All guides

Sea Wolf guide: how the game works and how to play it well

The stages of each site, how to think about averages, how to split the 30 minutes and the mistakes that cost the most points in Sea Wolf.

Updated October 28, 2026. 8-minute read.

Practice Sea Wolf for free
Video walkthrough. The full written guide is below.

Sea Wolf is a game about constraints, not biology. At each of the 3 sites you pick 3 microbes whose averages fall within the required ranges, and you win it by practicing the flow and looking after your 30 minutes.

What it is and what it measures

Sea Wolf (also called Ocean Cleanup or Ocean Treatment) is usually the second game in the Solve, after Redrock. It lasts about 30 minutes and has 3 ocean sites polluted with plastic. At each site you build a treatment with exactly 3 microbes.

  • Averages: each microbe has 3 numeric attributes (density, energy and size, on a scale of 1 to 10). The average of your 3 microbes on each attribute has to fall within the range the site asks for.
  • Traits: each site asks for one desired trait and forbids one undesired trait. It is enough for one of your 3 microbes to have the desired trait, and none of them can have the undesired one.
  • It isn’t biology: the names are just labels in the game. You don’t need to know anything about microbiology.

The scoring isn’t public. One guide describes each site as starting at 100% and dropping 20% for every condition that isn’t met (an average out of range, a missing desired trait or an undesired trait in the mix). Others talk about a weighted score. Treat it as a reference, not the official formula. What they all repeat: covering every condition is worth more than nailing a single one.

The stages of each site

The 3 sites repeat the same mechanics, so what you learn at the first one you reuse at the other two.

  1. Profile. You read the site’s requirements (3 ranges, one desired trait and one undesired trait) and choose exactly 2 characteristics to filter by.
  2. Categorize. You see microbes one at a time and decide whether they are useful for the current site, the next one, or whether to discard them.
  3. Build the pool. You start with 6 microbes. Over 4 rounds you are offered 3 options and pick 1, until you have 10. No option is perfect, and that is on purpose.
  4. Treatment. From the 10, you pick the final 3: averages within every range, at least one with the desired trait and none with the undesired one.

Stages 3 and 4 weigh the most in the result. It’s worth getting through the first two quickly and saving time for the others.

Strategy, step by step

Think like a portfolio manager: you are looking for the best set of 3, not the perfect microbe.

  • Profile: choose the 2 most restrictive characteristics, not the most eye-catching ones. If a range is narrow (for example, density from 6 to 8) or the undesired trait would rule out many microbes, filter by that.
  • Categorize: take out what clearly doesn’t work first, but don’t over-discard. You need variety to be able to balance the averages later.
  • Pool: in every round ask yourself what you’re missing. If your density is already at the high end of the range, pick the microbe that brings it down, not the one that pushes it up further.
  • Track the averages as you go: one trick is to turn the average range into a sum range (multiply by 3). That way you compare the sum of the 3 values without dividing. For example, an average of 3 to 5 is the same as a sum of 9 to 15.
  • Treatment: if you kept an eye on the averages, this step takes a couple of minutes. Before submitting, check the three conditions: averages in range, desired trait present and undesired trait absent.
  • Stop optimizing in time: a reasonable treatment that meets everything is worth more than an “almost perfect” one that takes time away from site 3.

Time management

The most common mistake is spending too long on a single site. An unfinished site scores badly, so split the 30 minutes from the start. A reference from one guide:

Site First few times (min) With practice (min)
Site 1 12 7
Site 2 9 7
Site 3 9 7
Buffer 0 9

Site 1 takes longer because that’s where you learn the mechanics. If you reach 11 or 12 minutes on a site and the averages don’t fit, submit your best option and move on.

Common mistakes

  1. Filtering by the wrong thing: choosing the first 2 characteristics you see instead of the most restrictive ones.
  2. Looking for the perfect microbe: one that fixes one attribute but breaks another is worse than an “average” one that balances everything.
  3. Not tracking the averages: if you don’t watch them in the pool, in the treatment you end up doing all the arithmetic with the clock breathing down your neck.
  4. Misreading the desired trait: it is collective. One of the 3 having it is enough; don’t discard microbes just because they don’t have it.
  5. Optimizing a single thing: hitting one range perfectly while the undesired trait slips in.
  6. Ignoring the clock: splitting the time 10-10-10 and running out of it at site 3.
  7. Skipping the tutorial: it isn’t timed. Use it to read the rules and write down the essentials before the clock starts.

How to practice with the simulator

  1. Play first in learning mode, where you see live what fits and why.
  2. Move on to test mode with the full clock, like the real day.
  3. Aim to get from about 12 minutes per site down to around 7.
  4. After each game, review why your combination worked or failed, not just the score.
  5. Train the mental average of 3 numbers from 1 to 10 until it takes you seconds.

One guide estimates that 10 to 15 practice sites are enough to make the flow automatic.

To see all of this in action, Sea Wolf solved walks through a full game, site by site.

Sources

McKinsey doesn’t publish the rules or the scoring, so the figures come from prep sites (several sell courses) and don’t always agree.