# Scenario Lab Knowledge base to support Scenario Lab, a robust economic and scenario analysis platform. **Windham’s Scenario Lab** is a new analytical platform that enables investors to evaluate and stress test portfolios based on a solid scientific foundation that eliminates the biases and guesswork associated with conventional approaches to scenario analysis. {% embed url="" %} Scenario Lab: A Powerful Scenario Analysis Tool {% endembed %} It relies on a powerful statistic called the Mahalanobis distance to determine the likelihood of prospective economic scenarios. This statistic was originally derived in the early 20th century by an Indian mathematician to analyze human skulls, and it has since been applied to measure turbulence in the financial markets, diagnose diseases, detect anomalies in self-driving vehicles, and construct an index of the business cycle. **Scenario Lab** supports all the steps of scenario analysis. ### Features * Generates pre-defined or customized economic scenarios
* Assigns probabilities to economic scenarios by measuring their statistical similarity to current economic conditions or average economic conditions for a chosen history
* Shows how to modify scenario specifications to reconcile empirically driven scenario probabilities with an investors' subjective views
* Maps prospective economic scenarios onto asset class returns
* Generates a variety of portfolio metrics for determining the most suitable portfolio ### Scenario Lab Innovations * Replaces guesswork of assigning probabilities to prospective scenarios with a mathematically rigorous algorithm for measuring statistical similarity
* Enables investors to evaluate portfolio risk and return {% hint style="info" %} Although **Scenario Lab** rests on a system of complex mathematical formulas, it is simple to operate and easy to understand. {% endhint %} # Release Notes Change log for communicating new features, fixes, and revisions for Scenario Lab and Stress Test Lab. ![Scenario Lab, a new tool for Enhanced Scenario Analysis](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHHuUuUl4YqZSukE1wN%2F-MHHxkqbB2MEozlPBsDo%2Fimage.png?alt=media\&token=d7151ace-7587-4030-a06b-1d2d31bb88c9) ## 1.2.0.11448 (2022-04-25) We have added features and made improvements since our launch of Stress Test Lab. ### Added * Modify probabilities of prospective stress scenarios and reconcile scenario estimates with your views. * Review the covariance matrix on the examine stress scenario screen. ### Updated * Squashed bug that sometimes did not let users switch anchor scenarios. ## 1.2.0.11419 (2022-03-07) We are proud to announce our newest innovation, Stress Test Lab! Stress Test Lab allows you assess the probability of stress scenarios. This work is based on recently [published research](https://www.risk.net/journal-of-risk/7931516/severe-but-plausible-or-not) in the Journal of Risk. ### Added * Smart stress scenarios allow users to select data that lies above or below a certain percentile range for an economic variable. * Adaptive anchoring allows a user to forecast probabilities based on their views of economic conditions persisting or reverting to its historical norms. * Adjust the plausibility model to align theoretical assumptions with the empirical distribution. ### **Updated** * Improved cyber security. * Improved chart presentation. * Improved performance of the application so you can analyze scenarios faster. * Improved table presentation. * Replaced scientific notation with rounding. ## 1.1.0.10829 (2020-12-08) More improvements just in time for the holiday season, warm wishes from the Windham Labs team! ### Added * Expanded data set available for analysis. * Licensing and security improvements. ### Updated * Optimized loading of instruments and economic variables, 10x improvement! * Improved database infrastructure. ## 1.1.0.10755 (2020-11-16) The Labs team have been busy generating new modeling ideas and implementing them for your analysis needs. We've added our recent research insight into a different approach to regression - for mapping relative scenario probabilities onto asset class returns. ### Added * Access the [Partial Sample Regression](https://jpm.pm-research.com/content/early/2020/07/06/jpm.2020.1.167) as a regression model for mapping scenario probabilities onto asset class returns, and configure a relevance threshold. * Added a Help