[philsnow]

> In saying “every point,” we obviously have to turn a continuous problem into a discrete problem, and the choice of what level of resolution to use is non-trivial. Right now, when I set that resolution to only make a calculation for each two-yard by two-yard square, I can get the map done in a couple minutes. When I set the resolution at one foot by one foot, the process usually takes over an hour. TL;DR: I can make maps for shots from general areas on a hole pretty quickly, but as I get closer to mapping “every point” on a hole, it gets exponentially more time consuming.

Are you making an assumption that you need to have the same resolution at every point on the course? Maybe there are broad areas of fairway that have similar scores (so lower res is fine) and specific "sharp" areas (you'd probably need finer resolution on the hole-wards side of bunkers vs the side that's farther away from the hole?).

There's lots of techniques in computer graphics for figuring out when you can downsample, and you're going to have lots of opportunities to tune those techniques to the problem of golf course strategy analysis (trees, as you mention, probably cast "shadows" of uncertainty and those shadows would need better sampling).

(disclaimer: I golf <1x / year so I don't even know all the words you used in the article)

[scoofy]

OP here. You make a very good point. Most of the holes definitely don't need a consistent gradient across every point on the hole. However, some will. The goal is to reveal that underlying nature of the hazards, and this includes contouring. When we get into the really nitty-gritty aspects of contouring's outsized effects on architecturally interesting areas, the main worry I have is that you can't know where to skimp on the resolution until after you've got the result.

I do think, however, that there should be some relationship between resolution and the "stickiness" of the surface (with higher friction variable, e.g. heavy rough). Given the higher friction areas, there will be less movement in rollout. Less movement in rollout means that the net effect of the contouring is less significant, which means that the resolution is less important, and we can probably save time in these areas.

I'll really have to think about this. It's a good idea.