> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dash360.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Risk Analysis
> Run the unified cost and schedule Monte Carlo simulation: configure the uncertainty cascade, choose a simulation mode, read the contingency decomposition and WBS rollup, and compare scenarios and saved runs.
## Overview
The Risk Analysis page runs Dash360's unified Monte Carlo engine. From a single set of iterations it produces both a project cost distribution and a project finish-date distribution, driven by two sources of variability:
* **Estimating uncertainty**: the spread in activity durations and resource-assignment costs, sampled from their Uncertainty Classes.
* **Risk events**: the risks in the Risk Register, each rolled against its probability and cost or schedule impact.
The page is organized into two cards: **Cost and Schedule Uncertainty** (the Uncertainty Cascade, where you decide which classes apply where) and **Monte Carlo Simulations** (where you run, read, save, and compare simulations).
**URL:** `/Risk/ScheduleRiskAnalysis/Index`
For the full, end-to-end explanation of every variable that feeds the simulation (modes, distributions, calendars, risk routing, correlation, and both contingency decompositions), see [How the Dash360 Monte Carlo Works](/user-guide/concepts/monte-carlo-methodology). This page focuses on operating the tool; that page explains the math.
**Availability (Pro vs Lite):** The Uncertainty Cascade, simulation-mode selection, Activity-to-Activity correlation, the Schedule results, Compare Simulations, and the deeper result views (contingency decomposition, Monte Carlo by WBS, and sensitivity charts) are **Risk Pro** features. In the **Lite** tier the page is limited to a Risk-Events-Only cost run with the percentile summary; Pro-only sections show a **PRO** badge with an upgrade link.
## Video walkthrough
A full tour of the Risk Analysis page: setting the uncertainty cascade, running a simulation, reading the cost and schedule distributions, the contingency decompositions, sensitivity charts, and comparing saved runs.
## Prerequisites
* A project must be selected.
* Uncertainty Classes should be defined for the project. See [Uncertainty Classes](/user-guide/admin/uncertainty-classes).
* For risk-driven results, the project's Risk Register should contain risks with probabilities and Risk (pre-response) and Mitigated Risk (post-response) impacts.
***
## Uncertainty Cascade
The **Cost and Schedule Uncertainty** card (the Uncertainty Cascade) decides which Uncertainty Class applies to each part of the project. It resolves top down: a class set at the **Project** level applies everywhere unless a **WBS** overrides it, which a **Work Package** can override, which an **Activity** or **Resource Assignment** can override. This lets you set a sensible project-wide default and refine only where needed.
The cascade is shown as a tree (Project, WBS, Work Package, Activity, Resource Assignment), expanded fully by default so you can see the whole structure at once. Each node has a **Cost Uncertainty** and a **Schedule Uncertainty** column showing its effective class and where that class came from (set directly, or inherited from a parent), plus a **Cost Overlay** column on the rows that carry that flag.
Each class cell is a slider: a lock toggle on the left, then an **OFF** stop, then the Uncertainty Classes as color-graded stops (conservative on the left, aggressive on the right). Drag the thumb or click a stop to set the class.
* **Set a class** on a node to override what it inherits. The override cascades to that node's children unless they have their own override.
* **Lock / Unlock** a node with the lock icon: a locked row holds an explicit class; unlocking clears the override and returns the row to inheriting from its parent (inherited rows show in italics).
* **OFF** sets an explicit no-uncertainty (deterministic single point) class. Unlike unlocking, OFF is itself an explicit assignment that stops inheritance and cascades a deterministic value to descendants; use the lock icon, not OFF, when you want to go back to inheriting the parent's class.
* **Cost Overlay** column (Activity and Resource Assignment rows): whether that item's cost scales with its sampled duration. See [The Cost Overlay column](#the-cost-overlay-column) below.
* **Correlation** column (Activity rows): link activities so their duration draws move together in the simulation (Activity-to-Activity correlation). This is a Pro feature. See [Correlation and Activity Mapping](/user-guide/risk/correlation-and-mapping).
