VanAurum Logo

Documentation


Welcome to VanAurum's Documentation Page

This page contains information, tutorials, and glossaries that explain the content on this site and how to use it. We're constantly updating this page as content changes, new functionality is added, and as the site evolves.

Still have questions? If the answers you are looking for aren't on this page, don't hesitate to e-mail us at support@vanaurum.ai.


Documentation


Table of Contents




What is VanAurum?


Introduction:

This is the second generation of the VanAurum platform. The first platform was our first stab at using machine learning to create high impact market analytics. The first platform was very much "display-based", meaning we generated machine learning analytics on the back-end and posted them to the site daily.

Since then, we have learned a lot about what's needed to build a scalable machine learning platform. This version of VanAurum attempts to integrate as many of those learnings as possible. That being said, this version of the platform was built with the next version in mind. Our ultimate goal at VanAurum is to create an autonomous AI-based financial platform that can help everyone become a more informed investor, trader, or analyst. In a world where financial opinions are cheap, I want to make access to financial facts even cheaper. I think machine learning can help us get there.

What does it do, and who should use it?

VanAurum seeks to do two things well:

  1. Quickly assign quantifiable historical context to any market event.
  2. Use machine learning to synthesize superior technical indicators
There's a lot there, so let's unpack that a bit.

Most analysts follow market indicators or technical indicators that they feel best describe market movements. A piece that's often missing, or difficult to come by, is the meaning and implication of specific technical events. For example, suppose net speculalive positions in Gold have just reached extreme lows. One next logical question might be, "What has that meant in the past?". VanAurum can quickly quantify that for you. We collect as much financial information we can, and the number of market relationships you can quantify increases exponentially with the addition of new data.

VanAurum is designed for traders, investors, and analysts that want to make decisions rooted in quantifiable facts.



VanAurum Insights


Introduction:

We've been building a financial database since we first launched the inaugural version of our platform in 2017. In order to build scalable machine learning systems for financial analysis you need high quality data sources, and a database schema that is conducive to quickly performing cross-asset analysis. Most of the effort surrounding the inaugural version of our platform was (and is) centered around building a more reliable and scalable data architecture. VanAurum Insights is a SQL interface that we've designed to allow members to gain financial insights quickly and easily. Billions of permutations of historical market conditions can be analyzed and quantified with concise analytics.



Example: Simple cross-asset analysis between Gold and the US Dollar
Our database is designed so that our machine learning system can quickly build training datasets that compare return profiles against market action in other sectors. VanAurum Insights is designed so that our members can obtain the exact same insights - on demand. Consider the following situation where one would like to know the 42-day returns for Gold when the US Dollar's RSI indicator has been above 70.


The analytics from VanAurum Insights are broken down into sections:

  • Search description: This is an auto-generated natural language description of the search query that gets assembled by the user. This serves two main functions:
    1. Provide feedback to the user to aid in using the tool, and
    2. Build a training dataset for the natural langage to SQL (NL-SQL) machine learning system that will launch with our third version of the platform. The auto-generated descriptions provide high quality examples in the training dataset from which noise can be added to assist with generalization.
  • Histogram of returns: Because your search query represents a subset of the total set of returns, we present the statistical distributions for the entire set of returns and the returns belonging only to your query.
  • Recent occurrences: We provide you with the number of instances of your query within the last three years. This information is overlaid on a price chart of your asset of interest.
  • Summary Statistics: With each query, we provide the following summary statistics:
    1. Number of occurrences: This is the number of trading days in our price history where your queried condition asserted itself. This is displayed alongside the full number of trading days for that asset in our database.
    2. Mean value of return set: Shows the average return over your specified return period when your queried market condition asserted itself. This is compared against the average return for your chosen period for the entire history of the asset.
    3. Standard deviation: Standard deviation of the returns (of your specified period) for your query and the full history of the data.
    4. Minimum return: The minimum return over your specified period for both your specified query and the full history of returns.
    5. 25th percentile: The level at which 25% of the total occurences are below, and 75% are above.
    6. 50th percentile: The level at which 50% of the total occurences are below, and 50% are above (Median).
    7. 75thth percentile: The level at which 75% of the total occurences are below, and 25% are above.
    8. Maximum return: The minimum return over your specified period for both your specified query and the full history of returns.
  • T-Statistic: Presented as a signal-to-noise ratio for your query. Compares the mean returns of your specified market condition with the mean returns of the full price history, adjusing for the number of occurrences and standard deviation.

Responsive image

Using the SQL interface
The main area of the SQL interface has six boxes - the minimum number of entries required to get analytics from the database. You do not need to worry about what you can pick and choose - anything you see in the drop-down menus is a valid analysis option. In this section we will walk you through what these boxes mean, and how to use them to build your personal searches.


Example 1: Using the SQL interface for single-asset analysis
In this example we're going to look at finding single-asset analytics for Bitcoin. What do we mean by "single-asset"? This means we're going to find analytics for the asset of interest using its own technical indicators, rather than the indicators from another asset or other economic data.

