close

Yo may want to upgrade you're browser.

You're using an old version of Internet Explorer to browse SpiderXL right now. SpiderXL will work better for you if you upgrade or switch to another browser.

NumXL Release Notes

(January 30th, 2010) - NumXL 1.0 (SP2) - Maintenance Release

  1. No code/feature change.
  2. Fixed Issue: The spreadsheet examples (tutorial & case studies) were not updated to reflect latest changes.

(January 26th, 2010) - NumXL 1.0 (SP2) - Maintenance Release

  1. Fixed Issue: GARCH/EGARCH out-of-sample volatility forecast does not mean-revert to long-run value.
  2. Fixed Issue: How can I construct an EGARCH model and to forecast the long-run average volatility.
  3. Fixed Issue: Cannot Use Keyboard Shortcuts to Select Ranges.
  4. Fixed Issue: A missing leading zero in significance level field locks up excel.
  5. Fixed Issue: NumXL VBA compile errors.
  6. Minor Enhancements:
    • Revised the arguments list for EWMAi and EWXCF functions
    • Revised the arguments list for GARCH/EGARCH/GARCHM_FORESD
    • Support for term structure volatility forecast
    • NumXL dialog is now pops up on Excel main window upon startup (and never hidden).
    • A splash window pops up now for 3 seconds upon excel startup

(January 5th, 2010) - NumXL 1.0 (SP1) - License support

  1. No code/feature change.
  2. Activated the license management subsystem. The downloaded add-in operates in professional mode during the trial period. Afterward, the add-in revert to Lite function-restricted mode.

(October 6th, 2009) - NumXL 1.0 - Security release

  1. No code/feature change.
  2. Complying with software industry security policy and practices, Spider Financial signed (using CA-issued digital certificate) NumXL 1.0 package to ensure authenticity and integrity.

(October 1st, 2009) - NumXL 1.0

  1. New Excel function (EWXCF) to compute the correlation time series using exponential weighted volatility and covariance.
  2. New Excel functions to calibrate the user's model(s). The need for such functions arises in model's parameter stability and back-testing using rolling windows.
  3. New support for building a hybrid models (e.g. ARMAi for the mean and GARCH for the vol, GLM+ARMA for the mean and EGARCH for the vol, etc.)
  4. Added Generalized Linear model (GLM) to its list of supported model. The GLM is introduced to presemt linear models for exogneous factors (e.g.calendar events, economical data, etc.)
  5. MINOR: NumXL clears the messages posted to EXCEL status bar upon loading.
  6. NumXL UI added a calibration (i.e. model fitting) functionality to its menu and toolbar. Upon selection, NumXL initializes the Solver's Add-in fields with the selected model's values and launches it on screen.

July 24th, 2009 - NumXL 1.0 RC

  1. New Toolbars and menu - A new Add-in (NumXLUI) is written in VBA and is designed to assist our users to automate the process of data analysis. The new Add-in includes Dialog box and wizard to capture user's input and insert the corresponding analysis blocks (with equations) into Excel worksheet.
  2. New Tutorial document is available now to introduce basic principles of data analysis using NumXL. An online version is available too.
  3. Redesigned the function categories to better-help our users find NumXL function.
  4. Fixed minor issues in the help file related to the tree view.
  5. Fixed minor problems in the Installer with regard to registering the add-in with Excel.
  6. Removed Functions - TSAVG, TSSD, TSSKEW and TSXKURT: With the introduction of RMNA, our users can use EXCEL built-in functions to acheive the same goal. For INSTANCE, TSAVG(Data) is now equivalent to AVERAGE(RMNA(Data))
  7. Added new Function - RMNA: RMNA removes missing values (#NA) from a given data set while preserving the original order of the series.
  8. Performance revision - Revised the INTERPOLATE function to accept an array of target values and return array of values.

June 30th, 2009 - NumXL 1.0 Beta 

  1. Added API to perform basis statistical test for the mean (TEST_MEAN),
    standard deviation (TEST_STDEVi), skewness (TEST_SKEW) and
    excess-kurtosis (TEST_XKURT).
  2.  Added support for leptokurtic (fat-tailed distributed) innovations (Student's t-distribution, GEDi) into GARCH/EGARCH/GARCH-Mi models.
  3. Added case study for S&P500 monthly, daily and weekly returns analysis.
  4. Added case study for Russell 2000 monthly, daily and weekly return analysis.

April 15th, 2009 - NumXL 1.0 Alpha

  1. Initial release.
  2. Revised the calculation of ACFi to the following formula:
    3. \[ r_{h} = \frac{\sum_{k=h}^N{(y_{k}-\bar y)(y_{k-h}-\bar y)}}{\sum_{k=h}^N(y_{k}-\bar y)^2} \]
  3. Added support for Lagging (LAG) and Difference (DIFF) Operations in NumXL. As a result of, the returned series contain missing values and "#N/A" are inserted for those values.
  4. NumXL Functions recognizes "#N/A" values as missing values, and adjust appropriately for their presence.
  5. Added Normal distribution test for a sample using Jacque-Bera test. The test is primarily based on skewness and kurtosis of the sample data.
    JB = {S^*}^2+{K^*}^2 \sim \chi^2(2)
    S^* = \frac{\textrm{Skewness}}{\sqrt{6/T}}
    K^* = \frac{\textrm{Kurtosis}}{\sqrt{24/T}}
  6. Added a two tailed test for ACF and PACFi
  7. Added a test (WNTest) for white noise (Portmanteau statistics). The test primarily checks for serial correlationsin the sample. We implemented the Ljung-Box (1978) modified Q*(m)\sim \chi^2(m) statistics. The default and recommended value for m= ln(T) , but the user can supply his own. The function returns the P-Value of the test.
  8. Added support for 3 statistical normality tests (jarque-Bera, Wilk-Shapiro, and durnik Chi-square).
  9. Added ARMA model support APIs: ARMA_AICi for (measure goodness of fit), ARMA_LLFi (log-likelyhood), ARMA_RESID (standardized residuals), ARMA_FORE (forecast), ARMA_FORECI (confidence interval limits).
  10. Added GARCH model support APIs: GARCH_AIC for (measure goodness of fit), GARCH_LLF (log-likelyhood), GARCH_RESID (standardized residuals), GARCH_FORE (forecast), GARCH_FORECI (confidence interval limits).
  11. Added EGARCH model support APIs: EGARCH_AIC for (measure goodness of
    fit), EGARCH_LLF (log-likelyhood), EGARCH_RESID (standardized residuals),
    EGARCH_FORE (forecast), EGARCH_FORECI (confidence interval limits).
  12. Added GARCH-M model support APIs: GARCHM_AIC (measure of goodness of fit), GARCH_LLF (log-likelihood functions), GARCHM_RESID (standardized residuals), GARCHM_FORE (Forecast), GARCHM_FORESD (volatility forecast), etc.
  13. Added <MODEL>_CHECK (ARMA_CHECK, GARCH_CHECK, EGARCH_CHECK, GARCHM_CHECK) to examine the selected model coeffient for stability.
»