* **Reports** column: once a simulation has been run or loaded, per-row icons open that node's Cost Monte Carlo, Schedule Monte Carlo, and Schedule Delay Drivers charts. See [Per-node reports (Reports column)](#per-node-reports-reports-column) for details.
The cascade is resolved at simulation time, so changes take effect on the next run.
### The Cost Overlay column
**Cost Overlay** ties a cost to a duration: when it is on, an iteration that samples a longer duration also scales that work's cost. It is only physically true for level-of-effort work (labor, support staff, equipment rental); a firm-fixed-price subcontract or a purchased material costs the same however long the schedule runs. See [Uncertainty and Cost Overlay](/user-guide/estimating/uncertainty-and-cost-overlay#cost-overlay) for the full model and the arithmetic.
The column shows one clock icon per row, and it appears only on the two row types that carry the flag. **Project, WBS, and Work Package rows are blank**, the same way Schedule Uncertainty is blank on Resource Assignment rows.
Read the icon two ways at once:
* **Color** is the answer: **orange** means this cost will stretch with its duration, **gray** means it will not. Scanning for orange shows you every cost that is coupled to the schedule.
* **Fill** tells you where an orange came from: **solid orange** was switched on for that row, **outline orange** is inherited from a mapped activity. Every gray is an outline, whether the off was set on the row or inherited; the tooltip and the picker say which.
So there are three things to see: solid orange (on, set here), outline orange (on, inherited), outline gray (off).
Hovering names the state in full, for example "Cost Overlay is ON, inherited from activity ESP-M035". Clicking opens a small picker with the states written out and the current one checked:
| Row type | Choices | Notes |
| ----------------------- | ---------------- | --------------------------------------------------------------------------------------------------------------------------------- |
| **Activity** | On, Off | An activity is always explicitly one or the other; it has nothing to inherit from. |
| **Resource Assignment** | Inherit, On, Off | **Inherit** (the default) follows the activities the assignment is mapped to, and the picker shows what it currently resolves to. |
Inheritance is **permissive**: an assignment mapped to several activities overlays if **any** one of them has overlay on. An assignment mapped to no activity has nothing to inherit, so it reads "No mapped activity" and never overlays; set it to **On** explicitly if it should.
Because inheritance flows through the activity mapping rather than the tree, setting overlay on an activity changes the resolved state of every assignment mapped to it, wherever those assignments sit in the tree. The cascade reloads after each change so those rows update immediately.
Turning overlay on across a large part of the project can materially raise the P80 and P90 contingency, because it couples cost growth to schedule growth. Flag only the work that genuinely scales with time. Every change is recorded in the activity's or assignment's history so the estimate basis can be traced.
Overlay only changes a number when the activity's duration actually varies, which means the activity needs a schedule Uncertainty Class to sample from. This makes it a Pro-tier control in practice: the Lite tier runs risk events only, so nothing samples a duration and overlay has nothing to scale.
## Filtering which risks are included
A floating filter bar (the **Filters** pill, collapsed by default on the **Project:** row) decides which risks enter the simulation. The **same filters drive both the Monte Carlo run and the Scenario Planner**, so the planner always shows exactly the risk set being simulated.
* **Risk Status** (multi-select): defaults to **Active**. Add Proposed, Realized, Retired, or Deprecated if you want them in the run.
* **Classification**: **Threats Only** (default), Opportunities Only, or All. Threats add cost and extend the schedule; opportunities reduce cost and pull the schedule in.
Click the pill to expand the bar, **Change Filters** to edit, and **Clear all** to reset to the defaults. The bar also shows a read-only **Project** chip.
## Running a simulation
The **Run New Simulation** area holds the controls for a run:
* **Mode** (the **Show** choice): pick what the simulation includes before you run it (see below).
* **Simulation Iterations**: how many Monte Carlo iterations to run (default 5000). More iterations give smoother, more stable percentiles at the cost of a longer run.