Below is a screenshot of a complete query, and the natural language description that we provide as feedback. The six boxes of the main search area are as follows:

  1. Asset of interest: This is the first box from the left ('BITCOIN_USD' in this example). This is the asset whose return profile you're looking for analysis on.
  2. Return Period: This is the return period you are interested in ('RETURNS5' in this example, meaning 5-day returns). Return periods ranging from 1 day to 252 days (one year) are available.
  3. Primary comparison asset: This is the asset that you want to compare your primary asset against. In this example, because we are comparing Bitcoin against its own technical indicators, you choose the same asset. If you were wanting to compare Bitcoin against the technical indicators of another asset or economic data point you would make that selection here.
  4. Primary comparison field: All of the technical indicators and variants that are available for your comparison asset will appear in this list. In this example we are using '10DMA_50DMA_RATIO', which is the value of the 10-day simple moving average divided by the 50-day simple moving average.

  5. Primary comparison operator: Specifies how you want to split the technical condition. In this example, we want all 5-day returns where the 10-day moving average is above the 50-day moving average (the ratio is greater than or equal to 1.0)
  6. Primary comparison value: The value of the comparison field that you want to compare against.
Responsive image


Example 2: Using the SQL interface for cross-asset analysis
In this example we are going to look at finding cross-asset analytics for Gold. What do we mean by "cross-asset"? This just means we are going to find analytics for the asset of interest using the indicators from another asset or economid data point. In this example we are going to find the average one-quarter (63 trading days) return for gold when the USD Dollar's 20-day RSI indicator had a reading above 70.

Below is a screenshot of a complete query, and the natural language description that we provide as feedback. The six boxes of the main search area are as follows:

  1. Asset of interest: This is the first box from the left ('BITCOIN_USD' in this example). This is the asset whose return profile you're looking for analysis on.
  2. Return Period: This is the return period you are interested in ('RETURNS5' in this example, meaning 5-day returns). Return periods ranging from 1 day to 252 days (one year) are available.
  3. Primary comparison asset: This is the asset that you want to compare your primary asset against. In this example, because we are comparing Bitcoin against its own technical indicators, you choose the same asset. If you were wanting to compare Bitcoin against the technical indicators of another asset or economic data point you would make that selection here.
  4. Primary comparison field: All of the technical indicators and variants that are available for your comparison asset will appear in this list. In this example we're using '10DMA_50DMA_RATIO', which is the value of the 10-day simple moving average divided by the 50-day simple moving average.

  5. Primary comparison operator: Specifies how you want to split the technical condition. In this example, we want all 5-day returns where the 10-day moving average is above the 50-day moving average (the ratio is greater than or equal to 1.0)
  6. Primary comparison value: The value of the comparison field that you want to compare against.
Responsive image




The VanAurum AI Sentiment Index (VaiSi)


Introduction:

The VanAurum AI Sentiment Index is the most popular of our anlytics from the inaugural platform. We have enhanced this analysis and made it more scalable on our end so that we can offer VaiSi analytics for ten times as many assets (and growing).

VanAurum builds training datasets from our financial database and then performs an iterative ranking and "feature reduction" process. VanAurum ranks the top ten indicators that, historically, have best delineated positive and negative market returns over periods ranging from 21 days to 84 days (four months). VanAurum then synthesizes a new market oscillator using a constrained optimization algorithm (more on this below) that assigns weights to each indicator. The result is VaiSi.

The objective of this process to create a technical picture of the market that is numerically rooted and unbiased by emotion, to the greatest extent possible.


Example: VaiSi Indicator for Gold

The analytics from VaiSi are broken down into two graphs with a total of four sections:
  • Top graph (VaiSi Universe): This bar chart of the VaiSi values for all the asset's in VanAurum's database, ranked highest to lowest:
    • Red bars indicate assets with VaiSi values above the 80 threshold. VanAurum considers these assets to be overbought.
    • Blue bars indicate assets with VaiSi values between 20 and 80.
    • Green bars indicate assets with VaiSi values between 0 and 20. VanAurum considers these assets to be oversold.
    • This is intended to provide the user with a high-level snapshot of VanAurum's assessment of its universe of assets.
  • Bottom graph (Asset specific VaiSi analytics):
    • Top Section: This is a chart of the closing price of the asset chosen from the user in the dropdown menu.
    • Middle (Heat map): This area is a heatmap of the indicators ranked highest by VanAurum's machine learning algorithms. These indicators can, and likely do, vary for each asset. The dark blue areas represent an extremely oversold reading for a particular indicator. Bright yellow indicates an overbought reading for for that particular indicator. Where indicators selected by VanAurum are not numerically bound (i.e, between 0 and 100) VanAurum converts the datapoint to a ranked percentile adjusted for statistical outliers. The intention of this area is to provide the user with visual feedback for the ten indicators deemed most significant by VanAurum.
    • Bottom: The bottom section is the VaiSi indicator synthesized by VanAurum using the top ranked indicators in the middle section. Note that VaiSi indicators are slow moving indicators and are not suitable short trading periods. The intention of the VaiSi indicator is to provide a high level of indication of when a market is likely to enter a multi-month rally or correction. The performance, pace, and behaviour of VaiSi indicators will likely vary for each asset.

