Upload
bzinchenko
View
100
Download
4
Embed Size (px)
Citation preview
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
Quant Trade®
Precision Oscillator Suite
Indicators for Bloomberg Professional
Quant Trade Technologies, Inc. 2 N Riverside Plaza Suite 2325, Chicago, Illinois 60606
http://www.quant-trade.com/
Document ID:
Precision Oscillator Suite.doc
General information:
Version: 1.2 State of document: Status: released
Last modified: 07 June 2014 Created: 21 May 2014
Project: Precision Suite Author: Boris Zinchenko Project lead: Boris Zinchenko
Recipients:
Accompanying documents:
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 1/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
Proof of change
Change Modified section(s) Log message ID # Date Version
1 29 May 2014 1.0 All Document created BZ 2 07 June 2014 1.2 Price Beam Reflects new indicator EL
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 2/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
Table of contents
1 Abstract 4 2 Analysis 5
2.1 Introduction ...................................................................................... 5 2.2 Background ...................................................................................... 5 2.3 Raw Data Streams ............................................................................. 6 2.4 Stable Price Aggregates ..................................................................... 7 2.5 A Solution ........................................................................................ 7 2.6 Summary ......................................................................................... 7
3 Technical indicators 9 3.1 Price Beam ....................................................................................... 9
3.1.1 Setting Chart Background ....................................................................... 9 3.1.2 Preparing Chart Template ...................................................................... 11 3.1.3 Sharing Chart with Colleagues ................................................................ 11
3.2 Precise Average .............................................................................. 13 3.3 Precise Bollinger Bands .................................................................... 14 3.4 Precise Exponential Moving Average (Precise EMA) .............................. 15 3.5 Precise Forecast Oscillator (Precise FOSC) .......................................... 16 3.6 Precise Linear Regression ................................................................. 17 3.7 Precise Moving Average Convergence/Divergence (MACD) .................... 18 3.8 Precise Percentage Price Oscillator (PPO) ............................................ 19 3.9 Precise Price Oscillator ..................................................................... 20 3.10 Precise Rate-of-Change (ROC) .......................................................... 21 3.11 Precise Standard Moving Average (SMA) ............................................ 22 3.12 Precise Standard Deviation ............................................................... 23 3.13 Precise Standard Error ..................................................................... 24 3.14 Precise Triple Exponential Moving Average .......................................... 25 3.15 Precise Triangular Moving Average .................................................... 26 3.16 Precise Triple Exponential Average .................................................... 27 3.17 Precise Time Series Forecast ............................................................. 28 3.18 Precise Weighted Moving Average (WMA) ........................................... 29
References 30
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 3/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
1 Abstract We offer a simple and powerful collection of improved technical indicators
inherited from classical oscillators widely used throughout modern technical
analysis. These oscillators take advantage of full intra-bar information
provided by the Bloomberg charting package. They allow for up to four times
more precision against their classical counterparts.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 4/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
2 Analysis
2.1 Introduction A standard bar on a price chart typically has four channels: open, high, low
and close. Many technical indicators use only one of these channels for the
statistical inference (analysis). In the image below, we illustrate the benefits
of using all four channels to improve trading signal precision and illustrate it
on a simple example of a Bollinger Bands calculation. In this suite we include
all of the indicators in the figure below, as well as reconstruct many other
indicators on the same improvement principle.
The figure illustrates a simulated intra-bar probability channel associated
with price components provided on the chart. The blue price gradient depicts
the associated probability levels extrapolated from the price sample within
each bar. Over this price distribution we plot Bollinger bands calculated by
the standard method against the improved bar price resolution method
(blue, yellow, red.) One can see that the bands using full price information
are more smooth and consistent with the price channel.
2.2 Background Here we present a new method of applying trading indicators to financial
charts. By combining classical statistics with uniquely modified chart
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 5/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1 indicators, you will be able to understand classical charts from the viewpoint
of random processes.
All of you are familiar with trading charts. As a common tool to traders, we
find a myriad of variations from the very basic to the exotic. Regardless of
the chart used, the goal is the same. You want to have as much useful
information as possible on your chart so you can make good trading
decisions.
While financial charts can be vivid and attractive, they tend to contradict
modern statistics. The charts and the indicators used with them can
introduce unstable measures increasing the risk of your trades. Here we will
explain some of the pitfalls of typical charting from a simple mathematical
point of view. We will then introduce a variation on classical trading
indicators that help to overcome these shortcomings.
2.3 Raw Data Streams Pure trading data for any market instrument comes from exchanges in the
form of bid & ask prices in unison with a last traded price. This is high-
frequency data, which you can watch on a real-time trading terminal. Most
traders rarely watch, archive, and analyze this data for the simple reason
that it is too abundant. Even modern computers have difficulty in analyzing
and archiving this granular data.