* **Scenario** dropdown and **Edit** button: optionally run against a saved scenario (see [Scenario Planner](#scenario-planner)).
* **Random Seed** (optional): leave blank for a fresh random seed, or paste a seed to reproduce an earlier run exactly.
* **Run**: starts the simulation. The chart is hidden while the run is in progress and appears when results are ready.
### Simulation modes
You choose the mode before running, because changing it requires a new run:
| Mode | What it samples | Use it to answer |
| ---------------------------------------- | ------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- |
| **Combined (Uncertainty + Risk Events)** | Estimating uncertainty and risk events together, with cost-schedule correlation | "What is my total exposure?" This is the default and the basis for the contingency decomposition. |
| **Risk Events Only** | Risk-register events against deterministic baselines | "How much exposure comes from my risks alone?" Matches the legacy cost Monte Carlo behavior. |
| **Uncertainty Only** | Estimating uncertainty against the baseline, no risks | "How much spread comes from my estimate quality alone?" |
The chosen mode is echoed in the chart subtitle so a saved or exported chart is self-describing.
### Reproducibility (Random Seed)
Every run uses a random seed, shown with the results and saved with the simulation. To reproduce a run, copy its seed into the Random Seed input (the **Reuse seed** button does this for a loaded simulation) and run again with the same mode and inputs. This is also the clean way to do before/after comparisons: hold the seed fixed and change one input.
## Scenario Planner
The **Monte Carlo Scenario Planner** is a what-if workbench. It lets you model the effect of turning specific **risks** off, or turning individual **response actions** on or off, without changing your live risk data, then run the simulation against that saved scenario. Open it with the **Edit** button next to the **Scenario** dropdown on the Run New Simulation row.
The planner lists the **same filtered risk set the simulation uses** (Active + Threats Only by default), so the scenario you build and the run you launch always match.
### Loading and saving scenarios
The **Load Scenario** dropdown starts on **Default (Current State)**, which mirrors your live data with every risk and active response action turned on. Use the controls next to it to manage scenarios:
* **Add** (the **+** icon): start a new scenario. You must give it a name before you can toggle anything; the planner prompts you if you try.
* **Rename** (pencil) and **Delete** (trash) icons appear once a saved scenario is loaded.
* Every toggle **auto-saves** to the current scenario, so there is no separate save step once it is named.
### The summary bar
Three totals across the top update live as you toggle:
| Metric | Meaning |
| ------------------------------------------ | ------------------------------------------------------------------------------ |
| **Risk Total Exposure** | The probability-weighted cost exposure of the active risks before any response |
| **Scenario Mitigated Risk Total Exposure** | The exposure after the scenario's active response actions are applied |
| **Reduction** | Risk minus Mitigated Risk: the value the scenario's responses buy you |
### The risk table
Each risk is a row, with its response actions nested beneath it. The columns are **Active**, **Risk ID**, **Title / Action**, **Status**, **WBS**, **Work Package**, **Risk Probability (%)**, **Risk Cost Impact**, **Mitigated Risk Probability (%)**, **Mitigated Risk Cost Impact**, and **Mitigated Risk Exposure**.
* **Deactivating a risk** (click its **Active** badge) excludes it **entirely** from the scenario: its row is struck through and it drops out of the totals and the simulation. Use this to ask "what if this risk never materializes?"
* **Response actions** appear as sub-rows under each **Mitigate** risk, each with its own **Active** toggle and its **Probability Reduction**, **Cost Reduction**, and **Details** (the action's status and owner). Only the **active** actions' reductions are subtracted from the Risk values to compute the Mitigated Risk numbers, so **turning an action off restores its reduction** and raises that risk's Mitigated Risk exposure. Use this to test "what if we drop this mitigation?"
* The **Mitigated Risk Probability / Mitigated Risk Cost / Mitigated Risk Exposure** values are color-coded: **green** when the response improves the number, **red** when it is worse, and black when unchanged.
The selected scenario's name is recorded with the run and shown in the results, so a saved or compared simulation always says which scenario produced it.