Responsive image Responsive image


Committment of Traders Reports


Introduction:

Every week we generate analytics from the CFTC Committment of Traders reports. We do not cover every contract, but we cover the core reports that align with our asset coverage list. The analytics we generate for the COT report are "conventional" analytics - no machine learning here. We generate them as a standalone report because in our experience the data derivatives from these reports are consistently ranked high by our machine learning algorithms.

Example: Gold Committment of Traders Report

The analytics from VanAurum Insights are broken down into sections:

Data Series Descrption

Details


Total speculator long / commercial short contracts
Total speculator long contracts is the sum of all speculative positions, large and small.

Net Speculator Positions as % of Open Interest
Calculated as (Speculator long positions - Speculator short positions)/(Open Interest). This is a great datapoint because it normalizes for changes in the average open interest over time.

Total Speculator Shorts
Sum total of all speculator short positions, large and small.

Change in Speculator Longs
Week-over-week change in the total number of speculator long positions.

Total speculator long / commercial short contracts
Week-over-week change in the total number of speculator short positions

Blees Differential
The Blees Differential is calculated by subtracting the Blees Long Rating from the Blees Short Rating. The Blees Long Rating is calculated by normalizing the most recent speculator long positions against the 78-week (1.5 years) maximum and 78-week minimum number of contracts. A Blees Long rating of 100 means that speculator long contracts are the highest they have been in 78 weeks. Conversely, a reading of 0 would indicate that speculator long contracts are the lowest they have been in 78 weeks. The same is calculated for speculator short positions.

Blees long and short ratings are bound between 0 and 100. Blees Differential can therefore be within a range of -100 to +100. A reading of +100 would indicate extreme bearishness, and -100 would indicate extreme speculator bullishness. A reading of +100 would indicate that speculator shorts are at 78-Week highs, while at the same time speculator longs are at 78-Week lows.

Change in Open Interest
Week-over-week change in open interest

Net Speculator Positions as Percentile
Net speculative positions divided by open interest, ranked as a percentile.

Responsive image



Description of assets and their naming conventions


How we choose our assets:

At VanAurum we believe the 80/20 rule, to some extent, applies to all systems - natural and synthetic. By that logic, there must be some aspects of the “active market” that have a disproportionate level of influence over others. Extending that one step further, we think a good way to describe an “active market” is by trading activity and history. Accordingly, if you are going to start training a machine learning system for financial analysis, our view is that the most actively traded assets (by volume) and the duration of their trading history are good places to begin.

By being selective in this manner, with a relatively small portion of market assets you can analyze a disproportionately large percentage of overall market activity. This has a secondary benefit of alleviating the data analysis burdens on the machine learning system. You also have other constraints with your data, like quality - in machine learning, as with other things, garbage in equals garbage out.

Master Data List (Updated Daily)


VanAurum is currently tracking 147 data feeds that cover the majority of the largest international markets. This is broken down accordingly:


  • CFTC Trader Reports: 16
  • Commodities: 22
  • Sector ETFs: 30
  • Major Indices: 10
  • Currencies: 10
  • Ratios: 21
  • Key Interest Rates: 25
  • Economic Data Points: 2
  • Yield Curves: 5