Most data vendors, software developers, and traders consider raw data to be
noisy, excessive, and exorbitant. The truth of the matter is however, that
more raw data makes statistical analysis more reliable. This is particularly
true when trying to develop a rule or algorithm for predicting the future. This
is why many types of predictive tools fall short of desired performance in
market research. In many cases it is not the model, but the amount and
type of data that is used.
One simple example of how data can affect analysis is made clearly evident
with an example of typical data from a bar chart. Usually this data is painted
on the screen with one time stamp and open, high, low, close (OHLC) data.
Any experienced trader will recognize that a single time stamp for these four
data points is not accurate. In fact, most traders tolerate it as a limitation of
the equipment and tools they are using. If we were to time stamp each price
separately, we would instantly have much more useful data that can
generate more statistically accurate results. To make matters worse, most
technical indicators are designed to use one price, such as the closing price
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 6/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1 or some mathematical derivation of multiple prices using simple arithmetic
averages.
2.4 Stable Price Aggregates In regular statistics, a moving average is simple and not very useful for
aggregate data. However, when you face the danger of misshaped
aggregate data, you need to find something stable to rely on. One possible
solution is to replace the original data with a more stable average
approximation. This is better, even at the cost of some loss in data
precision.
In trading, precision is the key to success, so how can we improve the
resolution of standard price data? The best statistical practice suggests using
a median as an accurate measure of the middle point. The median is derived
from nonparametric statistics and appears more robust on data likely to
have anomalies in distribution.
2.5 A Solution One way to improve precision without acquiring additional data is to use all
available data from the provider and to smooth the data. So in the case
above, we take all four data points lumped under one time stamp and sum
up averages over several bars. Not only will you have four times the amount
of data, but the relative role of high and low prices in the average will
steadily diminish. The reason is that with many values, the high and low
prices fall within the range of a cumulative sample of several bars taken
together.
2.6 Summary We have considered a number of issues that compromise the accuracy of
price data and statistical analysis that most traders use. To alleviate some
of these issues, we have discussed how standard open, high, low, close price
data can be interpreted as a price probability channel. Using this method, it
is possible to improve the precision of standard statistical trading indicators
by modifying how they calculate the available data under one time stamp.
The Precision Oscillator Suite is a composite of seventeen classical statistical
indicators that have been re-engineered to reflect the price probability
channel. In addition to increased precision, you will find that these indicators
smooth price data much more efficiently. These indicators are designed for
the Bloomberg trading platform. You may insert and manipulate them into
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 7/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1 any given chart in the same fashion that you would insert any built-in
Bloomberg indicator.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 8/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3 Technical indicators In the following sections we shortly describe each modified indicator as
compared to its classical counterpart. These new variations are found in the
Bloomberg package. Each indicator begins with the word, “Precise” and is
labeled as a Quant Trade indicator.
3.1 Price Beam Precise Price Beam is the fundamental basic indicator that constitutes the
core foundation of all other indicators in this suite. It visually shows the
stochastic distribution of the underlying price points in the form of a visual
heat map. Lightly colored regions reveal a higher concentration of the price
probability function, while darker regions correspond to less likely price
distribution ranges.
Price points are more concentrated in the brighter high probability channels
and gradually fade out to less probable dark outer regions where a trade is
highly unlikely to occur. Due to this feature, the Precision Price Beam
indicator should always appear on a uniform black background. Black
designates a uniform outer field of zero probability, on which the channel of
positive probabilities is depicted.
3.1.1 Setting Chart Background A user must take specific steps to ensure the proper chart background:
First, select menu “Edit” (97) and pick the option “Chart Colors/Styles”
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 9/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
The chart “Property Manager” will appear
Take the following steps:
1. Select the “Color/Style” tab.
2. Select “Solid” option on “Chart Background.”
3. Select “Black” as the background color.
4. Select a color for the bars on the chart. We recommend “Green” for
the up bar color and “Red” for the down bar color.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 10/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1 5. Hit “Update” to confirm.
Upon completion of these steps, the Precision Price Beam indicator will now
appear on the proper background.
3.1.2 Preparing Chart Template To avoid repeating the above steps for every chart, you may want to create
a predefined template with these settings and then reuse it with charts for
other symbols.
1. Go again to the “Edit” (97) menu.
2. Pick the option “Create Theme From Chart.”
3. Select a name for new theme. For instance, “Price Beam.”
4. Confirm by “Save Theme.”
You can now select this theme for another chart and also set it as a default.
3.1.3 Sharing Chart with Colleagues You may want to share the template with your colleagues. To do so, go to
chart “Actions” menu (96) and select “Share” item.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 11/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
Now the chart sharing dialog will appear. Fill out the fields and hit the
“Send” button when you are ready. The chart will now be shared.
You may pick “Firm” in the “Share with” field to share company wide.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 12/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.2 Precise Average Precise Average is the average of current bar using all price channels.
