Below are some demonstrations of various key features of *
QuanTek*, and also of its precursor

We have been performing some tests of the **QuanTek****Adaptive Filter**, varying the parameters to achieve
best performance. You may check the most recent demonstrations of
these tests on the following pages. These tests will be updated frequently, and they demonstrate how
the **Adaptive Filter** performs under a variety of
market conditions.

**Correlation Test - AAPL (2018-12-07):** This is a
test of the performance of the **Default Adaptive Filter**
using the **Correlation Test - Filters** dialog with a
variety of settings, using **AAPL (2018-12-07) **data:

The **Adaptive Filter** was calculated using a
**Time Horizon** setting from **1** to
**128** days. Then the **correlation**
with **N-day future returns** with **N**
from **1** to **128** days was measured.
The measurement was also performed using four different **
Correlation Scales**, which are the time intervals over which
the **correlation** is measured. The four **
Correlation Scales** are **128 days** (6 mo.),
**256 days** (1 yr), **512** days (2 yr.),
and **1024 days** (4 yr.). The results are compared to
the **expected return** due to the **2048-day
trend line**, extrapolated into the future.

**
Demo: Displays of Technical Indicators** (Revised September
10, 2006) This demonstration describes some of the most important displays, and the **Technical
Indicators** which are associated with them. A description is also given
of how some of these indicators are calculated and how to interpret them.
In particular, the **Price Projection** and **Harmonic Oscillator**
indicators, and the **Momentum** indicators
and **Trading Rules** indicator are explained, and the tests used to
establish their validity by computing the **correlation** with **future
returns** are described.

**Demo: Linear Prediction Filters **
(Revised September 8, 2006) This demonstration shows the
results of testing two of the **Linear Prediction** filters using the **Correlation
Test - Filters** test, for **correlation** of the **Price Projection**
with **future returns**. The test is done for **MSFT**, **XOM**,
and **AAPL**
stocks. The result is a high degree of correlation for a **time horizon**
greater than 10 days, but for a shorter **time horizon** the correlation
tends to be buried in **stochastic noise**.

Demo: Momentum Indicators -
Design and Testing
(Revised November 13, 2006) This demonstration
shows how to design and test a set of three **Momentum** indicators, and save the results to an individual
security data file. The **Technical Indicators** dialog is used to
design the indicators, and the **Correlation Test - Indicators** dialog is
used to test for **correlation** with **future returns**, for a variety of
settings of the **time horizon** and **Lead Time**. The **Lead Time**
is then used to optimize the indicator for maximum **correlation** with **future
returns**. The final results of the optimized **Momentum** indicators
are shown in their splitter window. Then a set of **Trading Rules** is
specified using the **Trading Rules Filtering & Momentum Weights**
dialog, and finally the **Trading Rules **are tested in a
simulated trading scenario using the **Diagnostic Test**. This test
uses the default **Momentum** indicators, which coincide with the **Harmonic
Oscillator** indicators (**Smoothing time scale** equal to the **time
horizon**).

**Demo: Momentum Indicators -
No Smoothing** (Revised October 7, 2006) This demonstration
investigates the design and testing of a set of **Momentum** indicators in
which the **Savitzky-Golay** smoothing filter is not used. In this
case, the **Velocity** indicator is just the **Price Projection (returns)**
itself, so this is equivalent to a test of the **Price Projection**.
The **Relative Price** and **Acceleration** type indicators are also
tested. A set of three **Momentum** indicators consisting of the **Relative
Price**, **Velocity -- Price Projection (returns)**, and another **Velocity**
indicator is finally chosen. Then these three indicators are tested using
the **Diagnostic Test** for four different trading scenarios. The **time
horizon** for all these tests is set to **20 days**, the optimum value for
this particular stock -- **MSFT**.

**Demo: Periodogram and Wavelet
Spectrum** (Revised June 27, 2006) This demonstration
describes the **Periodogram **and** Wavelet Spectrum** tests and their
meaning and interpretation. The **Periodogram** and** Wavelet Spectrum**
are the basis of two of the **Linear
Prediction **filters used for the **Price Projection** in ** QuanTek**.
The filter based on the

**
Demo:
Fractional Difference Filter** (Revised May 28, 2005)
This demonstration tests the **Fractional Difference Filter**
in * QuanTek* using the

**Demo: Trend Persistence and
Smoothness** (Revised May 26, 2005) This demonstration
illustrates the "ideal" behavior of the ** QuanTek** trading
rules and indicators, on the assumption that the "signal" is a slowly
varying, smooth function buried in stochastic noise. This is illustrated
by means of an artificial

**Demo: Features of
the Main Graph ( StockEval)**
(Posted December 1, 2001) This describes the special features of the

**Demo: Harmonic
Oscillator Indicator ( StockEval)**
(Posted December 1, 2001) This describes the three parts of the

**Demo: Diagnostic Test of Returns ( StockEval)**
(Posted December 1, 2001) This describes the results of the

*Go back to ***Home Page**