## Reading the Monte Carlo Cost Results
### The cost distribution chart
The chart overlays a histogram of iteration costs (the frequency bars) with the cumulative S-curve. The S-curve answers "what is the probability the project comes in at or below cost X?" Percentile lines (P10, P50, P80, and so on) can be toggled on from the Percentile Summary table.
### Risk and Mitigated Risk
When risks carry both Risk and Mitigated Risk impacts, the chart can show two curves:
* **Mitigated Risk**: the outcome assuming your planned mitigations and response actions are in place. This is the primary view.
* **Risk**: the outcome before mitigation.
Use the toggles above the chart to show either curve or both. The gap between them is the value of your mitigation plan.
The **Risk** and **Mitigated Risk** labels are configurable per project (an administrator can rename them); "Risk" and "Mitigated Risk" are the defaults. They are used consistently on the chart toggles, tables, and comparison views throughout this page.
### Percentile Summary
The Percentile Summary table lists the cost at each confidence level (P10, P50, P70, P75, P80, P90). When both Pre and Post curves are shown, it also shows the Reduction and % Improved between them. Each row has a Show checkbox to draw that percentile's line on the chart.
Use **+ Add** to add a custom percentile, either by confidence level (for example, P85) or by a target cost value (which the table resolves to the percentile it falls at).
### Contingency Decomposition
The **Contingency Decomposition** panel answers "where does my contingency come from?" at a confidence level you choose. It appears after a **Combined** run, since the breakdown needs all sources present.
It shows:
* **Combined P\[X]**: the total project cost at the selected confidence level.
* **Deterministic baseline**: the point-estimate cost with no uncertainty and no risks.
* **Contingency (P\[X] minus baseline)**: the buffer you would hold.
The contingency is then split by source:
| Source | Meaning |
| ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **Estimating uncertainty** | The portion driven by the spread in your cost estimates |
| **Cost overlay** | The portion driven by time-driven costs scaling as the schedule runs long (shown only if items are flagged for [Cost Overlay](/user-guide/estimating/uncertainty-and-cost-overlay#cost-overlay)) |
| **Risk events** | The portion driven by the Risk Register |
| **Interaction / correlation** | The remainder, so the parts sum exactly to the combined contingency |
#### Choosing a confidence level
The **Confidence Level** picker is populated from the percentiles in the Percentile Summary table (including any custom percentiles you have added) and defaults to **P80** (a commonly cited level for cost contingency). Your choice is remembered per project. Changing it updates the numbers instantly, with no need to re-run, because the breakdown is computed from the cached iteration data.
#### Understanding the Interaction line
The three sources are estimated independently, so they do not always sum exactly to the combined contingency; the difference is shown as **Interaction / correlation**. This line is usually **negative**, and that is expected and useful: it is unlikely that estimating uncertainty and risk events both land on their high-side outcomes in the same iteration, so the contingency you actually need is less than the sum of the parts. A negative interaction is the diversification benefit of analyzing the sources together. A positive interaction would mean the sources tend to spike together (positive correlation).
### Monte Carlo by WBS
The **Monte Carlo by WBS** tree rolls the results up the WBS structure, so you can see each WBS element's contribution to the total at each percentile. Use the **View** selector to switch the rollup between the **Mitigated Risk** (default) and **Risk** contributions. Use **Hide \$0 Rows** to suppress empty branches, and the **Collapse All / Expand All** link to toggle the tree.
Within every iteration the dollars roll up exactly, so the **Mean** column ties out top to bottom. The **percentile** columns do not sum, and should not: a parent's P80 comes from a different iteration than each child's P80, so a parent's upper percentiles are lower (and lower percentiles higher) than the sum of its children. That diversification is the point of project-level Monte Carlo. The `?` icon on the rollup explains this; use the **Mean** when you need a column that adds up.
### Cost Sensitivity (Top Drivers)
After the WBS rollup, two tornado charts rank what is driving the spread in project cost, so you know where to focus attention.