Settings:
Classical analog: Typical price
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 13/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.3 Precise Bollinger Bands Precise Bollinger Bands uses “Precise Standard Deviation” and “Precise
Average” for more precision of the standard Bollinger Band calculation.
Settings:
Classical analog: Bollinger Bands
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 14/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.4 Precise Exponential Moving Average (Precise EMA) Precise Exponential Moving Average has improved precision due to use of all
four price channels on the chart as one sequence for averaging.
Settings:
Classical analog: Exponential Moving Average
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 15/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.5 Precise Forecast Oscillator (Precise FOSC) Precise Forecast Oscillator compares the actual price to the time series
forecast and calculates a percentage between negative 100% and positive
100%. This version uses “Precise Linear Regression” as the forecast base.
Settings:
Classical analog: Forecast Oscillator (FOSC)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 16/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.6 Precise Linear Regression Precise Linear Regression is the extension of the linear regression based
indicators calculated across all price channels to improve precision of the
regression.
Settings:
Classical analog: Linear Regression
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 17/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.7 Precise Moving Average Convergence/Divergence (MACD) Precise Moving Average Convergence/Divergence (MACD) is the trend
following momentum based on two “Precise Average” indicators.
Settings:
Classical analog: Moving Average Convergence/Divergence (MACD)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 18/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.8 Precise Percentage Price Oscillator (PPO) Precise Percentage Price Oscillator (PPO) is based on two “Precise Average”
indicators expressed as a percentage.
Settings:
Classical analog: Percentage Price Oscillator (PPO)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 19/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.9 Precise Price Oscillator Precise Price Oscillator shows the variation between two “Precise - Averages”
for the price of a security.
Settings:
Classical analog: Price Oscillator
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 20/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.10 Precise Rate-of-Change (ROC) Precise Rate-of-Change (ROC) is the difference between “Precise Average”
indicators on the current bar and a previous bar a number of periods ago.
Settings:
Classical analog: Rate-of-Change (ROC)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 21/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.11 Precise Standard Moving Average (SMA) Precise Standard Moving Average (SMA) is the Standard Moving Average
calculated on all bar channels.
Settings:
Classical analog: Standard Moving Average (SMA)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 22/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.12 Precise Standard Deviation Precise Standard Deviation is the Standard Deviation calculated on all price
channels.
Settings:
Classical analog: Standard Deviation
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 23/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.13 Precise Standard Error Precise Standard Error is Standard Error shows how far prices oscillate
around a linear regression line. Linear regression is calculated on all price
channels.
Settings:
Classical analog: Standard Error
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 24/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.14 Precise Triple Exponential Moving Average Precise Triple Exponential Moving Average is the combination of a single,
double and triple “Precise Exponential Moving Average” indicator.
Settings:
Classical analog: Triple Exponential Moving Average
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 25/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.15 Precise Triangular Moving Average Precise Triangular Moving Average (TMA) is a double smoothed (i.e.
averaged twice) weighted moving average using all price channels.
Settings:
Classical analog: Triangular Moving Average (TMA)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 26/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.16 Precise Triple Exponential Average Precise Triple Exponential Average (TRIX) displays the percentage Rate of
Change (ROC) of a triple EMA over all price channels.
Settings:
Classical analog: Triple Exponential Average (TRIX)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 27/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.17 Precise Time Series Forecast Precise Time Series Forecast (TSF) calculates probable future values for the
price by fitting a linear regression line over a given number of price bars and
following that line forward into the future. It is calculated across all price
channels.
Settings:
Classical analog: Time Series Forecast (TSF)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 28/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
3.18 Precise Weighted Moving Average (WMA) Precise Weighted Moving Average (WMA) is average value of a security's
price over a period of time with special emphasis on the more recent
portions of the time period. It is calculated across all price channels.
Settings:
Classical analog: Weighted Moving Average (WMA)
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 29/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1
References
[1] Boris G. Zinchenko, Unraveling the Mystery of Stock Prices, Stocks & Commodities, March 2010 [2] Mosteller, Frederick and Rourke, Robert E.K. Sturdy Statistics: Nonparametrics and Order Statistics Reading, MA: Addison-Wesley, 1973. [3] L. D. Landau, L. M. Lifshitz, Quantum Mechanics: Non-Relativistic Theory, Volume 3, Third Edition, Elsevier, 2003
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 30/31
Trading Algorithms Fractal indicators, predictors and trading strategies
Version 1.1 Legal notice The product names used in this document are for identification purposes only. All trademarks and registered trademarks are the property of their respective owners. The following trademarks may or may not be marked in this document: Quant Trader is a trademark or registered trademark of Quant Trade Technologies, Inc. in United States and/or other countries. Other company, product, and service names may be trademarks, registered trademarks, or service marks of other owners.
Quant Trade Technologies, Inc. ■ October 11, 2014 ■ 31/31