menu item under the user profile. ### Updated * Improve administration panel configuration. ## **1.1.0.10710 (2020-10-28)** ### Added * Partial sample regression model to the calculation engine (see the our [research paper in the Journal of Portfolio Management September 2020](https://jpm.pm-research.com/content/early/2020/07/06/jpm.2020.1.167)) * New endpoints for Historical Events and authentication. ### Updated * Prospective change to reference anchor scenario when calculating expected scenario returns. * Squashed bug when creating prospective scenarios with the same name as anchor scenarios. ## 1.1.0.10661 (2020-10-07) Hello World! Scenario Lab's first version gets introduced to the market today. ### Added * Export diagnostics data for advanced and internal users. ### Updated * Fixed asynchronous request conflict on View Historical Impact screen. * View Historical Impact now defaults to grouping by attribute across all portfolios. * Independent variables treatment (yearly changes) in multi-factor regressions. * Updated research portal reference. * Minor UI revisions. ## 1.1.0.10621 (2020-10-05) You can now review historical events and their impact across your portfolios and use a new index of the business cycle, the KKT Index - a recession probability indicator. ### Added * [The Kinlaw-Kritzman-Turkington (KKT) Index](https://www.statestreet.com/ideas/articles/kkt-index.html), a recession probability indicator, is now included in Scenario Lab. * **View Historical Impact** screen: analyze historical events across portfolios. ### Updated * Improved server security. * Minor control behavior fixes across data selection screens. * Stability fix on Examine Scenarios screen. ## 1.0.0.10544 (2020-07-31) Introducing Scenario Lab internally, we believe this is going to be a neat tool for analysts. ### Added * Store and encapsulate case studies into "case files". * Select economic variable and asset data from a Windham-curated list of instruments. * Review timeseries and descriptive statistics about timeseries data. * Specify and customize prospective scenarios as projections of economic variables. * Analyze and update probabilities of prospective scenarios given anchor scenarios. * Inspect economic variable returns mapped onto asset class returns. * Observe each portfolio’s expected return based on (scenario) probability-weighted returns. # Getting Started A first guide on approaching scenario analysis using Windham's Scenario Lab ## Launching Scenario Lab Navigate to Windham's app portal and launch Scenario Lab to log in to the application. ```bash https://apps.windhamlabs.com/start/ ``` ## Navigation The Scenario Lab application follows a fluent design framework. Navigation menu is accessible via the hamburger icon on the top-left of the screen. ![](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHHxo0zSpl7UnNDATqU%2F-MHHyglyrVFUlToiB81T%2Fimage.png?alt=media\&token=a6be85f3-ba6c-4998-87f0-ee242bdb1d1a) # Navigation Learning how to navigate Scenario Lab ## Main Navigation ☰ Scenario Lab uses a fluent design framework. Navigation is accessible via the hamburger icon on the top-left of the application. ![Navigation Icon in Scenario Lab](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHHxo0zSpl7UnNDATqU%2F-MHHzjU6NdPeKz5R6hEz%2Fimage.png?alt=media\&token=cc015729-39a9-448b-9dd6-3390c205c770) Expanding the navigation menu will show the available screens. ![Navigation menu expanded](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHHxo0zSpl7UnNDATqU%2F-MHHyglyrVFUlToiB81T%2Fimage.png?alt=media\&token=a6be85f3-ba6c-4998-87f0-ee242bdb1d1a) ## Status Panel Throughout the application, the bottom panel will display useful information such as the name of the casefile you are currently working on, the common date range of the data, and the previous and next screens available relative to your current location. ![Status panel and Navigation Breadcrumbs](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIZN1mCXR5IT_VFRl9%2F-MHIam21lP0BB2oLLmtI%2Fnavstatus.gif?alt=media\&token=f83bb3d3-48ba-4e8b-9f57-2a2083556944) ## Workflow from Research Paper The application abstracts and generalizes the framework by [Kritzman, et. al. (2020)](https://www.statestreet.com/content/dam/statestreet/documents/ss_associates/JPM%20Enhanced%20Scenario%20Analysis.pdf): 1. Define a set of prospective scenarios as a combination of economic variables.