**Cost Risk Sensitivity Analysis (Top 10 Drivers)** ranks the **risk events** by their effect on project cost: one signed bar per risk, sorted by magnitude. Threats point right and are shown in **red** (they add cost); opportunities point left and are shown in **blue** (they save cost). The axis is the signed cost impact (threat positive, opportunity negative). Clicking a bar opens that risk for editing, so you can go straight from "this risk moves my number the most" to adjusting it.
**Cost Sensitivity by WBS (Top 10 Uncertainty Drivers)** ranks the **estimating uncertainty** instead: each bar is one WBS element's **P10 to P90 baseline-cost range**, so a longer bar means a wider, less certain estimate. Where the risk tornado shows which discrete risks matter most, this shows which parts of the estimate itself carry the most uncertainty and might warrant tighter estimating or a basis-of-estimate review.
## Reading the Monte Carlo Schedule Results
The **Schedule** tab presents the same run from the finish-date angle: instead of "how much will the project cost?", it answers "**when will the project finish?**". It comes from the very same iterations as the cost results; because cost and schedule are sampled with shared random draws, a long-duration iteration also tends to be a high-cost iteration (the cost-schedule correlation), which is why running them together is more realistic than treating them separately.
The Schedule tab mirrors the Cost tab section for section, but everything is expressed in **finish dates and working days** rather than dollars.
### The finish-date distribution chart
The chart overlays a histogram of iteration finish dates (the frequency bars) with the cumulative S-curve. The S-curve answers "what is the probability the project finishes **on or before** date X?" Toggles above the chart show the **Risk** curve, the **Mitigated Risk** curve, and the frequency bars; percentile lines can be turned on from the Percentile Details table.
### Risk and Mitigated Risk
When risks carry both Risk and Mitigated Risk schedule impacts, the chart can show two finish-date curves:
* **Mitigated Risk**: the finish distribution assuming your planned mitigations and response actions are in place. This is the primary view.
* **Risk**: the finish distribution before mitigation.
The gap between the two curves is how much earlier your response plan is expected to bring the project in. (In **Uncertainty Only** mode there are no risks, so Risk and Mitigated Risk are identical and the toggles stay hidden.)
### Percentile Details
The Percentile Details table lists the **finish date** at each confidence level (P10, P50, P70, P75, P80, P90), with a Description and, when both curves are shown, the **Delta in days** between Pre and Post. Each row has a Show checkbox to draw that percentile's line on the chart.
Use **+ Add** to add a custom percentile either **by confidence level** (for example, P85) or **by a target finish date** (which the table resolves to the percentile that date falls at), so you can answer "what is my confidence of finishing by this committed date?"
### Schedule Contingency Decomposition
The schedule analog of the cost [Contingency Decomposition](#contingency-decomposition). It shows the number of days between the deterministic baseline finish and the finish date at the confidence level you choose, then splits that schedule buffer by source:
| Source | Meaning |
| --------------------------------- | --------------------------------------------------------------------------------------------------- |
| **Activity-duration uncertainty** | Schedule growth from the Monte Carlo on activity durations alone (no risk events) |
| **Risk events** | Schedule growth from the risks alone (deterministic durations plus Mitigated Risk schedule impacts) |
| **Interaction / correlation** | The remainder, so the parts reconcile to the combined contingency |
There is **no cost-overlay source** here, since that is a cost-only effect. As on the cost side, the **Confidence Level** picker (default **P80**) is populated from the percentiles in the table and recomputes instantly from the cached iterations, and the **Interaction / correlation** line is usually negative, the diversification benefit of analyzing duration uncertainty and risk events together.
#### Calendar days versus working days
Each figure is shown in **calendar days** first, with the **working-days** equivalent alongside (for example, "134 calendar days (95 working)"):
* **Calendar days** is simply the elapsed time between the two dates. It is unambiguous, and it matches the finish dates and the Delta column used elsewhere on the page (and the Schedule Delay Drivers exposure).
