The game can be reformulated in an equivalent one: The player A chooses an element x from the set S (with ∣S∣=N) and the player B asks the sequence of questions. The j-th question consists of B choosing a set Dj⊆S and player A selecting a set Pj∈{Qj,QjC}. The player A has to make sure that for every j≥1 the following relation holds:
x∈Pj∪Pj+1∪⋯∪Pj+k.
The player B wins if after a finite number of steps he can choose a set X with ∣X∣≤n such that x∈X.
a.
It suffices to prove that if N≥2k+1 then the player B can determine a set S′⊆S with ∣S′∣≤N−1 such that x∈S′.
Assume that N≥2k+1. In the first move B selects any set D1⊆S such that ∣D1∣≥2k−1 and ∣D1C∣≥2k−1. After receiving the set P1 from A, B makes the second move. The player B selects a set D2⊆S such that ∣D2∩P1C∣≥2k−2 and ∣D2C∩P1C∣≥2k−2. The player B continues this way: in the move j he/she chooses a set Dj such that ∣Dj∩PjC∣≥2k−j and ∣DjC∩PjC∣≥2k−j.
In this way the player B has obtained the sets P1,P2,…,Pk such that (P1∪⋯∪Pk)C≥1. Then B chooses the set Dk+1 to be a singleton containing any element outside of P1∪⋯∪Pk. There are two cases now:
Case 1. The player A selects Pk+1=Dk+1C. Then B can take S′=S∖Dk+1 and the statement is proved.
Case 2. The player A selects Pk+1=Dk+1. Now the player B repeats the previous procedure on the set S1=S∖Dk+1 to obtain the sequence of sets Pk+2,Pk+3,…,P2k+1. The following inequality holds:
∣S1∖(Pk+2⋯P2k+1)∣≥1,
since ∣S1∣≥2k. However, now we have
∣(Pk+1∪Pk+2∪⋯∪P2k+1)C∣≥1,
and we may take S′=Pk+1∪⋯∪P2k+1.
b.
Let p and q be two real numbers such that 1.99<p<q<2. Let us choose k0 such that
(qp)k0≤2⋅(1−2q)andpk−1.99k>1.
We will prove that for every k≥k0 if ∣S∣∈(1.99k,pk) then there is a strategy for the player A to select sets P1,P2,… (based on sets D1,D2,… provided by B) such that for each j the following relation holds:
Pj∪Pj+1∪⋯∪Pj+k=S.
Assuming that S={1,2,…,N}, the player A will maintain the following sequence of N-tuples: (x)j=0∞=(x10,x20,…,xN0). Initially we set x10=x20=⋯=xN0=1. After the set Pj is selected then we define xj+1 based on xj as follows:
xij+1={1,q⋅xij,if i∈Pjif i∈/Pj.
The player A can keep B from winning if xij≤qk for each pair (i,j). For a sequence x, let us define T(x)=∑i=1Nxi. It suffices for player A to make sure that T(xj)≤qk for each j.
Notice that T(x0)=N≤pk<qk.
We will now prove that given xj such that T(xj)≤qk, and a set Dj+1 the player A can choose Pj+1∈{Dj+1,Dj+1C} such that T(xj+1)≤qk. Let y be the sequence that would be obtained if Pj+1=Dj+1, and let z be the sequence that would be obtained if Pj+1=Dj+1C. Then we have
T(y)=i∈Dj+1C∑qxij+∣Dj+1∣
T(z)=i∈Dj+1∑qxij+∣Dj+1C∣.
Summing up the previous two equalities gives:
T(y)+T(z)=q⋅T(xj)+N≤qk+1+pk, hence
min{T(y),T(z)}≤2q⋅qk+2pk≤qk,
because of our choice of k0.