2. Calculate the Mahalanobis distance of prospective scenarios from the prevailing and normal values and convert the statistical distances into a measure of relative likelihoods.
3. Map the scenario probabilities onto asset class returns and review. ## Workflow in Scenario Lab In Scenario Lab, we break down the framework described in our research into four user interface modules. ### Administration * **Choose Casefile**\ We organize data and scope of an analysis into a casefile document. Manage and organize your library of casefiles (analyses) on this screen. ### Design * **Select Economic Variables**\ Start building your casefile by selecting from dozens of commonly used economic variable time series. The software includes both quarterly and monthly time series.
* **Select Assets**\ Choose from an extensive list of asset class index returns to include in your casefile. ### Review * **Review Economic Variables**\ Inspect time series properties of selected economic variables corresponding charts.
* **Review Assets**\ Inspect time series properties of selected asset class index returns along with interactive visualizations. ### Analysis * **Classify Scenarios**\ Determine criteria to partition the data for initializing weak, normal, and robust scenarios.
* **Examine Scenarios**\ Estimate scenario probabilities, revise scenario estimates, or define new scenarios.
* **Estimate Expected Returns**\ Map scenario probabilities onto asset class returns.
* **Evaluate Portfolios**\ Contrast summary statistics across portfolios or export data and results. # Choosing Casefiles Creating and managing your projects in Scenario Lab ![Managing your projects](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXzhrJ6XpA7OXmNBsz%2F-MHY-fAJleQ3xIaaVsLG%2Fcasefilemanagement.gif?alt=media\&token=1e8918db-e01c-47d7-bd93-55de352c22a8) ### Creating a New Casefile In our suite of applications, we refer to each project as a casefile. A case file will save and store all information pertaining to a particular analysis. To create a new case file, click on + Create Casefile to get started and follow the on-screen instructions. ![](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXzhrJ6XpA7OXmNBsz%2F-MHY0foCGpq13ANQjz0i%2Fimage.png?alt=media\&token=df3f6ba7-c90a-486f-ae7e-179f7e823099) ### Organizing Casefiles ![Traffic light icons indicate actions available for managing an item](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXzhrJ6XpA7OXmNBsz%2F-MHY11syBTuyYjyIL_tb%2Fimage.png?alt=media\&token=125df8a3-64c9-4e62-a411-c3f30889d2a6) You can organize a collection of casefiles into groups. You may find this helpful to file away ideas that are prototypical, toy examples, client ready analyses, etc. ![Creating New Groups](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXzhrJ6XpA7OXmNBsz%2F-MHY1hJWIxo3ekTdZWXi%2Fimage.png?alt=media\&token=4628eed9-0369-4669-92cc-248d75aff711) To create a new group, simply edit the details of your case file and type in a new group name. # Design # Select Economic Variables Specify and select economic variables relevant for your analysis This screen allows you to select from a database of available economic indicators for use in the analysis. Scenario Lab defaults with both quarterly and monthly time series of economic variables. The top table shows the list of available economic variables to select from. Filter using the search box. The table shows the date range of available time series data for each economic variable and its source. ![Database of Economic Variables](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIUa7r-p9Xg1Q1OrEt%2F-MHIVFgt4yO7K-A8yPkO%2Fimage.png?alt=media\&token=ecbbcc1a-0d9c-47e4-bacd-8ca71ad9572b) {% hint style="info" %} Prior to selecting economic variables, review and specify the data periodicity that you want to work with. {% endhint %}
Specify data periodicity