Asset Code

Asset Type

Description


ALUMINUM_D

Commodity

Continuous futures contract price for Aluminum


AUD_D

Currency

Continuous futures contract price for the Australian Dollar


BITCOIN_D

Cryptocurrency

Mid pricing levels for continuous Bitcoin prices


CAD_D

Currency

Continuous futures contract price for the Canadian Dollar


CHF_D

Currency

Continuous contract price for Swiss Franc futures


COFFEE_D

Commodity

Continuous contract price for coffee futures


COPPER_COT_FO

Committment of Traders

CFTC trader positioning analytics for Copper


COPPER_D

Commodity

Continuous contract price for copper futures


CORN_COT_FO

Committment of Traders

CFTC trader positioning analytics for Corn


CORN_D

Commodity

Continuous futures contract price for Corn


COTTON_COT_FO

Committment of Traders

CFTC trader positioning analytics for Cotton


COTTON_D

Commodity

Continuous contract price for cotton futures


DJIA_D

Index

Continuous futures contract price for the Dow Jones Industrial Average


EEM_D

Sector ETF

iShares Emerging Markets Index Fund ETF


EFA_D

Sector ETF

iShares MSCI EAFE Index Fund ETF


EM_HY_CREDIT_YIELD_D

Interest Rate

emerging market high yield corporate bond yields


E-MINI_RUSSELL_2000_COT_FO

Committment of Traders

CFTC trader positioning analytics for the MSCI Russell 2000 E-mini contract


E-MINI_SPX_500_COT_FO

Committment of Traders

CFTC trader positioning analytics for the S&P500 E-mini contract


ETHEREUM_D

Cryptocurrency

Mid pricing levels for continuous Ethereum prices


EURODOLLAR_D

Currency

Continuous contract price for Eurodollar futures


EURO_FX_D

Currency

Continuous contract price for Euro FX Futures


EURO_STOXX_50_D

Index

Continuous contract price for Euro Stoxx 50 index futures


EWA_D

Sector ETF

iShares MSCI Australia ETF


EWI_D

Sector ETF

MSCI iShares Italy ETF


EWL_D

Sector ETF

MSCI iShares Switzerland ETF


FEDFUNDS_30_D

Economic Data

the 30-Day Federal Reserve Funds Rate


FRZN_CONCENTRATED_ORANGE_JUICE_COT_FO

Committment of Traders

CFTC trader positioning analytics for frozen orange juice futures contracts


FTSE100_D

Index

Continuous contract price for FTSE 100 futures


FXI_D

Sector ETF

the iShares China large cap stock index ETF


GASOIL_D

Commodity

Continuous contract price for gas oil futures


GASOLINE_D

Commodity

Continuous contract price for gasoline futures


GBP_D

Currency

Continuous futures contract price for the British Pound


GDX_D

Sector ETF

the VanEck senior gold miners ETF


GDXJ_D

Sector ETF

the VanEck junior gold miners ETF


GOLD_COT_FO

Committment of Traders

CFTC trader positioning analytics for Gold


GOLD_D

Commodity

Continuous futures contract price for Gold


HEATING_OIL_D

Commodity

Continuous contract price for heating oil futures


HYG_D

Sector ETF

iShares High Yield Corporate Bond ETF


IBB_D

Sector ETF

the iShares Nasdaq Biotechnology sector ETF


ISRA_D

Sector ETF

VanEck Israel stock market index ETF


IYT_D

Sector ETF

iShares Dow Transportation ETF


JAPANESE_YEN_COT_FO

Committment of Traders

CFTC trader positioning analytics for japenese yen futures contracts


JPY_D

Currency

Continuous contract price for Japanese Yen futures


LEAN_HOGS_D

Commodity

Continuous contract price for lean hog futures


LIT_D

Sector ETF

GlobalX Lithium and Battery Technology Stock ETF


LIVE_CATTLE_COT_FO

Committment of Traders

CFTC trader positioning analytics for live cattle futures contracts


LIVE_CATTLE_D

Commodity

Continuous contract price for live cattle futures


MEXICAN_PESO_D

Currency

Continuous contract price for the Mexican Peso


NASDAQ_100_COT_FO

Committment of Traders

CFTC trader positioning analytics for Nasdaq 100 futures contracts


NASDAQ100_D

Index

Continuous contract price for Nasdaq 100 futures


NATURAL_GAS_D

Commodity

Continuous contract price for natural gas futures


NIKKEI225_D

Index

Continuous contract price for Nikkei 225 index futures


NZD_D

Currency

Continuous contract price for New Zealand Dollar futures


OATS_D

Commodity

Continuous futures contract price for the Dow Jones Industrial Average


ORANGE_JUICE_D

Commodity

Continuous contract price for orange juice futures


PALLADIUM_COT_FO

Committment of Traders

CFTC trader positioning analytics for Palladium futures


PALLADIUM_D

Commodity

Continuous futures contract price for Palladium


PLATINUM_COT_FO

Committment of Traders

CFTC trader positioning analytics for Platinum futures


PLATINUM_D

Commodity

Continuous futures contract price for Platinum


PLND_D

Sector ETF

VanEck Poland stock market index ETF


RATIO_COPPER_GOLD_D

Ratio

The copper-to-gold price ratio


RATIO_FXI_SPX_D

Ratio

The ratio of the Chinese Large Cap Stock ETF to S&P500 ratio.


RATIO_GDX_GOLD_D

Ratio

The GDX-to-Gold price ratio


RATIO_GDX_SPX_D

Ratio

The ratio of the VanEck Senior Gold Miner ETF to S&P500 ratio.


RATIO_GOLD_SILVER_D

Ratio

The gold-to-silver price ratio


RATIO_GOLD_UST10Y_D

Ratio

The ratio of Gold to 10-Year US Treasury Bonds.


RATIO_GOLD_WTI_D

Ratio

The gold-to-WTI crude oil price ratio


RATIO_HYG_GASOLINE_D

Ratio

The ratio of the HYG High Yield Corporate Credit ETF to Gasoline.


RATIO_HYG_WTICRUDE_D

Ratio

The ratio of the HYG High Yield Corporate Credit ETF to WTI Crude Oil.


RATIO_IBB_SPX_D

Ratio

The ratio of the Biotechnology Sector ETF to S&P500 ratio.


