We know which players won, now how did the holes perform? With four rounds of data, we can compute how many different scores each hole gave to each player, and compare that to how many different scores the hole gave to the field. What we want is a hole that gives a lot of different scores to the field, while giving any particular player the same score from round to round.
Now, because each player's "correct score" usually falls in-between integers, the best a hole can do is give out an average of 1.62 different scores to each player. At the other extreme, if a hole were to give out scores by rolling a four-sided die it would give out about 2.54 different scores to each player – so 2.54 basically indicates total randomness.
In general, the wider the scoring spread across the field, the more difficult it is for a hole to give consistent scores to each player from round to round. So, the holes that are up and to the right of the other holes did better. For example, hole #1 did better than #4 because #1 was more consistent to each player while giving out just as many different scores across the field.
Hole #18 seems to have given out scores to each player in a more-than-random way. But, that's because it's really two holes (with different pars) so we really can't read anything into this result. Holes #5, #13, and #17 also used two different pin positions, so ignore them, too.
We can say holes #1, #9, #10, and #12 gave the best performances. Holes #4, #8, #3, #14, and #15 underperformed – in this group of elite holes.
