The complete guide

Everything you need to know about McKinsey Solve

Solve (you may know it as the Problem Solving Game, or Imbellus) is McKinsey's gamified screen, and for most candidates it's the first hurdle after the application goes in. Here's what's in it and how to get ready.

Format & timing

Solve is a set of ecology-themed mini-games. They're looking at how you gather information, structure a problem and decide under time pressure. You don't need to know anything about consulting to do well. As of 2026 the lineup runs in this order:

  • 1 · Red Rock Study: data analysis (~30 min)
  • 2 · Sea Wolf: the "ocean cleanup" optimization game (~30 min)
  • 3 · Sustainable Future Lab: a behavioral, branching scenario (~20 min), included on longer invites

A 65-minute invitation means Red Rock and Sea Wolf. An 85-minute one adds the Sustainable Future Lab. Worth knowing: McKinsey records how you got there, not just where you ended up. How you explore, what you go back and change, how you pace yourself. It's all captured.

Ecosystem Building was dropped from the live assessment in 2025. Plenty of prep material out there still treats it as current, which is worth knowing if you're following an old guide. We've kept ours, clearly labelled, because it's a decent reasoning workout. You won't be asked to play it.

📊 1 · Red Rock Study

Time: ~30 minutes · Type: data analysis / quant

Red Rock is the first game and the one that feels most like an actual case. Two parts, and the flow only runs one way: once you leave a phase, it locks behind you.

Part 1: Study (Investigation → Analysis → Report)

  • Investigation: read a page of text, tables and charts. Only the Research Journal is available (no calculator yet). You drag/collect the values you'll need into it, because they won't be available later.
  • Analysis: the calculator appears. Solve 2–4 quantitative questions using your collected values; each answer is logged back into the Journal.
  • Report: use your answers to fill blanks in a written summary, choose a chart type, and complete a table. You cannot return to earlier phases.

Part 2: Cases

Six short, unrelated questions on the same theme. About two-thirds of everything in Red Rock is math: percentages and percentage points, ratios, weighted averages, mean/median/mode, simple probability. None of it goes past arithmetic. What catches people out is speed, and losing track of which number feeds which calculation.

→ Deep dive: the full Red Rock walkthrough with timing splits
→ Practice Red Rock now

🦠 2 · Sea Wolf

Time: ~30 minutes · Type: analytical / optimization

You clean up three contaminated ocean sites. Each microbe has three attributes (1–10) and one trait; each site defines target attribute ranges plus one desired and one undesired trait. You work each site through a fixed, one-way sequence of steps:

  • Step 1 · Build profile: pick the two characteristics that matter most for the site.
  • Step 2 · Categorize microbes: for each microbe, use it here, save it for the next site, or reject it (locks on confirm; saved microbes reappear at the next site's Step 0).
  • Step 3 · Build prospect pool: you're given six prospects, then add four more by picking 1 of 3 across four rounds.
  • Step 4 · Create treatment: choose exactly three microbes. Efficiency starts at 100% and loses 20% for each attribute mean out of range, 20% if no desired trait is present, and 20% for every microbe carrying the undesired trait.

Because it's the mean that matters, one extreme microbe can be pulled back into range by the two beside it. That's the whole game. Picking the three best individuals is not the same as picking the best three.

→ Deep dive: the full Sea Wolf strategy guide
→ Practice Sea Wolf now

🌱 3 · Sustainable Future Lab

Time: ~20 minutes · Type: behavioral / judgment

The newest module (added in 2026). You lead an environmental research team through 13 prompts: one drag-and-drop priority ranking, then 12 branching scenario decisions. Each choice triggers a consequence and reshapes the next situation you face.

There are no correct answers here, which is exactly why people flounder. What McKinsey looks at is consistency: whether you weigh trade-offs the same way from one decision to the next, across five traits (prioritization, deciding under uncertainty, reading messy information, balancing trade-offs, and handling stakeholders). Decide what kind of leader you are in the first two questions, then don't drift.

→ Deep dive: the full Sustainable Futures Lab guide
→ Practice Future Lab now

🐙 Ecosystem Building Legacy · retired 2025

Status: removed from the live assessment in 2025 · Type: constraint satisfaction

Ecosystem used to be the centrepiece. You built a food chain of eight species from a big pool and placed it somewhere every one of them could survive: each needs something to eat, has to tolerate the habitat, and has to end up with positive calories once feeding resolves from the top down. It's not in a real sitting any more. It's still a good constraint-satisfaction drill, so we've left it up.

→ Practice the legacy Ecosystem game

FAQ

What games are actually on the test now?

As of 2026: Red Rock Study, Sea Wolf, and (on longer invitations) the Sustainable Future Lab. The Ecosystem game was retired in 2025.

Is using a practice simulator cheating?

No. It's the same as working through sample questions before any other test. Our scenarios are generated, so you're building the speed and habits the assessment measures rather than memorizing answers you'll see again on the day.

Are these the exact games McKinsey uses?

No, and be suspicious of anyone claiming otherwise. We rebuilt them from publicly documented mechanics. The real numbers, artwork and scoring weights are McKinsey's, they're not published, and they change.

Do I need anything installed?

No. It all runs in the browser. No Excel, no download, and you don't need an account to start.