The bottom table shows a list of selected economic variables saved into the casefile. The bottom grid allows you to rename economic variables for display, examine the common data date range, and remove unwanted economic variables. ![Review your selections](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIUa7r-p9Xg1Q1OrEt%2F-MHIZ0cL1IGLsU9cCL_3%2Fselectvariables2.gif?alt=media\&token=f64fdd6b-52fd-400e-9790-242502a2913b) # Select Assets Specify and select asset class time series returns to include in your analysis This screen allows you to select from a database of available asset class returns for use in the analysis. Scenario Lab defaults with both quarterly and monthly time series of asset class returns. The top table shows the list of available asset class instruments to select from. Filter using the search box to narrow your selection. Click on Add Assets to include your selection into your casefile. ![Searching and selecting assets](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIbQBITNue73srbzPD%2F-MHIcRC61RUt4nZeHUnM%2Fselectassets.gif?alt=media\&token=4800b585-9b8b-4b25-a2e1-4b42370b517f) The bottom grid control shows a list of selected asset classes saved into the casefile. You can rename asset classes for shorter or custom display names, examine the common data date range, or remove irrelevant asset classes. ![Review your selection](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIcXfhEvboDiI2fl03%2F-MHIdDZM7MQQcCQQsX5n%2Fimage.png?alt=media\&token=fb73a49c-c941-4fba-9118-6c0e991a4123) # Review # Review Economic Variables Inspect time series properties of selected economic variables Once you've selected your economic variables from Scenario Lab's database, you can review the details of its time series data on the Review Economic Variables screen. Click on an economic variable row to update charts. Scenario Lab will plot the histogram of the empirical distribution of the variable as well as the corresponding time series plot. ![Reviewing your economic variables](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIdHbz7fIHGhhqiW6p%2F-MHIdn74xuBTiZEM8tNq%2Freviewvariables.gif?alt=media\&token=4b7a0824-bcb5-4f48-99fa-e4c4fc58c4eb) The visualizations are particularly helpful to understand the cyclical behavior of the selected variables. It also can inform you of its variability and range of outcomes experienced in the past. The table summarizes the minimum, maximum, and the first four statistical moments for each selected economic variable. # Review Asset Classes Examine time series properties of selected asset classes Next, you can review the details of asset class time series data on the Review Asset Classes screen. This screen is similar to Review Economic Variables. Click on an asset class row to update visual exhibits. Scenario Lab will plot the histogram of the empirical distribution of an asset class as well as a scatter plot of its time series. ![Reviewing asset class data in Scenario Lab](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIezGYxRpfgAkTeqBm%2F-MHIgc1EnfuAjIy4cY7d%2Freviewassets.gif?alt=media\&token=db8f4c9e-3591-47db-ab1e-5b855a537071) The visualizations are particularly helpful to understand the empirical properties of the selected variables. This is particularly useful to examine fat tails or asymmetry in the data. The table summarizes the annualized return, annualized risk (standard deviation), skewness, and kurtosis for each asset class in your casefile. # Analysis # Classify Scenarios Determine criteria to partition the data for initializing weak, normal, and robust scenarios. ![Initializing prospective scenarios in Scenario Lab](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHIi41I0ubZ4N9fXxqU%2F-MHIihkcMdEkrOkTmdkL%2Fimage.png?alt=media\&token=52c1b456-04f0-44bd-b9a3-7fff4069e6f7) To help you get started with defining a finite set of prospective scenarios, the Classify Scenarios screen allows you to initialize a Weak, Normal, and Robust scenario by 1. Specifying a reference economic variable to partition market environments.
2. Specifying the directional economic favorability of the variable. \ *(e.g. low GNP year-on-year changes are a weak indicator, while high unemployment values are a weak indicator)*
