McKinsey Sea Wolf Game

TL;DR: Sea Wolf is a constraint-solving and optimisation task in McKinsey Solve. Candidates in 2026 describe picking microbes to clean up three contaminated ocean sites, in about 30 minutes in total, and most say it comes second, after Redrock. It is a separate game from Ecosystem Building, which many older guides get wrong. McKinsey itself names no games and says "there's no need to prepare", so treat every walkthrough, this one included, as candidate reports. The reasoning it rewards (map the constraints before you commit, then optimise inside them) is still worth warming up.

What McKinsey itself says

McKinsey's Solve page calls Solve "a gamified assessment created to showcase your problem-solving abilities" and says "There's no need to prepare—approach it with curiosity and creativity." Its Solve FAQ (September 2026) adds the practical rules, checked on 4 November 2026:

• No business or gaming knowledge needed. Each task starts with an untimed tutorial.
• Length is in your invitation. McKinsey doesn't publish a standard duration.
• PC or Mac only. It can't be taken on an iPad or phone. Sound isn't needed, and a mouse is optional.
• Once a year. "Solve scores are valid for one year, and you may complete the assessment only once every 12 months."
• Scenarios vary. McKinsey says Solve "is designed so that no two individuals have the same parameters and combinations of data", and you "may also be asked at random to take an additional task in person and/or to explain your logic".
• No tools or help. Applications, websites, AI tools, pre-prepared notes and screenshots are not allowed, except as an approved accommodation.
• Languages: English, Spanish, Portuguese or Japanese.

McKinsey doesn't name Sea Wolf, Redrock or any other game on these pages. Everything below about the modules comes from candidates and prep sites.

Where it sits in Solve (candidate reports)

Candidates in 2026 most commonly describe this shape:

1. Redrock Study: research and data interpretation, usually first.
2. Sea Wolf: constraint-solving and optimisation, usually second.
3. Sustainable Futures Lab: judgement and prioritisation, on longer invitations.

Invitation length is the clearest signal. A prep site updated on 2 November 2026 describes 65 minutes as Redrock plus Sea Wolf, 85 minutes as those two plus a 20-minute Sustainable Futures Lab, and 95 minutes as those two plus a newer 30-minute version (RoadToOffer). Another, updated on 23 September 2026, describes a 95-minute, three-module battery in that order (StrategyCase). Earlier versions of this page said Redrock must be finished before Sea Wolf opens. We could not confirm that rule in any current source, so we no longer state it.

A note on older material: much of it describes Ecosystem Building as the main event and treats "sea wolf" as one of its nicknames. That doesn't match what candidates report now: the 2026 breakdowns above don't list Ecosystem as a current module. The ecosystem guide explains what it was and why it is still useful background.

What the Sea Wolf task looks like (candidate reports)

Prep sites that rebuilt the game from candidate accounts describe the same core: you treat three ocean sites contaminated with plastic, and for each you choose three microbes whose combined attributes and traits fit the site's requirements (StrategyCase). StrategyCase says each microbe has three number attributes (density, energy and size, on a 1 to 10 scale) and traits such as aerobic or heat resistant, and each site lists traits it wants and traits it doesn't. RoadToOffer describes sorting cultures into current, later or reject over several rounds before the final choice, and both put Sea Wolf at about 30 minutes for all three sites (RoadToOffer). StrategyCase suggests about 12 minutes for the first site and 9 for each of the others.

These are reconstructions, not McKinsey's specification, and McKinsey says the parameters differ for every candidate.

What Sea Wolf measures

The scenarios change and McKinsey publishes no rules, so treat any specific walkthrough, including the one above, as indicative rather than a script. What is consistent is the kind of thinking being read.

Working within hard constraints. You are given rules that cannot be broken and resources that will not stretch. The first job is understanding what the constraints actually forbid, before you begin arranging anything. Candidates who start optimising before they have mapped the limits routinely have to unwind their own work.

Optimising, not just satisfying. A solution that merely obeys every rule is the floor, not the goal. There is generally a better and a worse legal answer, and the gap between them is where the score lives.

Holding several interacting variables at once. Change one thing and two others move. This is the part that loads working memory hardest, and the part that punishes an approach kept only in your head.

Time discipline. The clock does not forgive exploration you never converted into a committed decision.

Process is read alongside outcome

This is the single most useful thing to internalise about Solve as a whole, and it applies squarely here.

Prep sites and former consultants widely report that Solve records how you worked, not only what you ended with: the order you did things in, what you revisited, whether your decisions followed a consistent logic, whether you ran out of clock. McKinsey doesn't publish its scoring, but its FAQ does say you may be asked "to explain your logic", which fits. Two candidates who land on similar answers can be read very differently: one methodically, one by thrashing until something stuck.