* **Working days** is that same buffer expressed on the **project's default calendar**, that is, its work week and holidays. This is the figure schedulers usually hold as reserve. The two differ by the calendar's working ratio (roughly 5 in 7 for a standard five-day week, so 134 calendar days is about 95 working days).
The working-days figure is always counted on the **one** default calendar. If the activities on the driving path use different calendars, working days is still measured against the default calendar (and falls back to a plain five-day week only if no default calendar is set), so it can slightly misrepresent a finish that was actually driven on another calendar. Calendar days has no such dependency, which is why it leads. When in doubt, read the calendar-days figure.
### Schedule Sensitivity (Top Drivers)
The Schedule tab ranks what drives the finish date with two tornado charts, parallel to the cost [Cost Sensitivity](#cost-sensitivity-top-drivers) view:
* **Schedule Risk Sensitivity (By Risk)** ranks the **risk events** by their effect on the finish date: threats that **extend** the finish point right (red), opportunities that **compress** it point left (blue). It is shown for Combined and Risk-Only runs.
* **Schedule Sensitivity Analysis (By Activity) — Top 10 Drivers** ranks the **activities** by how many days their duration variation moves the project finish, so you can see which activities the schedule is most sensitive to.
### Schedule Criticality
The **Schedule Criticality — Top 10 Activities** chart shows the **percentage of iterations in which each activity sat on the critical path** (its total float was effectively zero). It complements the by-activity sensitivity tornado: sensitivity measures **how much** an activity moves the finish when its duration changes, while criticality measures **how often** the activity is actually on the binding path. An activity that is both highly sensitive and frequently critical is a prime candidate for schedule risk mitigation.
### Schedule modeling rules
A few schedule-modeling rules are worth knowing (the `?` icon on the Schedule tab summarizes them, and the [methodology page](/user-guide/concepts/monte-carlo-methodology) covers them in full):
* Schedule **constraints are ignored** (a constrained date never moves, which would distort the analysis).
* **Start-to-Finish (SF)** links are not modeled; convert them to Finish-to-Start.
* Activities use their **own calendar**, else the **project default**, else a **5-day work week**.
* A risk's schedule impact extends its **mapped critical-path activities**; a risk with **no activity mapping** extends the **project finish** directly.
## Per-node reports (Reports column)
The Uncertainty Cascade tree has a **Reports** column (just after Correlation) that lets you drill into the Monte Carlo result for any single node, not only the project total. Each row shows up to three icons, and they become active once you have **run or loaded** a simulation (they appear grayed out until then):
* **Cost Monte Carlo** (green): the node's total-cost distribution. Shown on Project, WBS, Work Package, and Resource Assignment rows.
* **Schedule Monte Carlo** (blue): the node's finish-date distribution. Shown on Project, WBS, Work Package, and Activity rows.
* **Schedule Delay Drivers** (purple): the schedule drivers of the node's finish. Shown on the schedule-driving rows (Project, WBS, Work Package, Activity).
Hover an icon for a quick preview; click it to open the full modal.
### The Cost and Schedule modals
The Cost and Schedule modals are the **same charts as the main results**, scoped to the node you picked: the frequency histogram with the cumulative S-curve, the Risk and Mitigated Risk curves with their Show and Frequency toggles, the deterministic baseline and percentile crosshair lines, and a Percentile Details table (including any custom percentiles created in the main results).
All three reports share a **single modal** with a tab strip at the top: **Cost**, **Schedule**, and **Schedule Delay Drivers**. A tab appears only when that report exists for the node, so a Resource Assignment shows just **Cost**, an Activity shows **Schedule** and **Schedule Delay Drivers**, and a Project, WBS, or Work Package shows all three. Whichever icon you click opens the modal on that report's tab, and you can switch between the node's reports without leaving the modal.
