4 sheets tagged with "winner":

This spreadsheet was the winner of the first round of the Resolver One Spreadsheet Challenge, in January 2009.

It numerically solves Poisson partial differential equations (PDEs), and is set up to help you easily perform an electrostatic analysis of a microchip - something that would normally require software costing $20,000!

It could easily be adapted to do other analyses best performed with PDEs - for example, you could use it to work out seepage underneath a dam, or to solve thermodynamic problems, like heat conduction through a body.

To use it, download the ZIP file (using the button to the right), unpack it, and then load the file "finite_difference.rsl" into Resolver One. A simple model is loaded; click on the "Solve model" to perform the analysis. You can then graph the solution using the "Visualization" buttons below.

Two other models are provided as demonstrations in the "models" directory, which you can load up with the "Load Model" button. You can also create your own by populating the grid with appropriate numbers and clicking on the "Save Model" button.

This spreadsheet was the winner of the second round of the Resolver One Spreadsheet Challenge, in February 2009.

I wanted to develop a program to enable everyone to carry out complex statistical analysis and stochastic modeling in Resolver One by harnessing the power of R.

In order to run complicated statistical analysis/modeling I use R (an open source clone of the powerful commercial language S-Plus). R is the most successful statistical library currently existing on the market. R has different powerful libraries for data mining, genetics, portfolio optimization, asset allocation, etc.

Although R is the strongest tool for statistical analysis, unlike SPSS it lacks an easy to use yet a powerful spreadsheet. On the other hand building a bridge between R and Resolver one is extremely difficult and involved a number of trial and errors and using Interop objects. By using the provided framework one can enjoy the power of all of the R libraries with the comfort and the power of Resolver One. The accompanying examples will help the reader to use similar strategies to solve complicated statistical problems and to develop exciting results and 2D/3D graphs using OpenGL capabilities and drawing. Users are able to develop fly-through 3D visualizations from Resolver One by using RGL library in R (see screenshot). I have made an effort to provide a complete framework with directions for how to write a program in this framework and how to do error handling between R and Resolver One. All the complex libraries are provided in binary format and users only need to call them from within their Resolver One. It will then enable them to use all the functionality of R inside their Resolver One spreadsheets and send them information successfully back and forth between these two programs.

The same component can be used to call SciLab functions (a clone of MATLAB, a scientific computing environment), using the same procedure provided in the article, users are able to use SciLab similar to the way they use R inside Resolver One.

This spreadsheet can be used to evaluate a player's Texas Holdem hand and show which other hands would beat you using a Monte Carlo simulation. It can calculate odds on the flop, turn or river, and will display pre-calculated odds for pre-flop hands. It also includes a built in help system.

The spreadsheet uses a .NET library to evaluate and score a players hand and compute the hand value (ie.. Full house, 6s full of 2s). There is a grid of potential opponent pocket hands, which holds the 169 possible types of starting hands. Each cell in the grid contains a list of pocket hands for that group (stored as a list of 52 digit binary masks) and the Invalid/Win/Tie/Lose information for that group. Groups of hands with a high percent chance to win will appear as red cells, where as hands with a lower percent chance to win will appear in gray or white.

New Features:
-Built in help system to outline how the spreadsheet is used
-Support for partial hands using Monte Carlo simulation
-Displays High Win percentages as shades of red
-Displays Moderate Win percentages as shades of gray
-Displays low win percentages as white cells
-Uses cached worksheet to store initial hands and pre-flop win percentages that are pre-calculated
-Automatically recalculates odds after changing cards
-Uses separate files for each user code section, as well as a single CachedCodeModule file for cached code. This allows for use of a full IronPython IDE during development

Performance:
-Currently is set to run 300 tests per cell on the flop or 100 test per cell on the turn.
-Total hand evaluations on the flop: 350,000

This is an example for implementing the Microsoft Solver Foundation in Resolver One. This framework provides a suite of solvers and helpers. Here the classical Sudoku puzzle solved as a constraint program.

Requisite:

* Visual Studio Tools for Office 3.0 runtime (http://www.microsoft.com/downloads/details.aspx?FamilyID=54eb3a5a-0e52-40f9-a2d1-eecd7a092dcb&displaylang=en)
* Microsoft Solver Foundation Express edition (http://code.msdn.microsoft.com/solverfoundation)