
Minimax works from the leaves upward: figure out what each MIN node will do first, since MIN moves right after MAX, then let MAX pick the best of those outcomes.
List the three MIN nodes with the leaf values under each:
| MAX's move | Leaf values under that MIN node | MIN picks (smallest) |
|---|---|---|
| Strategy 1 | 8, 6, -1 | -1 |
| Strategy 2 | 1, 5, 7 | 1 |
| Strategy 3 | -4, -3, -12 | -12 |
MIN always wants to steer the game toward the lowest possible utility for MAX, so each MIN node's value is just the smallest of its three children. That gives $-1$, $1$, and $-12$ for MAX's three options.
Now step up one more level. MAX gets to choose which of these three backed-up values to walk into, and MAX always wants the largest one. Comparing $-1$, $1$, and $-12$, the largest is $1$, coming from strategy 2.
So playing optimally, MAX should choose strategy 2 at the root, since every other strategy leads to a worse (smaller) guaranteed outcome once MIN responds optimally.
\[ \boxed{\text{Strategy } 2} \]