In Go, at least, I'd think it would be much easier to deliberately orient yourself with conscious thought since you are effectively doing a very concrete form of calculation. Sports match prediction, on the other hand, is a lot more vague and nebulous than, say, a life and death problem. It probably is easy to get misguided by irrelevant factors, or give relevant factors too much or too little weight.

Well, the same should go for chess, and expertise seems to be a factor. Isn't there chess research that shows that, given a problem position, chess masters immediately looked at the right answer while average players did not? OC, that is not the result of conscious processing, but neither is it the result of extended unconscious processing. So it is not conclusive, but it does indicate the importance of unconscious processing.

I wasn't trying to suggest that unconscious processing wasn't significant in Go. I was only pointing out that unlike in match prediction, I doubt that conscious processing is going to make things worse most of the time.