Sea Wolf is the ocean-site optimization game most candidates see as the second module of McKinsey Solve. Reported length is about 30 minutes. You match marine species to sites using temperature, depth, nutrients, and — in later rounds — how species interact. The scoring pattern candidates describe is harsh on rule breaks and late-round instability, not on whether you found a pretty coral palette.
It is not Redrock (data). It is not the old food-chain Ecosystem Building game, though that can still appear in a longer three-game sitting — ecosystem building guide. Confirm your invite.
Last full pass: 24 August 2026. CoachNed is independent practice material, not affiliated with McKinsey. Independent browser practice: Sea Wolf-style sim. Not a leaked McKinsey file.
The only strategy that survives late rounds
- Read every species card and every site card before placing. Process score cares that you opened the panels.
- Eliminate anything that violates a hard constraint. Temperature out of range, depth illegal, nutrient need unmet: dead. Do not "almost" place it.
- Then optimize among the legal set (complementary pairs, avoid two high-nutrient competitors on the same rock).
- Work site by site in messy rounds so you do not orphan a species with no legal home.
Early rounds are clean matches. Late rounds add competition and symbiosis. Bank time early; spend it when interactions appear.
| Round stage | What is easy | What starts to kill the score |
|---|---|---|
| Early | One-to-one trait matches | Skipping a card and guessing |
| Middle | Overlapping ranges | Single-trait matching ("cold water = this site") |
| Late | Interactions, scarce nutrients | Two competitors on one site; unfinished placements |
Pacing note from the Solve guide: difficulty climbs, so do not perfectionist-paint round 1.
Worked example: four species, three sites
Sites
| Site | Temp | Depth | Nutrients |
|---|---|---|---|
| Inlet | Warm | Shallow | High |
| Shelf | Cool | Mid | Medium |
| Trench | Cold | Deep | Low |
Species (simplified cards)
| Species | Temp | Depth | Nutrients | Interaction |
|---|---|---|---|---|
| Kelp bass | Warm | Shallow or mid | High | Competes with anything else that needs high nutrients |
| Glass squid | Cold | Deep | Low | None |
| Filter sponge | Cool or cold | Mid or deep | Low or medium | Provides cover; cannot sit with bass |
| Grazing urchin | Warm or cool | Shallow | High or medium | Competes with bass for high nutrients |
Legal filter first (do this on paper in the real game with the UI).
- Bass: only Inlet (warm + shallow/mid + high). Shelf is cool. Trench is cold/deep/low. Bass cannot go anywhere else.
- Squid: only Trench.
- Sponge: Shelf (cool, mid, medium) or Trench (cold, deep, low). Not Inlet (warm, high, and "cannot sit with bass" anyway).
- Urchin: Inlet (warm, shallow, high/medium) or maybe Shelf if we stretch "cool + shallow" — Shelf depth is mid, urchin requires shallow → Inlet only.
Conflict. Bass and urchin both legal only on Inlet, and both want high nutrients → they compete. You cannot legally stack them if the card says high-nutrient competition hurts the score.
Placement. Inlet: bass only (or urchin only — pick the one the round's objective weights; if the prompt is "maximize biomass under stability," bass + empty urchin is better than a penalty). Shelf: sponge. Trench: squid. Urchin has no remaining legal site that is shallow without Inlet. Leaving urchin unplaced may score better than a constraint violation. That is the Sea Wolf punchline: an incomplete legal set beats an illegal pretty set.
If a later round adds a warm shallow medium-nutrient lagoon, the urchin gets a home and the competition disappears. Until then, do not force it onto the Shelf.
What this trains that cases also use
Filter, then optimize is the same move as "kill the infeasible branches before you model synergy." It is not a substitute for interviewer-led cases. After Solve, you still need PEI: PEI guide.
Redrock is the opposite tempo: slow read, honest bases. Practise both if you have both.
Failure modes that are Sea Wolf-specific
- Single-trait matching. Cold water is not enough if depth is wrong.
- Placing competitors "for coverage." The card told you they fight.
- Never opening the interaction panel. Process plus product both suffer.
- External tools, screenshots, second monitor. Integrity and process flags. Solve guide.
- Studying only Ecosystem Building blogs from 2021. Different game. See ecosystem building only if your invite or the three-game variant includes it.
- Buying "real Sea Wolf maps." Do not. Independent sims and constraint puzzles only.
A short prep block
Day 1. This page + invite timing.
Day 2. Logic-grid / constraint puzzles: "who cannot sit where." No biology textbook.
Day 3. Timed Sea Wolf-style rounds. After each, write one illegal placement you almost made.
Day 4. If Redrock is also on the invite, one data sitting so the switch does not surprise you. Then stop grinding games and sleep.
McKinsey does not publish a Sea Wolf cutoff. You will not see your score. Treat legal placements and calm panel reading as the controllable inputs.