RATIO_IYT_DJIA_D

Ratio

The ratio of the IYT iShares Dow Transportation ETF to DJIA.


RATIO_PLAT_COPPER_D

Ratio

The ratio of Platinum to Copper.


RATIO_PLAT_GOLD_D

Ratio

The ratio of Platinum to Gold.


RATIO_PLAT_SILVER_D

Ratio

The ratio of Platinum to Silver.


RATIO_RSX_SPX_D

Ratio

The ratio of the Russian Large Cap Stock ETF to S&P500 ratio.


RATIO_SIL_SILVER_D

Ratio

The Silver miners to Silver price ratio


RATIO_SPX_GOLD_D

Ratio

The S&P500-to-gold price ratio


RATIO_SPX_UST10Y_D

Ratio

The ratio of the S&P500 to 10-Year US Treasury Bonds.


RATIO_TLT_WTICRUDE_D

Ratio

The ratio of the TLT 20+ Year Treasury Bond ETF to WTI Crude Oil.


RATIO_VDE_SPX_D

Ratio

The ratio of the Vanguard Energy Sector ETF to S&P500 ratio.


RATIO_VDE_WTICRUDE_D

Ratio

The ratio of the VDE Energy ETF to WTI Crude Oil.


ROUGH_RICE_D

Commodity

Continuous futures contract price for rough rice


RSX_D

Sector ETF

VanEck Russian stock market index ETF


RUSSEL1000_D

Index

Continuous contract price for Russel 1000 index futures


RUSSEL2000_D

Index

Continuous contract price for Russel 2000 index futures


SIL_D

Sector ETF

the GlobalX Funds Silver Miner ETF


SILVER_COT_FO

Committment of Traders

CFTC trader positioning analytics for Silver


SILVER_D

Commodity

Continuous futures contract price for Silver


SOY_BEANS_COT_FO

Committment of Traders

CFTC trader positioning analytics for soy bean futures contracts


SOYBEANS_D

Commodity

Continuous futures contract price for soybeans


SPX_D

Index

Continuous futures contract price for the S&P 500


SUGAR_D

Commodity

Continuous contract price for sugar futures


TLT_D

Sector ETF

iShares 20+ Year Treasury Bond ETF


TSX60_D

Index

Continuous contract price for TSX 60 futures


US10YR_TREASURY_REAL_YIELD_D

Interest Rate

the 10-year inflation-adjusted US treasury yield


US10YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 10-year US Treasuries


US1MO_TREASURY_REAL_YIELD_D

Interest Rate

the 1-month inflation-adjusted US treasury yield


US1MO_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 1-month US Treasuries


US1YR_TREASURY_REAL_YIELD_D

Interest Rate

the 1-year inflation-adjusted US treasury yield


US1YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 1-year US Treasuries


US20YR_TREASURY_REAL_YIELD_D

Interest Rate

the 20-year inflation-adjusted US treasury yield


US20YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 20-year US Treasuries


US2MO_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 2-month US Treasuries


US2YR_TREASURY_REAL_YIELD_D

Interest Rate

the 2-year inflation-adjusted US treasury yield


US2YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 2-year US Treasuries


US30YR_TREASURY_REAL_YIELD_D

Interest Rate

the 30-year inflation-adjusted US treasury yield


US30YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 30-year US Treasuries


US3MO_TREASURY_REAL_YIELD_D

Interest Rate

the 3-month inflation-adjusted US treasury yield


US3MO_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 3-month US Treasuries


US3YR_TREASURY_REAL_YIELD_D

Interest Rate

the 3-year inflation-adjusted US treasury yield


US3YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 3-year US Treasuries


US5YR_TREASURY_REAL_YIELD_D

Interest Rate

the 5-year inflation-adjusted US treasury yield


US5YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 5-year US Treasuries


US6MO_TREASURY_REAL_YIELD_D

Interest Rate

the 6-month inflation-adjusted US treasury yield


US6MO_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 6-month US Treasuries


US7YR_TREASURY_REAL_YIELD_D

Interest Rate

the 7-year inflation-adjusted US treasury yield


US7YR_TREASURY_YIELD_D

Interest Rate

Prevailing yields on 7-year US Treasuries


USD_D

Currency

Continuous futures contract price for the US Dollar


US_HY_CREDIT_YIELD_D

Interest Rate

high yield US corporate bond yields


US_INFLATION_YOY_D

Economic Data

Year-over-year US CPI inflation data


UST10Y_D

Treasuries

Continuous futures contract price for the 10-Year US Treasury Bond


UST2Y_D

Treasuries

Continuous futures contract price for the 2-Year US Treasury Bond


UST30Y_D

Treasuries

Continuous futures contract price for the 30-Year US Treasury Bond


UST5Y_D

Treasuries

Continuous futures contract price for the 5-Year US Treasury Bond


US_TREASURY_BONDS_COT_FO

Committment of Traders

CFTC trader positioning analytics for US Treasury Bonds


VCR_D

Sector ETF

the Vanguard consumer discretionary sector ETF


VDC_D

Sector ETF

the Vanguard consumer staples sector ETF


VDE_D

Sector ETF

Vanguard Energy Sector ETF


VHT_D

Sector ETF

the Vanguard health care sector ETF


VIS_D

Sector ETF

the Vanguard industrial sector ETF


VIX_COT_FO

Committment of Traders

CFTC trader positioning analytics for VIX futures


VIX_D

Index

Continuous futures contract price for the Volatility Index (VIX)