The **Project** row matches the main Cost and Schedule charts exactly. Rows below the project are summarized for compact storage (see [How per-node numbers roll up](#how-per-node-numbers-roll-up) below, and [Per-node distributions and bucketing](/user-guide/concepts/monte-carlo-methodology#per-node-distributions-and-bucketing) on the methodology page), so their percentiles can differ from the exact value by a fraction of a percent.
### How per-node numbers roll up
Cost and schedule roll up the tree differently, and the modals state this inline:
* **Cost** is additive in the mean: a parent's average cost equals the sum of its children. Its **confidence levels do not add up**, though: a parent's P80 is lower than the sum of the children's P80, because their risks rarely peak in the same iteration (diversification).
* **Schedule** is not additive at all: a parent's finish is its **latest-finishing child** (the critical path), so child finish dates never sum to the parent, and a child can even show a larger delay than its parent.
### Schedule Delay Drivers
The **Schedule Delay Drivers** modal ranks the predecessors that drive the selected node's finish, as horizontal stacked bars. Each bar is split into a **duration-uncertainty** segment plus one segment **per mapped risk event** (labeled with the risk's ID and title), so you can see how much of a driver's delay is baseline uncertainty versus specific risks. Controls at the top let you:
* Pick the **exposure P-level** (default P75). The bars rescale so they sum to the node's risk exposure (the P-level finish minus the deterministic finish) at that confidence level.
* Toggle **Risk** versus **Mitigated Risk** (shown when the project has mitigations).
Days here are **calendar days**, consistent with the finish dates and the Delta column elsewhere on the page. Like the Schedule Monte Carlo, this view follows the critical path and does not sum up the hierarchy.
## Saving and comparing simulations
* **Save**: give a run a title to save it. Saved runs store the full results, the seed, the mode, the percentiles, the contingency decomposition, and the finish-date distribution, so they reload exactly as they ran.
* **Load**: pick a saved run from the Previous Simulations list to restore its charts, tables, and decomposition. A reloaded run renders through the same path as a live run, so it looks identical.
* **Compare Simulations**: open the comparison view to put saved runs side by side (for example, a Risk run against a Mitigated Risk run, or two scenarios), so you can quantify the difference.
### Compare Monte Carlo Results
Click **Compare Simulations** to open the **Compare Monte Carlo Results** window. Add **2 to 5** saved simulations, and the window overlays their cumulative-probability S-curves and lines up their percentiles in a comparison table. It is the cleanest way to answer "how much better is plan A than plan B?"
Each comparison is built from a few controls at the top:
* **Add simulation**: pick a saved run from the dropdown and click **Add**. Each added run gets its own color, which is used consistently in the chart, the legend, and the table.
* **Results** (per run): choose **Mitigated Risk** or **Risk** for that run, so you can compare like-for-like or deliberately contrast a Risk run against a Mitigated Risk run.
* **Base**: the radio button marks one run as the **baseline** that the others are measured against. The base run is listed first, and every delta in the table reads "base vs other".
* **Remove** (trash icon): drop a run from the comparison.
#### Cost and Schedule tabs
The results are split into two tabs so you can compare a run on either dimension of the same iterations:
* The **Cost** tab overlays each run's cost S-curve and a **Percentile Comparison** table of the cost at each confidence level (P10 through P90, plus any custom percentiles). Deltas are shown in dollars and percent; **green** means the run is **lower** than the baseline (an improvement), **red** means higher.
* The **Schedule** tab does the same for the **project finish date**: it overlays each run's finish-date S-curve and lists the finish date at each percentile. Here deltas are shown in **calendar days**; **green** means the run finishes **earlier** than the baseline (an improvement), **red** means later.
In either tab, the **Show** checkboxes on the left of the table draw that run's percentile line on the chart, so you can mark, say, every run's P80 to see the spread at a glance.
The Schedule tab only includes runs that were saved **with** finish-date results. Older saved runs (from before schedule results were added) appear in the Cost comparison but are skipped on the Schedule tab; if fewer than two of the selected runs have schedule data, the tab shows a short message instead of a chart. Re-run and save those simulations to compare them on schedule.