As I understand it, the question is about the following game $G(\kappa)$ of length $\omega$ played on a set $S$ of cardinality $\kappa$. At the $n^\text{th}$ turn, first White chooses a function $\varphi_n:S\to\mathbb N$, and then Black (a.k.a. Dick) chooses a number $k_n\in\mathbb N$; Black wins if $\forall x\in S\ \exists n\in\mathbb N\ \varphi_n(x)\le k_n$. (More colorfully: at each turn, White cuts what's left of $S$ into countably many pieces, and Black eats any finite number of those pieces; Black wins if he ends up eating the whole thing.) For small values of $\kappa=|S|$ Black has a winning strategy, for large values White has a winning strategy, and for in-between values (if any) the game is undetermined. We want to find the cutoff points.
In the notation of Cichoń's diagram, $\mathfrak d$ is the minimum cardinality of a family of functions $F\subseteq\mathbb N^\mathbb N$ such that $\forall g\in\mathbb N^\mathbb N\ \exists f\in F\ \forall n\in\mathbb N\ g(n)\lt f(n)$, and $\operatorname{non}(\mathcal B)$ is the minimum cardinality of a non-meager (second category) set of real numbers. The "small uncountable cardinals" $\mathfrak d$ and $\operatorname{non}(\mathcal B)$ are "incomparable" in the sense that either one may be smaller than the other, or they may be equal.
As already shown in an answer by Alex Kruckman, if $\kappa\ge\mathfrak d$ then White has a winning strategy in $G(\kappa)$. I claim that the converse also holds, i.e., White has a winning strategy in $G(\kappa)$ if and only if $\kappa\ge\mathfrak d$. If Black has a winning strategy, it follows from Kruckman's observation that $\kappa\lt\mathfrak d$; I claim that we also have $\kappa\lt\operatorname{non}(\mathcal B)$, i.e., if Black has a winning strategy in $G(\kappa)$ then $\kappa\lt\min\{\mathfrak d,\operatorname{non}(\mathcal B)\}$. Of course Black has a winning strategy if $\kappa\le\aleph_0$, as was already noted by the OP. These observations (if correct) leave open what happens if $\aleph_1\le\kappa\lt\min\{\mathfrak d,\operatorname{non}(\mathcal B)\}$. It's easy to see that the least cardinal $\kappa$ for which Black has no winning strategy must have uncountable cofinality.
Theorem 1. White has a winning strategy in $G(\kappa)$ if and only if $\kappa\ge\mathfrak d$.
Proof. If $|S|=\kappa\ge\mathfrak d$ then we can choose functions $f_x\in\mathbb N^\mathbb N$ such that $\forall g\in\mathbb N^\mathbb N\ \exists x\in S\ \forall n\in\mathbb N\ g(n)\lt f_x(n)$. Now a winning strategy for White is to play $\varphi_n(x)=f_x(n)$ regardless of what Black does.
Conversely, suppose White has a winning strategy $\sigma$ in $G(\kappa)$; that is, on the $n^\text{th}$ turn, at each $x\in X$ White plays $\varphi_n(x)=\sigma(x;k_1,\dots,k_{n-1})$ where $k_1,\dots,k_{n-1}$ are the previous moves by Black. Now for each $x\in X$ define $f_x(n)$ recursively so that
$$f_x(n)=\max\{\sigma(x;k_1,\dots,k_{n-1}):k_i\lt f_x(i)\text{ for }i=1,\dots,n-1\}.$$
Thus $f_x(n)$ is at least as big as any number that the strategy $\sigma$ will ever assign to $x$ at the $n^\text{th}$ turn, provided $x$ has not already been "eaten". I claim that $\{f_x:x\in S\}$ is a "dominating family", whence $|S|\ge\mathfrak d$. Cosider any function $g:\mathbb N\to\mathbb N$. Since $\sigma$ is a winning strategy for White, there is a point $x\in S$ which "survives" a play $\varphi_1,k_1,\varphi_2,k_2,\dots$ in which White follows the strategy $\sigma$ and Black plays $k_n=g(n)$. Then for all $n\in\mathbb N$ we have (by induction on $n$) $f_x(n)\ge\sigma(x;k_1,\dots,k_{n-1})=\varphi_n(x)\gt k_n=g(n)$.
Theorem 2. If Black has a winning strategy in $G(\kappa)$ then $\kappa\lt\min\{\mathfrak d,\operatorname{non}(\mathcal B)\}$.
Proof. Suppose Black has a winning strategy in $G(\kappa)$. Plainly $\kappa\lt\mathfrak d\le\mathfrak c$ by Theorem 1; Black and White can't both have winning strategies. Now assume for a contradiction that $\kappa\ge\operatorname{non}(\mathcal B)$. We may assume that the game is played on a non-meager set $S\subseteq\mathbb R$ of cardinality $\kappa$, and we may assume that $S\cap\mathbb Q=\varnothing$.
Let $\sigma$ be a fixed winning strategy for Black. We consider only plays in which Black follows the strategy $\sigma$ and White's moves are restricted as follows: At the $n^\text{th}$ turn White chooses a number $q_n\in\mathbb Q$ and plays the function $\varphi_n:S\to\mathbb N$ defined by $\varphi_n(x)=\left\lceil\frac1{|x-q_n|}\right\rceil$. With Black following the fixed strategy $\sigma$, the play of the game is determined by White's chosen sequence of rational numbers $q_1,q_2,q_3,\dots$.
If $q_1,\dots,q_n$ is a sequence of rational numbers, let $S_{q_1,\dots,q_n}$ be the set of points $x\in S$ which are "doomed" after White has chosen $q_1,\dots,q_n$, that is, Black (following his strategy $\sigma$) is sure to "eat" $x$ at the next turn, regardless of White's choice of $q_{n+1}$. Since $\sigma$ is a winning strategy, each point $x\in S$ is in some set $S_{q_1,\dots,q_n}$; otherwise White could play so that $x$ is never "eaten". So $S$ is the union of the (countably many) sets $S_{q_1,\dots,q_n}$. Since $S$ is non-meager, there is some sequence $q_1,\dots,q_n$ such that the set $S_{q_1,\dots,q_n}$ is dense in some interval $I$. After White's moves $q_1,\dots,q_n$, every point in $S_{q_1,\dots,q_n}$ is "doomed" to be eaten on the next turn. However, if White chooses $q_{n+1}\in\mathbb Q\cap I$ this is impossible, since $\varphi_{n+1}(x)$ is unbounded on $S_{q_1,\dots,q_n}\cap I$