VNM_D

Sector ETF

VanEck Vietnam stock market index ETF


VNQ_D

Sector ETF

the Vanguard real estate sector ETF


VOX_D

Sector ETF

the Vanguard telecommunications sector ETF


VPU_D

Sector ETF

the Vanguard electric utilities sector ETF


VWO_D

Sector ETF

the Vanguard emerging markets stock index ETF


WHEAT_D

Commodity

Continuous futures contract price for wheat


WTI_CRUDE_OIL_D

Commodity

Continuous contract price for WTI Crude Oil futures


XLB_D

Sector ETF

SPDR Materials Sector ETF


XLF_D

Sector ETF

the SPDR financial sector ETF


XOP_D

Sector ETF

SPDR S&P Oil & Gas Explorers and Producers ETF


YC10Y2Y_D

Yield Curve

Spread between US 10-Year and 2-Year treasury bond yields


YC10Y3MO_D

Yield Curve

Spread between US 10-Year and 3-month treasury bond yields


YC10Y5Y_D

Yield Curve

Spread between US 10-Year and 5-Year treasury bond yields


YC1Y1MO_D

Yield Curve

Spread between US 1-Year and 1-month treasury bond yields


YC5Y1Y_D

Yield Curve

Spread between US 5-Year and 1-year treasury bond yields







Description of technical indicators and their naming conventions


Introduction:

In this section you will find descriptions and explanations for the market indicators used on this site and our philosophy behind a machine learning process called 'feature engineering'. VanAurum is designed to accelerate the process of technical analysis for our members. Because of this, most of our market indicators focus on refinements or interpretations of other technical indicators. In machine learning, this is a process called 'feature engineering'.



Feature Engineering:
Feature engineering is the process of creating new indicators by performing simple operations on existing indicators. These types of metrics exist everywhere - as it turns out, people are pretty good at making them. For example, in the health field there is a metric called the Body Mass Index (BMI). The BMI, in the field of machine learning, could be considered an 'engineered feature'. Why? Well, BMI is just a person's weight divided by the square of their height. When doctors are assigning risk levels for things like diabetes and heart disease, the BMI is a much more useful metric than a person's height or weight alone. The hybrid metrics often carry more information than their constituent parts.

In machine learning, these hybrid metrics are very useful if for no other reasons than they reduce the number of data points being analyzed. The other goal of feature engineering is to create a data series with a cleaner statistical distribution than perhaps is inherent in the data used to create the new feature. A good example of this in financial data is moving average ratios. A moving average on its own is just a number. Moving averages become more useful in time series analysis when they are viewed in relation to another time series. A moving average value on its own can take on any number - it is 'unbounded'. Ratios, on the other hand, will be normally distributed - or something resembling a normal distribution.

Most machine learning algorithms, with the exception of most decision tree algorithms, are highly sensitive to the distributions and scaling of data. Most machine learning algorithms would prefer values to represent 'standard deviations from the mean', or be bound between 0 and 1. For this reason, providing a machine learning algorithm with something like a single moving average would be meaningless to it (and for us as well!).

Another example of this is bollinger bands. Bollinger bands, depending on the period selected, are lines indicating the price level that would be a certain number of standard deviations from a rolling mean price level. People view these on charts and as the price approaches them we can visually ascertain that the price is approaching a certain statistical extreme. The bollinger bands themselves, however, are also just a single value like a moving average. An example of an engineered feature used on this site is called "PROXTOBOLLXX", which represents the proximity of an asset's price to a specified bollinger band (on VanAurum these default to two standard deviations). Because VanAurum does not have eyes, we need a way of encoding the asset's price with respect to a specific bollinger band - otherwise the data will not be useful. The resulting, engineered feature, turns the unbound bollinger band value into a nicely bound feature that is in a format preferred by machine learning systems.

In fact, most effective market-based technical indicators are great examples of engineered features. The slow stochastic oscillator, for example, is a nicely bound feature that compares current price against the high and low prices for a certain period.

Machine learning is not a free lunch. In a world where machine learning is ubiquitous, I suspect feature engineering will always be a role that can benefit from the human mind. This is where we spend most of our time at VanAurum when not writing code. Finding new data sources and attempting to expose VanAurum to useful features that it can learn something from.



Indicator List:


Technical/Market Indicator

Indicator Details


XXXDMA_YYYDMA_RATIO
These indicators represent the ratio of the two indicated simple moving averages, with the moving average denoted by 'XXX' being the numerator in the ratio. A ratio of 1.0 means the moving average are equivalent. A ratio below one means the first moving average is below the second moving average, and vice versa.

