7

Given a topological space $X$, Let $\mathcal O$, denote the set of all open covers of $X$.

We say that a space $X$ satisfies $S_1(\mathcal O,\mathcal O)$, if for every sequence of open covers $\{ \mathcal U_n : n \in \mathbb N \}$, there exists a sequence of open sets $\{ U_n : U_n \in \mathcal U_n \text{ for each } n \in \mathbb N \}$, which is an open cover of $X$.

Let $G_1(\mathcal O,\mathcal O)$ be the game with the following rules:

  1. There are two players: ONE and TWO.

  2. At each odd step $2n-1$, where $n \in \mathbb N$, ONE picks an open cover $\mathcal U_{2n-1}$ of $X$.

  3. At each even step $2n$, where $n \in \mathbb N$, TWO responds by picking an open set $U_{2n} \in \mathcal U_{2n-1}$

  4. If the sequence $\{ U_{2n} : n \in \mathbb N \}$ is an open cover of $X$ then TWO wins.

  5. If the sequence $\{ U_{2n} : n \in \mathbb N \}$ is not an open cover of $X$ then ONE wins.

  6. A strategy for ONE, is a function $F$, that, given any initial finite sequence of moves, $\{ \mathcal U_1, U_2, \mathcal U_3, U_4,...,U_{2n-2} \}$ of the game, produces the next move $\mathcal U_{2n-1}$.

  7. A winning strategy for ONE, is a strategy $F$, such that $\bigcup_{n=1}^{\infty} U_{2n}$ is never a cover of $X$.

  8. A strategy for TWO, is a function $F$, that, given any initial finite sequence of moves, $\{ \mathcal U_1, U_2, \mathcal U_3, U_4,...,\mathcal U_{2n-1} \}$ of the game, produces the next move $U_{2n}$

  9. A winning strategy for TWO, is a strategy $F$, such that $\bigcup_{n=1}^{\infty} U_{2n}$ is allways a cover of $X$.

We say that a topological space $X$ satisfies $G_1(\mathcal O,\mathcal O)$ if TWO has a winning strategy for $G_1(\mathcal O,\mathcal O)$ in $X$.

I am trying to prove the claim below and have mannaged to prove only one direction. Any help?

This claim is some lighter version of theorem 3 in this article.

Claim: $X$ satisfies $S_1(\mathcal O,\mathcal O)$ if and only if TWO has a winning strategy for $G_1(\mathcal O,\mathcal O)$ in $X$.

Proof: First direction: Suppose that TWO has a winning strategy for $G_1(\mathcal O,\mathcal O)$ in $X$. Then, Given a sequence $\mathcal U_{2n-1} = \{ \mathcal U_1, \mathcal U_3, \mathcal U_5, ... \}$ of open covers of $X$, Start with $\{ \mathcal U_1 \}$, and let $U_2 = F(U_1)$, where $F$ is the winning strategy for TWO. Then, take $U_4 = F(\mathcal U_1,U_2,\mathcal U_3)$, at the $n$'th step, $U_{2n} = F(\mathcal U_1,U_2, \mathcal U_3,U_4,...,\mathcal U_{2n-1})$. It is clear that $\bigcup U_{2n}$ is an open cover of $X$ since $F$ is a winning strategy. Second direction: Suppose that $X$ satisfies $S_1(\mathcal O,\mathcal O)$. Then, for every sequence of open covers $\mathcal U_n$ of $X$, there is a sequence $\{ U_n : U_n \in \mathcal U_n, n \in \mathbb N \}$, such that $\bigcup U_n$ is an open cover of $X$. We have to show that TWO has a winning strategy in the game $G_1(\mathcal O,\mathcal O)$. So, how can I build this strategy?

Thank you!

Rushabh Mehta
  • 13,845
topsi
  • 4,292
  • 1
    why do you think your claim is true ? – mercio Jul 03 '14 at 11:11
  • I am not sure.. as a matter of fact, I think that this is a part of my question.. But, I am not sure how to find a counterexample either.. – topsi Jul 03 '14 at 11:40
  • What is $\mathcal O$? You define things like $G_1(\mathcal O,\mathcal O)$ and $S_1(\mathcal O,\mathcal O)$; but they do not depend on $\mathcal O$ in any way. – Martin Sleziak Jul 03 '14 at 12:02
  • BTW the source of your post contains this link: http://www.ams.org/journals/proc/1999-127-02/S0002-9939-99-04513-X/S0002-9939-99-04513-X.pdf But it is not mentioned in the post anywhere. (I think that it should be mentioned. For example, it is shown in that paper that $S_1$ holds if and only ONE does not have a winning strategy.) – Martin Sleziak Jul 03 '14 at 12:06
  • Sorry, I will fix it now in my original question – topsi Jul 03 '14 at 12:09
  • 1
    I am not sure whether this helps, but in the paper M Scheepers: A Direct Proof of a Theorem of Telgársky; DOI: 10.1090/S0002-9939-1995-1273523-1, jstor it is shown that similar claim is not true for Menger's game. If I am not mistaken, the only difference between the two games is that here TWO selects a finite subset of $\mathcal U_n$ instead of just one set from this open cover. – Martin Sleziak Jul 03 '14 at 12:19
  • Since I did not want to post too many comments here, I've mentioned some related things I was able to find in chat. – Martin Sleziak Jul 03 '14 at 12:47
  • I will go over it. Thank you!! – topsi Jul 03 '14 at 13:37
  • @ShirSivroni I just discovered your interesting question. So what is the answer in the end? Did you manage to prove your claim, or is it not true after all? Thanks for your answer... – Etienne Jul 26 '14 at 13:53
  • 2
    @ShirSivroni After a very superficial look at Schheepers' paper, it seems that the best you can get is that $X$ satisfies $\mathcal S_1(\mathcal O,\mathcal O)$ if and only if player I does not have a winning strategy in the game. – Etienne Jul 26 '14 at 14:04

1 Answers1

2

First, let's note that the selection principle $S_1(\mathcal A,\mathcal B)$ is equivalent to ONE lacking a winning "predetermined" strategy in the selection game $G_1(\mathcal A,\mathcal B)$ (which uses the round number but ignores the moves of TWO). https://arxiv.org/abs/1809.10783

Your claim that $S_1(\mathcal O,\mathcal O)$ is equivalent to TWO having a winning strategy in $G_1(\mathcal O,\mathcal O)$ is false: a Lusin set is an example of a space where neither player has a winning strategy for $G_1(\mathcal O,\mathcal O)$. Therefore $S_1(\mathcal O,\mathcal O)$ holds while TWO lacks a winning strategy for $G_1(\mathcal O,\mathcal O)$. http://www.ams.org/journals/proc/1995-123-11/S0002-9939-1995-1273523-1/home.html