply do a multiple correlation of all in each of the time periods above.

This is the kind of stuff I like, things that can be shown. But to be statistically relevant, more samples should be present. A test going back to 1900s may reveal something completely different.

A better way to test the signals is to test for the past 20 yrs without the past 3 yrs and then see whether the method you found by using the prior 20 yrs is still applicable to the past 3. This way you can also tell whether the methods are not outdated or discovered by too many...

- kisa

[arbman]

BTW, you need to build the BB withds in percent changes, not in absolute values. Absolute values would not make any sense in a statistical scan since the prices are generally going up in this inflation based system and the absolute volatility amounts will be always larger than the historical ones under the conditions with the same percent changes... In that sense, you can understand the amount of liquidity out there by the lack of volatility, the volatility is a product of illiquidity...

[Jnavin]

Fib's carefully measured thinking on this subject is right here.

[MrOrange]

test

Edited by MrOrange, 28 January 2007 - 11:32 PM.

[dcengr]

Just a reminder to all. It is interesting to read about descriptions of methods, patterns, theories, etc, but it is IMPORTANT to BACKTEST and prove such methods, patterns, theories in a scientific manner. If I were to say "xx pattern is very bullish and so and so occurs when seen under yy conditions".. would you just believe me because I said it? Would you just BET your money on it without proof and take it for granted that I am right? Would it matter if I was a successful trader or not? If I sold you snake oil and told you it would cure all your physical ailments, would you buy it? People once did buy them, and they still do buy more sophisticated versions of it. One should be careful believing what's written unless PROOF is provided, or if PROOF is not provided, one should take some initiative to validate such PROOF if one feels it is worth while and wishes to bet using it. As I said before, I will do some more studies on this key reversal pattern, and will publish findings, that are back tested going back to 100 years of data, to see if there's statistical relevance to it. There's lies, darn lies, and statistics .

[fib_1618]

It is interesting to read about descriptions of methods, patterns, theories, etc, but it is IMPORTANT to BACKTEST and prove such methods, patterns, theories in a scientific manner.

Spoken like a true fundamentalist whose sole goal is to disprove technical analysis in any way, shape, or form.

One can back test all they want to, but until you can back test the backdrop of the mass psychology that accompanies any or all technical bar (or candlestick) structures (the why and when they appear when they do), I believe that proving things in a "scientific way" will never provide the "proof" that would be necessary to convince those who consider this kind of market analysis questionable at best, and voodoo at the very least.

Fib

[dcengr]

It is interesting to read about descriptions of methods, patterns, theories, etc, but it is IMPORTANT to BACKTEST and prove such methods, patterns, theories in a scientific manner.

Spoken like a true fundamentalist whose sole goal is to disprove technical analysis in any way, shape, or form.

One can back test all they want to, but until you can back test the backdrop of the mass psychology that accompanies any or all technical bar (or candlestick) structures (the why and when they appear when they do), I believe that proving things in a "scientific way" will never provide the "proof" that would be necessary to convince those who consider this kind of market analysis questionable at best, and voodoo at the very least.

Fib

Now you're putting words in my mouth. If I was a fundamentalist, why would I use TA? Has anyone here seen me post more fundamentals or technicals? If you were to take my messages and by sample of the type of messages I was posting, would you conclude that I was a fundamentalist or technician? This sort of testing is exactly what I believe many "TA theories" do not provide. A lot of "fancy words", a lot of "huffing about sounding like an expert", and ABSOLUTELY NO MEAT.

Backtesting has no relevance to fundamentals. It is a way of measuring the accuracy of observations. If one things a certain pattern has a certain outcome much more than another outcome, then examples would be appreciated, both for and against it.

Let me ask you something, Fib. Are you an expert in mass psychology? Do you hold a degree in psychology? Are you considered an expert by others considered as experts in mass psychology? What is your "profession"?

I have a degree in engineering. I've worked in the field for many years. I got paid for it, and I believe I was rather good at it. I don't go around claiming to be a medical doctor, however.

But if you want an advice on your health from me, I'll be glad to give it. You can, however, take the advice and do with it as you wish. Don't complain to me if the advice wasn't back tested

[fib_1618]

If you were to take my messages and by sample of the type of messages I was posting, would you conclude that I was a fundamentalist or technician?

You seem to have again misunderstood the direction to which I was referring to in my post.

Let's try a more direct approach.

I believe that you have shown a ponderous to continue to find something scientific as it relates to some forms of technical analysis when in actually is really an art. Nothing wrong with that, in fact, with your engineering background, it would be fully expected. But what you will eventually find out is that it won't get you very far in these same areas of market analysis to whatever you're attempting to understand and/or utilize in your own trading. When it comes to back testing, for example, there are many technicians out there (Larry Williams comes to mind) who do a lot of back testing of many market hypothesis, but this is more in the area of inter and intra market number crunching correlations than it is to actual specific price bar charting action in and of itself.

So....we're just disagreeing on when the use of back testing is appropriate in determining if there is a enough evidence to conclude that something like a "Key Reversal" is something in which can be used in our effort to make money or not. You made your point, and I made mine.

As far as any conclusions that I or others may have in as much how they view your posts, I believe you under estimate the quality and cleverness of the people who either post here or are just here to compare notes. So in this area, I will leave such conclusions to ones own determination.

Fib