Example: if 50DMA_200DMA_RATIO > 1.1, it means the the 50-day moving average is 10% above the 200-day moving average.

CLOSE
For data with Open, High, Low, Close data series', Close represents the settling price at the end of the trading period. If the asset or economic indicator being analyzed only has one value for the timeseries - like US PMI for example - this is also called CLOSE in our system to simplify naming conventions.

DAY_OF_WEEK
A numerical value betwen 1 and 5 [ 1=Monday, 2=Tuesday, 3=Wednesday, 4=Friday, 5=Friday ]

WEEK_OF_YEAR
A numerical indicator between 1 and 52 representing the week of the calendar year. 1=first week of January.

MONTH_OF_YEAR
A numerical value between 1 and 12 [1=January, 12=December]

PROXTOBOLLXX
This indicator represents the price proximity to the upper bollinger band. The bollinger band period is denoted by the 'XX'. It is calculated as:
(CLOSE-Lower Bollinger Value)/(Upper - Lower Bollinger Value). This means that a value of 1.0 means price is touching the upper bollinger band. A value of 0.0 means price is touching the lower bollinger band. This indicator can also be negative (price below lower bollinger band) and greater than 1.0 (price above upper bollinger band). Bollinger bands are calculated as 2 rolling standard deviations above or below the rolling XX-period moving average.

Examples:
  • If PROXTOBOLL42 = 0.5, this means price is half-way between the 42-day upper and lower bollinger band.
  • If PROXTOBOLL42 = -0.05, this means price is 5% below the lower 42-day bollinger band.
  • If PROXTOBOLL42 = 1.01, this means price is 1% above the upper 42-day bollinger band.

RETURNSXXX
The 'XXX'-period returns of an asset. You will see a lag in this data because the returns are retroactive (For real time returns see Rate of Change (ROC))

For example, if you plot the 42-day returns with an asset, you won't see any data for the most recent 42-days. This is because we won't know what the 42-day returns for today are until 42-days have elapsed.

The return values are in decimal format but represent percentages.

Example: if RETURNS68= 0.14, this means the 68-Day return was 14%.

ROCXXX (Rate of Change)
The 'XXX'-period rate of change (ROC) tells you how much, as a percentage, the asset has risen or fallen over the given period.

Example: if ROC21= 0.14, this means the 21-Day rate of change is 14%. The asset has gained 14% in 21-days.

RSIXX
Relative Strength Index. A momentum oscillator that measures the speed and change of price movements. The 'XX' denotes the period, in days, the RSI value is calculated on.

Example: if RSI14= 70, this means the 14-Day RSI indicator equals 70.

SLOWSTOCHSXXX
The 'XXX'-period slow stochastic oscillator. Measures the current price relative to the highs and lows of the period denoted by 'XXX'. Considered to be 'overbought' when above 80, and oversold below 20 - the meaning of overbought and oversold is up for debate, but VanAurum can quantify what the signficance has been in the past. This value is bound between 0 and 100.

BOPMAXX
The 'XX'-day balance of power (BOP) moving average. BOP is calculated as (Close-Open)/(High-Low). BOP is a measure of how influential sellers were on the day's price range relative to buyers. BOP moving averages are a measure of trends in influential buying or selling pressure for a particular asset or ratio. A BOP of 1 means price opened the trading day at the low of the day, and ended at the high of the day. Conversely, a BOP of -1 means price opened the day at the high, and closed at the low. In the rare event that the open, high, low, and close are all equal the BOP is set to 0.

UPDOWNSUMXX
The rolling 'XX' day net sum of up days and down days. Up-day = +1, Down-day =-1, Unch=0.

Example: if UPDOWNSUM42=12, this means over the last 42 days there have been 12 more up days than down days.






Description of economic factor data and their naming conventions


Introduction:

In addition to asset technical indicators and trader positioning analytics, VanAurum also pulls from a growing list of economic factor data - things like inflation, real interest rates, yield curves, and more. Our philosophy for selecting this data is similar to our asset data - we try to expose VanAurum to high quality, reliable data from the most influential markets. Most of our economic data revolves around yields and yield curve metrics. Most of our economic data series can be analyzed in the same way the asset data can, and we calculate a large cross-section of technical indicators for each economic data series as well.



Economic Factor List:
The following is a list of economic factors that VanAurum is currently exposed to.


