If I found myself playing golf against Tiger Woods, I'd try higher risk / higher reward shots, because if I stick to my typical game I'm surely going to lose. This would make my score worse on average, but who cares about that when once in a while the bet pays off and I win?
In golf, there are prizes for getting 2nd, 3rd, etc... So if you knew Woods was going to destroy you, you would still play your best game and try to get 2nd place.
Now, for the people who might get 2nd, 3rd, and maybe a few more top spots, they might want to try to best Woods. But people who have no chance would still play their regular game (if your theory is correct).
So I'd be interested to see if the drop in score correlated at all with the average score across previous games.
The article is a particularly idiotic example of uninformed speculation passing for science journalism.
Based solely on the above they draw their conclusions. I'm just pointing out that there is a simpler and more rational conclusion one could draw.