The practical consequence: do not abandon a sensible method halfway because it feels slow. Erratic strategy-switching reads worse than a steady approach that finishes slightly short of a perfect result. How the McKinsey Solve score is judged covers why there is no published pass mark and why chasing a target number is the wrong goal.

How to practise it

You can't rehearse Sea Wolf itself, and McKinsey says you don't need to. You can warm up the capacities it loads, and unlike the scenario, your progress there is measurable.

• Pathfinder puzzle — the closest analogue. Route and constraint reasoning, and specifically the look-before-you-move habit that constraint tasks reward and improvisation punishes.
• Inductive reasoning — inferring the governing rule from what a system does, rather than from instructions.
• Mental math — fast numerical judgement, since optimisation questions are usually comparisons rather than precise calculation.
• Task switching — holding one rule set while another competes for attention, which is what interacting constraints feel like in practice.

Each run is scored on the server against fixed norms and returns a percentile, so one cold pass across these takes about fifteen minutes and tells you which is your weakest. That is where your remaining time should go — see how we score for what those percentiles are and are not.

What first attempts look like. Our score benchmarks show how people did the first time they played each game. As of November 2026, the median first attempt on GamePrep was 66 out of 100 on inductive reasoning (209 players), 16 correct answers on mental math (347 players) and 60 out of 100 on task switching (160 players). On the pathfinder puzzle, the median was 5 puzzles solved (144 players), and most first-timers already reach the top of that scale, so treat it as a habit drill rather than a ranking. These are people practising, not McKinsey candidates, and none of it is a cut-off.

Preparing in the days you have

• Baseline first. Play each related game once, cold, and write down your percentiles. Guessing at your weak spot without measuring is how people drill the wrong thing.
• Practise finishing. Set a timer and force a decision before it runs out. Running out of clock mid-exploration is one of the most common self-inflicted Solve outcomes.
• Rehearse a default method, so that under pressure you have something to fall back on instead of improvising: read every constraint, write down what is forbidden, find any legal solution, then improve it.
• Read the in-assessment tutorial properly. McKinsey says each task opens with an untimed tutorial, so take the time. It is the only authoritative description of the version you are sitting.
• Fix the environment. A PC or Mac (McKinsey doesn't allow tablets or phones), a quiet room and notifications off. If something breaks, contact the support address in your invitation at the time: McKinsey says it can't reset your link otherwise.
• Work alone, with nothing open. McKinsey bans AI tools, websites and pre-prepared notes during Solve, and may ask you to explain your logic later.

For the wider picture, start from the McKinsey Solve guide, and Redrock since it comes first and gates this one.

FAQ

Is Sea Wolf the same as the ecosystem game? No. Older guides commonly said so, but candidates in 2026 report it as a distinct module: choosing microbes to treat contaminated ocean sites. Ecosystem Building is a separate scenario that 2026 breakdowns no longer list.

What order do the McKinsey Solve games come in? Candidates usually report Redrock first, then Sea Wolf, then Sustainable Futures Lab where it appears. McKinsey doesn't publish an order.

How long is the assessment? McKinsey puts the length in your invitation. Candidates report about 65 minutes for Redrock and Sea Wolf, and 85 or 95 minutes when Sustainable Futures Lab is added. Sea Wolf itself is reported at about 30 minutes for three sites.

Can I take Solve on my phone? No. McKinsey's FAQ says Solve is only available on a PC or Mac. A mouse is optional.

Can I get the answers to Sea Wolf? No, and they wouldn't help. McKinsey says no two candidates get the same parameters and data, and it may ask you to explain your logic afterwards. A memorised answer can't fake the working.

Is there a pass mark? None is published, and you won't be shown a score. McKinsey says your Solve results are considered together with the rest of your application and any other assessments. Scores are valid for a year, and you can take Solve only once every 12 months.

Does speed or accuracy matter more? A clean, systematic solution beats a fast one that violates a constraint. Managing the clock matters, but rushing into an illegal answer is worse than finishing slightly short.

What if my invitation names a game not listed here? Believe your invitation. The module set changes between cycles; this page reflects what candidates most commonly report in 2026, not a guarantee.

GamePrep is an independent practice platform and is not affiliated with McKinsey & Company. McKinsey does not publish Solve's rules, scoring or cut scores, and the assessment changes between cycles — your invitation and its in-assessment tutorial are the only authoritative sources. We provide related skill practice only — no leaked questions or answers.