3. Determining the cutoffs from the weak, normal, and robust ranges.
{% hint style="info" %} Scenario Lab automatically helps you initialize default upper and lower thresholds for a selected reference economic variable to the $$25^{\text{th}}$$ and $$75^{\text{th}}$$ percentiles of its empirical distribution. {% endhint %} The weak, normal, and robust scenario estimates are calculated by the conditional means of the sub-samples as defined by you. # Examine Scenarios Estimate scenario probabilities, revise scenario estimates, or define new scenarios. This is the core of Scenario Lab. This screen allows you to 1. Create new scenarios to extend beyond Weak, Normal, and Robust scenarios. 2. Review the underlying Covariance assumption. 3. Estimate probabilities of prospective scenarios. 4. Edit scenario probabilities and solve for scenario estimates that are consistent with specified views. ![](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MMVwbd8Yt-CZDKEkgmL%2F-MMVwnmGzlEJ8KMulTOF%2FscenarioLab600.png?alt=media\&token=c37a04b1-901b-42b7-9b19-8ec7c30dc2e6) The grid at the top of the screen shows the prospective economic scenarios (columns) and the corresponding economic variables values / views (rows). Scenario Lab measure the statistical distance of the prospective scenario estimates from this table to an anchor condition: past historical norm or current prevailing conditions. ![Selecting an anchor condition](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXhgx_Pv0uzt5QhJ1-%2F-MHXnvXGaXcc8T4gKwOc%2Fimage.png?alt=media\&token=a311c36b-22b3-47eb-b636-f550d337061c) Scenario Lab then converts these statistical distances from either the historical norm or current conditions into relative likelihoods. The relative likelihoods are then normalized and displayed as scenario probabilities in the bottom grid. ### Add or create new Prospective Economic Scenarios To add a new prospective economic scenario (column) in the top grid, click on on the create new scenario button and follow on-screen instructions. ![Create and specify addition prospective economic scenarios](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXhgx_Pv0uzt5QhJ1-%2F-MHXi7HkMxadCljPv5YE%2Fimage.png?alt=media\&token=b4124120-4ca9-4a5b-99b7-b30b57ab8768) You can create up and work with up to nine prospective economic scenarios ### Review Covariance To review the historical covariance used for the underlying analysis, click on Review Covariance to show a table of standard deviations and correlation coefficients across the economic variables. ### Revising Scenario Probabilities If you would like to edit or override the scenario probabilities with your own views, click on the corresponding scenario probability to edit: ![Revising a scenario probability](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXhgx_Pv0uzt5QhJ1-%2F-MHXnjkldXBwyEeTzf3t%2Fimage.png?alt=media\&token=affba41d-88f9-492a-b125-fea16ee130e2) Revising a scenario probability will engage a numerical solver and update scenario estimates in the top grid to reconcile the empirically-driven process with your new views. If you would like to reset any edits back to the initial set of empirical-driven probabilities and scenario estimates, click on Reset. ![Reset all edits](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXoPMiYtF3yOIbiLCN%2F-MHXofgQ2hKrF50O_mOM%2Fimage.png?alt=media\&token=ce686000-521f-4f07-95c5-72ef5dcbe787) # Estimate Expected Returns Map scenario probabilities onto asset class returns ![Estimate asset class expected returns](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXoiiPdGrKslodmPKt%2F-MHXoyKuJHSIbyTY8mVS%2Fimage.png?alt=media\&token=9e0a2d6e-3ae6-452a-a867-280ffd10b43c) This screen will show you the annualized expected historical returns, the mapped expected returns across the set of asset classes in your case file, and the scenario-weighted expected returns. Scenario Lab maps the prospective economic scenarios onto asset class returns by 1. Calculating the rolling yearly returns of asset classes (dependent variables) in excess of cash 2. Calculating the yearly changes in economic variables (independent variables) 3. Running a multi-factor regression Scenario-dependent asset class returns are calculated using the prospective change in each economic variable with the regression coefficients. The prospective change is calculated by evaluating the difference between the prospective scenario estimates for each economic variable and the anchor condition. A cash instrument is specified on this screen along with a configurable risk-free rate. The software platform marks the lower return instrument as cash absent of your specification. By multiplying the scenario dependent returns with scenario probabilities, we arrive at the probability-weighted-average