The four layers

A Topolog plan models four things at once: how long it takes, what it costs, how it can end, and how to make it better. That is a lot to absorb the first time you open the IDE. The fastest way in is to learn it in four layers, each one building on the last. Get comfortable with the time layer before you add money; read money before outcomes; optimize last, once the first three are second nature. Every tab in the IDE belongs to one of these layers.

Layer 1: Time

Start here, and stay here until it feels natural. A plan is a graph of tasks with dependencies. Each task carries an estimate and an uncertainty (its CV), and Topolog runs thousands of monte-carlo passes through the graph to produce a full distribution of finish dates rather than a single deadline.

Graph - the structure: what depends on what, and what can run in parallel.
Spectrum - the full distribution of finish dates. A narrow histogram means low risk; a wide one means a long tail. The Sentinels list beneath it breaks the finish down by which ending each run reached.
Critical path - the chain of tasks that shapes the right tail, so you know where pulling the date in is even possible.
Causal threads & parallelism - the resource-respecting schedule: who does what, and when.

Goal of the layer: quote a range instead of a single date, and know which one or two tasks actually drive it.

Layer 2: Money

Once you trust the time forecast, add the cash picture. Give the plan an Account, put costs on the tasks that spend, and revenue on the tasks that earn. The same monte-carlo pass that produced the spectrum now also tracks your balance across every run.

Money - the balance trajectory (P10 / P50 / P90), expected end balance, worst-case drawdown, and the chance of running out. The Cashflow breakdown lists what moves the balance and by how much, so a headline profit becomes a set of line items you can point at.
Pareto - when the plan has decisions (how much to spend, which option to take), the frontier shows the trade-off between spend and confidence.

Goal of the layer: understand the cost to finish, and whether the path you are on stays solvent.

Layer 3: Outcome

A plan rarely has one ending. Tasks produce outcomes, edges gate on them, and sentinels mark the end-states you can land in (typically success, partial, and failure). This layer is about reading how the plan can finish, not just when.

Sentinels (under Spectrum) - every ending with its probability and typical date. Expand one to see which outcomes distinguish it from the plan overall. Failure modes are usually compositional: you discover them here rather than guessing them up front.
Exchange - the deliverables you expect to be holding when the plan ends.

Goal of the layer: know your success paths and, more importantly, your failure modes, and why each one happens.

Layer 4: Optimization

Now the first three layers become levers. Change an estimate, reassign a task, or commit a decision, and watch the spectrum, the cashflow, and the frontier move in response. This is where a plan stops being a description and starts being a tool.

Decompose the highest-CV task on the critical path; its daughter tasks usually estimate more confidently and the spectrum tightens.
Re-price an over-optimistic gate’s prior and watch a false left peak collapse to where it should have been.
Pick the Pareto point that buys exactly the confidence you need, and no more spend than that.

Once the plan is active, recorded reality collapses uncertainty as you go, and all four layers sharpen together. Goal of the layer: move the plan toward earlier, cheaper, or more likely, on purpose.

Where to go next