McKinsey Sea Wolf Game: The Complete 2026 Guide

What is the McKinsey Sea Wolf game?

Sea Wolf, sometimes called the microbe game or ocean cleanup game, is the second module in McKinsey Solve, following the Redrock Study. It became a standard part of the assessment globally in early 2025 and runs 30 minutes in the 2026 format. The premise: three polluted ocean sites need cleaning, and for each one you assemble a team of three microbes whose combined properties meet that site's requirements.

Despite the biology costume, Sea Wolf has nothing to do with scientific knowledge. It is a constrained optimization exercise: every site defines numerical ranges its cleanup team must hit and qualitative traits it wants or cannot tolerate, and you search a roster of candidate microbes for a trio that satisfies everything at once. The skill being measured is structured decision making under constraints, with a heavy dose of fast arithmetic.

Sea Wolf appears in both Solve versions: the 65-minute test (Redrock plus Sea Wolf) and the 85-minute test, which adds the Sustainable Futures Lab. As with Redrock, an untimed tutorial runs before the clock starts.

How a Sea Wolf site works

Each site gives you the same three ingredients, and the entire game is the interaction between them.

Site requirements. A set of numerical attribute ranges, for example an oxygen output between 40 and 50, plus a trait constraint: traits the site wants on the team and traits it cannot tolerate.

The microbe roster. A list of candidate microbes, each with a value for every scored attribute and a set of qualitative traits. The roster is deliberately larger than you need, and no microbe is labeled good or bad.

Your selection. Exactly three microbes. The site evaluates the average of your trio's attribute values against each required range, and checks your trio's combined traits against the desired and undesired lists.

The averaging requirement is the whole test

The single most important fact about Sea Wolf: the site checks your trio's average, not each microbe individually. This changes everything about how you should search the roster.

Intuition says to look for microbes that individually sit inside every range and combine three of them. The test is built to punish that intuition. Optimal teams routinely include microbes that are far outside a range on their own, because an overshooting microbe and an undershooting microbe can average into the target together. Meanwhile, a microbe that looks individually perfect may carry an undesired trait that disqualifies it from the team entirely.

So the working method is: filter by traits first, since an undesired trait is an instant elimination that requires zero math, then think in terms of complements. If the oxygen range is 40 to 50 and you like a microbe at 62, you are now shopping for partners around the low 30s, not for more microbes near 45.

Try a Sea Wolf-style selection

A simplified, original example with one site and four candidate microbes. Pick the three that satisfy the site. Watch what happens to the averages as you select.

Notice that microbe A is far outside both ranges on its own, and microbe C sits perfectly inside them. Whether each belongs on the team is a different question entirely.

How Sea Wolf efficiency scoring works

Each site is scored as a cleanup efficiency that starts at 100 percent and falls with every constraint your team misses. Based on consistent candidate reports, the deduction structure works like this:

Two implications follow. First, partial credit is real: a team that hits two of three ranges scores meaningfully better than one that hits none, so a rushed but sensible selection beats an unfinished site. Second, like every Solve module, Sea Wolf also feeds the process score: how you explore the roster and how you spend time is recorded alongside what you submit. For how the product and process scores combine across the assessment, see the Solve overview guide.

Time strategy: pacing 30 minutes across 3 sites

Thirty minutes for three sites suggests ten minutes per site, and unlike Redrock there is no cheap section at the end to protect. The risk profile is different: the danger is perfectionism on an early site starving a later one.

Within a site, the order of operations is fixed: read the requirements, eliminate every microbe carrying an undesired trait, then build the trio around complements using the averaging logic. If a perfect trio will not appear, take the best near-miss; the deduction structure makes one missed range survivable and a timeout catastrophic.

The five most expensive Sea Wolf mistakes

1. Shopping for individually perfect microbes

   The roster is built so that in-range loners rarely form a valid team. The averaging requirement means complements win: a high and a low can be worth more together than two comfortable middles.

2. Skipping the trait filter

   An undesired trait is a flat 20 percent deduction that no arithmetic can rescue. Striking those microbes first shrinks the roster and prevents the most avoidable scoring loss in the game.

3. Recomputing averages from scratch

   For three microbes, sum the values and compare against three times the range bounds. Working with sums instead of dividing every candidate trio saves a surprising amount of clock.

4. Perfecting site 1

   Every site is worth the same. Five extra minutes polishing site 1 from 90 to 100 percent is a terrible trade if it leaves site 3 half-finished.

5. Forgetting the desired traits

   Undesired traits get all the attention, but where a site lists wanted traits, a team missing them leaves points behind too. Read the full trait rule, not just the red half.

How to prepare for Sea Wolf

1. Drill fast averages and sums

The arithmetic is addition and comparison done dozens of times under pressure. Practice summing three two-digit numbers mentally and comparing against bounds. Our free math drills build exactly this speed.

2. Internalize the complement habit

Take any target range and a value outside it, and practice naming the partner values that pull the average back inside. Once this is reflexive, the roster stops looking like noise and starts looking like pairs.

3. Run full sites under time

The trait filter, the complement search, and the submit-and-move-on discipline only become automatic by playing complete sites against a clock. Treat your first timed run as diagnostic: find which site cost you the most efficiency and whether the cause was a range miss, a trait oversight, or pacing.

Sea Wolf: frequently asked questions

The rest of the Solve assessment