estimate for each asset class. We always present this in the right-most column of the grid under the column "Scenario-Weighted". {% hint style="info" %} The framework of mapping prospective economic scenarios onto asset class returns are described with detail in the [Journal of Portfolio Management article, "Enhanced Scenario Analysis"](https://www.statestreet.com/content/dam/statestreet/documents/ss_associates/JPM%20Enhanced%20Scenario%20Analysis.pdf). {% endhint %} # Evaluate Portfolios Review and analyzed expected returns for your asset allocation needs ![Comparing portfolio outcomes](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MF21wY165I3tTs1ivrr%2F-MHXxWnHUZdgeqozlHLN%2F-MHXybZYOy5WFpGOajHc%2Fimage.png?alt=media\&token=25c0514d-34fb-49b2-8ac2-f67f3cb0480c) This screen allows you to specify portfolio asset allocation to evaluate expected return outcomes based on the scenario-weighted returns. The screen initializes default asset allocation weights by creating a conservative, moderate, and aggressive profile and guessing some initial weights based on the asset class selection. Edit the weights cells directly to specify your asset allocation portfolios and save it to the case file. {% hint style="info" %} The engineering team continues to work on expanding analytics of outcomes, bringing the wealth of research and capabilities that power the [Windham Portfolio Advisor](https://www.windhamlabs.com/products/windham-portfolio-advisor.html) to Scenario Lab's evaluation screens. {% endhint %} # Stress Test Lab The following section(s) describe functionality and screens pertaining to the Stress Test Lab license of Scenario Lab. {% content-ref url="/pages/HQaKIl6DODIhj4syDdXh" %} [Examine Stress Scenarios](/user-guide/stress-test-lab/examine-stress-scenarios) {% endcontent-ref %} # Examine Stress Scenarios An explanation on how to use the examine stress scenarios screen This is the core of Stress Test Lab. This screen allows you to 1. Create prospective stress scenarios and an alternative composite (non-stress scenario). 2. Review the underlying Covariance assumption. 3. Evaluate probabilities of stress scenarios relative to the alternative composite. ![](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2F05kCootzDUM3qOQdKA3Q%2FStressTestLabScreen.png?alt=media\&token=38d9304d-9212-4e43-a219-31f17ed222f7) The grid at the top of the screen shows the stress scenarios (columns) and the corresponding economic variables / values (rows). Stress Test Lab measures the statistical distance of each stress scenario to an anchor condition: past historical norm or current prevailing conditions ![Anchor selection](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2FgZm1I7KTM1Z0eFVBe5cn%2FProbsBasedOn.png?alt=media\&token=e6f4a975-82fe-4a84-8b85-e31a0ced5104) Stress Test Lab then converts these statistical distances from either the historical norm or current conditions into relative likelihoods. The relative likelihoods are then normalized and displayed as probabilities in the bottom grid. ### Create prospective stress scenarios and an alternative composite To add a new stress scenario (column) in the top grid, click on the new stress scenario button, and follow on-screen instructions. ![Add a New stress scenario](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2FE0gnW5umdzOV9n1uPB9A%2FNewStressScenario.png?alt=media\&token=ecd93273-2be1-40ff-9e0c-442205b6c47f) We provide a wizard option to help you set up each stress scenario. ![The Stress Scenario Wizard](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2FlBMtF7nRnDlvqtDQjMcn%2FStressScenarioWizard.png?alt=media\&token=9c363bf9-2e76-435e-b958-447b8274698e) ### Review Covariance To review the historical covariance used for the underlying analysis, click on Review Covariance to show a table of standard deviations and correlation coefficients across the economic variables. ### Evaluate probabilities of stress scenarios The Gaussian probability column assumes that the statistical distances are normally distributed. Unfortunately, real world data is rarely this well behaved. ![Stress Scenario Probabilities](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2Fk4ObzAQbNv3KNgTH0Tl0%2Fprobabilities.png?alt=media\&token=c6f38228-df1d-41ae-a4f3-5e534bbfb4b3) To account for non-normality, adjust the model to empirically align the statistical distances. We show the details of *j* adjustments as described in our research paper, you can review this by clicking the ellipses on the empirically aligned probability column. ![Empirical Alignment Information](https://3022362199-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MF21wY165I3tTs1ivrr%2Fuploads%2FLyCvN89tbJ3ALvfJ17dA%2FEmpiricalAlignmentWindow.png?alt=media\&token=4488d110-2bbc-49fb-aa31-9c40795242bf)