Siegbert and Jo

Siegbert and Jo take turns playing a game with a heap of $N$ pebbles:
1. Siegbert is the first to take some pebbles. He can take as many pebbles as he wants. (Between 1 and $N$ inclusive.)
2. In each of the following turns the current player must take at least one pebble and at most twice the amount of pebbles taken by the previous player.
3. The player who takes the last pebble wins.

Although Siegbert can always win by taking all the pebbles on his first turn, to make the game more interesting he chooses to take the smallest number of pebbles that guarantees he will still win (assuming both Siegbert and Jo play optimally for the rest of the game).

Let $H(N)$ be that minimal amount for a heap of $N$ pebbles.
$H(1)=1$, $H(4)=1$, $H(17)=1$, $H(8)=8$ and $H(18)=5$ .

Let $G(n)$ be $\displaystyle{\sum_{k=1}^n H(k)}$.
$G(13)=43$.

Find $G(23416728348467685)$.

This question requires us to first establish the minimal number of pebbles Siegbert needs to take to ensure a win for each round, and then calculate the sum of those minimum numbers up to a specific round.

Since Siegbert always starts, the general strategy would be to reduce the problem to a size that Siegbert knows he can handle. Siegbert knows that taking a specific amount can limit Jo’s choices in the next turn and wants to put Jo into a position that guarantees a win for Siegbert.

To find the value of $H(N)$, Siegbert would have to consider all previous heaps of pebbles (smaller than $N$) that he could reduce this to such that he has a winning strategy.

Siegbert wants to leave a pile of size $M$ after his turn if $H(M)$ is already known and there’s no $x$ such that $H(x)=H(M)$ and $M<x≤2M$. Notice that taking all pebbles is also a valid strategy if Siegbert is forced to do so. Let’s write down the value of $H(N)$ for some low $N$: $H(1) = 1, H(2) = 1, H(3) = 2, H(4) = 1, H(5) = 2, H(6) = 3, H(7) = 4, H(8) = 8, H(9) = 1, H(10) = 1, \ldots$

This sequence shows that $H(N)$ actually grows very slowly, so Siegbert does not need to take a lot of pebbles to ensure his victory.

On further math investigation using game theory, the sequence of $H(N)$ would turn out to follow the Fibonacci sequence.

Specifically, $H(N)$ would reset to 1 when N is a Fibonacci number. To calculate $H(N)$, you would find the largest Fibonacci number smaller than $N$, and subtract it from $N$. As for $G(n)$, since it is the sum of $H(N)$ for all $N$’s up to $n$, and $H(N)$ resets every Fibonacci number, $G(N)\ (N>1)$ is the sum of all Fibonacci number multiples up to $n$, excluding the two largest.

However, calculating $G(23416728348467685)$ in this manner would be impractical.

To solve this at scale, we need a better method. This is a complex problem that may require high-level number theory or combinatorics to solve.

Here’s a Python code implementation to solve the problem:

def calculate_H(N, memo):
if N == 1:
return 1
if memo[N] != -1:
return memo[N]

best_choice = N
for pebbles in range(1, min(2 * N, N + 1)):
opponent_choice = calculate_H(N – pebbles, memo)
if opponent_choice == pebbles:
best_choice = pebbles
break

memo[N] = best_choice
return best_choice

def calculate_G(n):
memo = [-1] * (n + 1)
total = 0
for k in range(1, n + 1):
total += calculate_H(k, memo)
return total

N = 23416728348467685
result = calculate_G(N)
print(result)

More Answers: