Daz Toolkits

Daz is an evolving collection of toolkits for the professional investment analyst.

Over time, the toolkits will fall into three types: Analytics, Data and Reports.  At this time only three Analytics modules are available.  Four Data modules are under development.  The Report modules will appear in the second half of 2010.

Why Tools?

Definition: Instruments used in a profession, e.g., tools of the trade.

Tools have been the helpers of craftsmen for millennia.  Each craft has its own characteristic tools, and each craftsman has his/her own distinct preferences.  The wrong tool is worthless.  The right tool is invaluable.  The hand knows which is which.

Daz is such a collection of tools.  These are discrete toolkits, each fit for a purpose – distinct and complementary to the other toolkits.  They are designed for the craftsman who knows his/her craft.  They provide the standard, expected, instruments as well as the more unusual instruments.  There are some bespoke instruments that have been created by a practitioner for his/her own use.  All of the tools are controlled directly by the craftsman, in his/her workbench: Excel.  The collection encourages the craftsman to seek out the right tool for his/her purpose not just the commonest tool.

Analytics

There are three collections of Daz analytic tools: DazStat, DazRatio and DazBeta.  They share a common design theme:

  • All of these toolkits run inside Excel as user-defined functions, just as any Excel-provided function
  • The function parameters are consistent within each module
  • The functions are differentiated by the first parameter so the craftsman can move between functions by changing one switch
  • Most functions return an array of secondary results as well as the default primary result
  • Each toolkit can return all primary and secondary results as a single array
  • The source returns may be arranged vertically or horizontally
  • The source returns may be in decimal (0.03) or percentage terms (3.0)
  • Leading and trailing zeroes are ignored
  • Every function can return its own label
  • Returns can be of any periodicity

DazStat is the fundamental toolkit.  Its focus is on the characteristics of the distribution itself.  In particular, DazStat emphasizes tests for randomness and conformity with the Normal distribution function.  If the returns do not conform to the Normal distribution function then most of the standard descriptive statistics used in conventional investment analysis will be invalid.

DazRatio focuses on risk-adjusted performance measures.  Different craftsmen have different preferences for these ratios so this toolkit provides many tools that may be compared.  All these tools allow the craftsman to specify a constant minimum acceptable return (MAR).

DazBeta is the most complex toolkit.  These tools include conventional and proprietary measures for the relationships between returns such as correlation and Beta.  These tools are quite subtle and benefit greatly from being part of a standard function library for consistency of analysis.  While DazRatio provides for a constant MAR (consistent with convention), DazBeta provided an alternative of specifying a time varying return array as dynamic MAR.

Data

DazUtility is another library that complements the analytics. In this case DazUtility is providing for the management and transformation of data that is then likely to be used in the analytics.

DazUtility works within Excel to provide transformations for the user’s data.  Some transformations are obvious: calculating returns from prices.  Some are less obvious: calculating weekly/monthly returns from daily returns and aligning the weekly/monthly return to each Friday and last calendar day of the month, regardless of the last market quote.  Some transformations work naturally with the Daz analytics toolkits. For example, separating the return series into runups, drawdowns and neutral markets based on a minimum qualifying period for runup/drawdown.  Not just doing that to one return series, but also have a collection of returns that are partitioned based on the benchmark return series.  Now the user can use the Daz analytics to see if the volatility or correlations were different depending on market phase.  A simple and powerful way to extend the power of the Daz analytics.