Why a solver is the wrong habit
Two reasons. The practical one: McKinsey's rules for Solve ban applications, websites, AI tools and prepared notes during the assessment. Whatever you practise with, you'll sit the real thing with your head and, per prep-site reports, pen and paper. The other reason: a solver hides the pattern. If a tool tells you the answer, you never learn to see that a 9/9/9 microbe with the wrong trait is worthless, or that a 1 on one attribute is fine next to two 9s. That recognition is the skill, and it takes about an evening to build.
If you already know the rules, skip ahead. If not, the Sea Wolf guide covers the four steps and the scoring. The short version: for each of three attributes, the mean of your three microbes' values must land in the site's target range, and that's what the shortcut is for.
Step 1: turn every range into a sum
Never divide by three under a clock. Multiply the range by three once, when you read the site brief, and work with sums from then on.
| Target range | Sum of three must be |
|---|---|
| 1 to 2 | 3 to 6 |
| 2 to 4 | 6 to 12 |
| 4 to 6 | 12 to 18 |
| 5 to 7 | 15 to 21 |
| 7 to 9 | 21 to 27 |
| 8 to 10 | 24 to 30 |
Write the three target sums on your paper before you look at a single microbe. From here on you're adding three single-digit numbers and comparing to a pair of bounds. That's it.
Step 2: cross out the microbes that can never work
Values run from 1 to 10, so the other two microbes in any trio contribute at most 20 and at least 2 to a sum. That gives you two quick tests for a single microbe:
- Too low to rescue: if the target sum's minimum minus 20 is more than the microbe's value, it's dead on that attribute. Target 8 to 10 (sum 24 to 30): 24 − 20 = 4, so anything below 4 is out. Target 7 to 9 (sum 21 to 27): anything below 1 is out, which means nothing is, so a low range floor gives you room.
- Too high to rescue: if the microbe's value plus 2 is more than the target sum's maximum, it's dead. Target 1 to 2 (sum 3 to 6): 6 − 2 = 4, so anything above 4 is out. Target 2 to 4 (sum 6 to 12): 12 − 2 = 10, so nothing is strictly dead, but an 8 needs partners summing to 4 or less, which is nearly the same thing.
Edge ranges (1 to 2, 8 to 10) are where these tests bite. Run them during the prospect-pool rounds, before a hopeless microbe ever gets into your pool. Then run the trait check: any microbe carrying the site's undesired trait is out regardless of numbers, and you need at least one carrying the desired trait.
Step 3: let two picks tell you the third
Once you've chosen two microbes, each attribute's target sum minus the two values you have is exactly the window the third microbe has to hit. Three subtractions, three windows, and you scan the pool for a microbe inside all three.
Example. Targets: Energy 2 to 4 (sum 6 to 12), Adhesion 5 to 7 (sum 15 to 21), Speed 8 to 10 (sum 24 to 30). Your first two picks, values in the order Energy / Adhesion / Speed: 3 / 5 / 9 and 2 / 6 / 8. Sums so far: 5, 11, 17. The third microbe needs Energy 1 to 7, Adhesion 4 to 10, Speed 7 to 10 (only up to 10 exists). Any microbe in the pool with Speed 7 or more, Adhesion 4 or more and Energy 7 or less finishes the trio. A 1 / 4 / 7 microbe would do it, and it looks terrible on its own.
Counter-example. Same targets, first two picks 4 / 7 / 10 and 4 / 8 / 10. Sums 8, 15, 20. The third needs Energy at most 4 (fine), Adhesion at most 6 (fine), Speed at least 4 (fine). Looks easy. But Adhesion is already at 15, the floor of the window, so the third microbe's Adhesion must be 6 or less, and anything above 6 breaks the sum. Two strong-looking picks have quietly narrowed the third pick to a small corner of the pool. When you notice that, consider swapping one of the two.
Worked mini-examples
A below-range value that's fine
Target 6 to 7 (sum 18 to 21). Trio values 2, 9, 9. Sum 20. In range. The 2 is a full four points below the floor and the trio is perfect. Anyone who rejected the 2 on sight threw away the answer.
An in-range value that isn't fine
Target 8 to 10 (sum 24 to 30). Trio values 8, 8, 7. Sum 23. One short. The 7 is one below the range and each 8 is exactly at the floor, so there's no slack to absorb it. Swap the 7 for a 9 (sum 25) or one of the 8s for a 10 (sum 25) and it clears.
Two low values on the same attribute
Target 2 to 4 (sum 6 to 12). Trio values 3, 1, 1. Sum 5. Below the floor. One value of 1 is a tool; two on the same attribute is a hole, because they need the third to contribute 4 or more on an attribute you're trying to keep low. This is the most common way a trio of "good" microbes fails.
A five-minute drill
Write three ranges from the table above on a card. Convert to sums. Now roll or pick nine random numbers from 1 to 10 as three microbes and check the trio in your head: sum, compare, next attribute. Then swap one microbe for a new random one and re-check. Ten trios take about five minutes. After two or three sessions the conversion and the sums stop costing attention, which is the whole point: in the real game you want your attention on traits and on the next site, not on dividing by three.
FAQ
What is the Sea Wolf averaging shortcut?
Multiply each target range by three and compare the sum of your three microbes' values to it, instead of dividing to get a mean. A range of 5 to 7 becomes a sum of 15 to 21. Same test, no division, and it makes the impossible microbes obvious.
Do I need a Sea Wolf solver or Excel calculator?
No. The arithmetic is three sums of three small numbers per site, and McKinsey's rules ban outside tools during the assessment anyway. A solver in practice teaches you to depend on something you won't have on the day. Learn the sum method instead; it takes an evening.
How do I know a microbe can never work for a site?
Take the target sum's minimum and subtract the most the other two microbes could contribute (20, since values run 1 to 10). For a range of 8 to 10 the sum must be at least 24, so any microbe below 4 on that attribute is out. Do the mirror check for low ranges.
Can a value outside the range still be a good pick?
Yes, often. The site scores the mean of three, so a 2 next to a 9 and a 9 gives a mean of 6.67. What matters is whether the sum lands in the target sum, not whether each value lands in the range.
Keep going
The full game, the four steps, the scoring and a complete site worked with numbers are in the Sea Wolf guide. The other games: Red Rock Study and the Sustainable Futures Lab. Or zoom out to the McKinsey Solve guide.
SolveForge is an independent practice tool and is not affiliated with, endorsed by, or connected to McKinsey & Company. "McKinsey" and "Solve" are trademarks of their respective owners. Practising here does not guarantee any assessment result.