Searching

Introduction

The searching functionality of the TandaPay simulation is a way to test how different variables effect the outcome. It's a powerful tool for analyzing this relationship, and can generate regression models of order ∈ [1, 10], and solutions for the regression model, which gives you an idea for what value would yield the outcome you're targeting.

Inputs

Target Percent: This is the percentage of (wins/losses/draws) you're targeting. This will effect the solution to the linear regression model.

Outcome to target: This is where you select whether you're targeting (wins/losses/draws).

Min Value: You can restrict the search space of the variable you're searching by setting a minimum value for the search space. The recommendation is to leave this at 0, as by default it simply searches through the entire search space.

Max Value: You can restrict the search space of the variable you're searching by setting a maximum value for the search space. The recommendation is to leave this at 100, as by default it simply searches through the entire search space.

Steps: How many samples do you want to take in the search space to create the table that will be printed, and use as data points for the regression model? (recommended: 20)

Order: What order do you want to use for the linear regression model? (recommended: 3)

Attribute to search: Which variable are you wanting to search? These will look slightly different than the more readable variable names in settings, but using common sense, you can see a clear association between these and the options listed in Settings.

Outputs

Table of values: Shows the result of each sample it took.

All Roots: All of the roots to the regression model, including those that are impossible or are nonreal answers.

Real Roots: Only the real roots of the regression model, but still includes impossible answers.

Possible Solutions: This is the section you're most likely interested in. It is the set of solutions for which the solution is both within the domain of acceptable values for this variable, and it's a real number. If you get multiple solutions here, try reducing the order of your linear regression model, or simply use common sense to pick the solution you're targeting.

Linreg Model: Despite being called "Linreg Model", it isn't necessarily showing you only a Linear Regression model. It will show you whichever linear regression model it used to calculate the previous results, and it will be the order you selected in the inputs for this search.

Notes

The sample size for each "Step" is set by "trial sample size" in settings.