Riggs vs Pritchard on Luck

At the Stirling Epistemic Value conference Wayne Riggs presented a nice alternative to Duncan Pritchard's modal account of luck. One of the core cases Wayne wanted his account to handle was the following (I do not remember the details but do by all means go read Wayne's very interesting paper).

Indiana Jones and Missouri Smith are captured by natives who are going to kill them. Suddenly there is a total solar eclipse and the natives believe the gods want them to set Jones and Smith free. Smith says "we were really lucky that that eclipse occurred", to which Jones replies "lucky? Are you out of your mind? I knew well in advance that it would occur and would never have allowed them to capture us if it didn't". It seems that the occurrence of the eclipse was lucky for Smith but not for Jones.

But how can that be? Well, according to Wayne, even if one has no control over a given event, an event can be lucky if one exploited one's knowledge of the occurrence of the event in particular ways. On Wayne's account, an event is lucky for x iff x has no control over its occurrence, and x did not exploit its occurrence for particular purposes (Wayne's analysis contains a third condition which I will leave out here).

I like the account. However, one issue which came up at the conference is that if one takes the control and exploitation conditions to be necesary and sufficient for luck, then control but lack of exploitation would be sufficient for non-luck. But that seems wrong. Suppose, for instance, that John would win the race if only Mary didn't participate. It is within John's control to break Mary's leg, but he doesn't want to do that. So he fails to exploit his control. However suppose, by sheer coincidence, that Mary breaks her leg and that John wins. In that event we would want to say that John's victory is lucky and that Mary's breaking her leg is unlucky. So maybe we should simply get rid of the control condition and make do with the exploitation condition.