Consider a vorpal sword in a multiuser game. Does it exist? I think not. The computer simulates a sword, but there is no simulated sword.
Suppose you disagree and think there were such an object. The object then seems, prima facie, to be both a non-defective sword and an immaterial object unable to cut any chunk of matter. This seems to be a contradiction. To get out of the contradiction, you probably need to say that the vorpal sword is not a sword—it's a virtual sword. And it can virtually cut some virtual chunks of matter. For many positive material properties that a non-virtual sword would have to have, the virtual sword has a virtual counterpart. But not for all of them. While a real sword would have to either be made by a human smith or not made by a human smith, if the game designers failed to specify the sword's history, then it will neither be true that the sword was virtually(made by a human smith) nor that the sword was virtually(not made by a human smith). There is no violation of excluded middle here.
But there may, nonetheless, be some properties that the virtual sword shares with real swords. The virtual sword may be owned by Jane, who also owns a real sword. The virtual sword may be beautiful, just as a real sword. And, of course, on the view in question, just as the real sword has existence, so does the virtual one.
But now consider this puzzle about identity. Suppose two people playing the game have vorpal swords with exactly the same identities. They put them down in a box. They close the box. They shake the box. Then they each take a vorpal sword out of the box. Do they get their own swords back, or have they swapped them? Suppose the game designers failed to specify the physics of what happens in the box—that's, after all, always possible in the case of virtual reality (it may not be specified just how long the intestines of the dragon are, though it may be specified that it has intestines). There is no entity without identity. If the virtual swords were real, there would be a fact about whether they went to different owners or not. I suppose we could say that after the swap they each come to have the same future properties as the other: virtual sword A has the property of virtually being owned by Jim or by Susan but not both, and so does virtual sword B, and neither has the property of virtually being owned by Jim or of virtually being owned by Susan. Each sword has the property of virtually having different present properties from the other sword, but neither sword actually has any different present properties from the other, after the swap.
Now, you might think, and you would I think be right, that all this is absurd. You might think that while there virtually are vorpal swords there are no virtual vorpal swords. But why not? One reason is that there are puzzles about identity. But there equally are puzzles about the identity of real swords. Another reason, and perhaps a more compelling one, is that whether Jim has a vorpal sword might just be a matter of the value of a single bit in the attributes of Jim (probably more in a sophisticated game) along with context. But likewise a real sword seems to be just a matter of the positions of particles, or maybe just of values of a field. Social constitution goes into virtual reality. But likewise into artifactual reality. And so on.
Conversely, various reasons for believing in real swords apply to virtual ones. The real swords enter into explanations of phenomena. So do the virtual ones. ("Why did George lose the fight? Because his opponent wielded a vorpal sword.") The real swords are thought of and felt about as if they were real. So do the virtual ones after people have been playing the game long enough. The real ones are apparently perceived. So are the virtual ones.
What does this mean? I think this will push in one of two directions. One way is if one holds on to the intuition that something whose existence is "a matter of the properties of other things" does not count as a thing. Thus, the vorpal sword is constituted by bits in the owner's attributes along with context is not a thing. But neither is a real sword if it turns out that its existence is a matter of the properties of particles or fields. Then one will conclude that artifacts, as such, do not exist. Given the undeniability of our own existence, we will have to conclude that we are different in some important sense from swords and bicycles—our relation to the underlying physical world is different.
Alternately, one might conclude that virtual items exist. But where will one draw the line, then? Once one allows virtual items to exist, one will almost surely have to allow fictional items to exist. Sherlock Holmes will exist. Will his gall bladder? Where will one stop? This way lies mad ontological profligacy.