Economic Factor Code

Details


US1MO_TREASURY_YIELD
Yield for 1-Month US Treasuries

US3MO_TREASURY_YIELD
Yield for 3-Month US Treasuries

US6MO_TREASURY_YIELD
Yield for 6-Month US Treasuries

US1YR_TREASURY_YIELD
Yield for 1-Year US Treasuries

US2YR_TREASURY_YIELD
Yield for 2-Year US Treasuries

US3YR_TREASURY_YIELD
Yield for 3-Year US Treasuries

US5YR_TREASURY_YIELD
Yield for 5-Year US Treasuries

US7YR_TREASURY_YIELD
Yield for 7-Year US Treasuries

US10YR_TREASURY_YIELD
Yield for 10-Year US Treasuries

US20YR_TREASURY_YIELD
Yield for 20-Year US Treasuries

US30YR_TREASURY_YIELD
Yield for 30-Year US Treasuries

US_INFLATION_YOY
US official inflation data year-over-year. Data series is monthly, but interpolated onto a daily scale.

US_HY_CREDIT_YIELD
US high yield corporate bond index (Prevailing yield)

EM_HY_CREDIT_YIELD
Emerging market high yield corporate bonds index (Prevailing yield)

YC10Y5Y
10-year to 5-Year US treasury yield curve. Calculated by subtracting 5-year yield from the 10-year yield.

YC10Y2Y
10-year to 2-Year US treasury yield curve. Calculated by subtracting 2-year yield from the 10-year yield.

YC10Y3MO
10-year to 3-Month US treasury yield curve. Calculated by subtracting 3-month yield from the 10-year yield.

YC5Y1Y
5-year to 1-year US treasury yield curve. Calculated by subtracting 1-year yield from the 5-year yield.

YC1Y1MO
1-year to 1-month US treasury yield curve. Calculated by subtracting 1-month yield from the 1-year yield.

USXXX_TREASURY_REAL_YIELD
Real yields for the treasury denoted by 'XXX'. Calculated by subtracting prevailing US inflation rate from the corresponding treasury yield.




Billing & Payments


Membership Levels

We offer three levels of membership: Monthly, quarterly, and semi-annual. All membership levels offer the same level of access, with longer term memberships offering a discount to the equivalent monthly rate. All membership levels come with a 14-day trial period that begins as soon as your account is created. We authorize all accounts through our payment provider, Stripe, as a way of preventing duplicate sign-ups and trial abuse.


The Account Page

The account page contains all of the information pertaining to your VanAurum user profile, payment information, and membership. On the account page you will find the following:

  • Payment Information:
    • Customer ID (Your encrypted customer token provided by Stripe.)
    • Credit card type
    • Last 4 digits of your credit card
    • Credit card expiry date
  • Membership Information
    • Account Status: Indicates whether your account is active, trialing, or inactive. Only active or trialing accounts have access to site analytics.
    • Marked for cancellation: Indicates if your account is set to terminate at the end of the current billing period. If "True", you will have site access until the end of the current billing period at which point your account will be deactivated.
    • Membership Plan: The nickname of your current membership level: Monthly, Quarterly, or Semi-annual.
    • Amount: The amount that you will be billed at the beginning of each billing period.
    • Membership ID: VanAurum reference code for your membership level.
    • Start of cycle: Represents the day that the membership was instantiated. Note that this is not always the day of your last payment.
    • End of this cycle: Represents the last day of this billing cycle. The day after this will be your next invoice date. If you are on a free trial, this date will represent the end of your trial period, after which regular billing will commence.
    • Coupon name: If you were given a coupon or special offer, the name of it will be shown here.
    • Coupon percent off: The discount (as a percentage) offered by your currently effective coupon.
    • Coupon duration: The period of time the coupon will be effective.
  • Profile Information:
    • Username: Your account name and unique site identifier.
    • E-mail: The address we use for all correspondence.
    • Password: Not visible because we do not have access to it. Instead, there is a link that allows the member to change passwords if they wish.

Cancelling Membership

Members can cancel their membership at any time by clicking on the cancel membership link on the account page. If you cancel your account within the free 14-day trial period your access will end immediately and your card will not be charged. If your trial period is over and you cancel during an active billing cycle, your access to the site will remain active until the end of your current billing period (you will not be charged again). Note that we do not offer prorated refunds on cancellations.


Changing or reactivating your membership.
If you have cancelled your membership and would like to reactivate, you can do so at any time by clicking on the reactive membership link on the account page. Note that this link will only be visible to members who have cancelled their account, or whose account is currently marked for cancellation. Note that free trials are not offered with account reactivations.

Members with an active membership can change their membership level at anytime by clicking on the change membership level link on the account page. When you change membership levels, you will not be billed for the new subscripton level until the end of your current billing period. For example, if a member has a monthly membership with a billing period that ends on December 10th and they change to a quarterly membership on December 1st, they will not be charged until December 11th (as usual). The invoice on December 11th will be for the full amount of the new quarterly membership that they switched to.


Webhooks
Because we do not store credit and payment information on our servers, we rely on something called webhooks to update our servers with recent account activity such as membership changes, credit card updates, payment failures, and other things. If you have made changes to your account on our accounts page and do not see the changes immediately reflected it could be because of a delay in the webhook being sent to our server. In this event, please wait for a period of time and refresh your account page. If the issue persists, please contact us with the issue at billing@vanaurum.ai