If you are very confident in your theory, seeing an outcome that matches your hypothesis should only update your confidence a very small amount. If you see an outcome that does not match your hypothesis, you should update your confidence significantly, since you had assigned a very low probability that that outcome would occur.