Rethinking SM

[Anthony]

Part of the problem is that realistically what you use for SM is going to depend on the use case.

In cases where cross-sectional area is critical (perception, random saturation of an area such as an attack with bombardment) averaging (SM for length) and (SM for width) and adding 2 is mostly correct, and generally has the same effects as Tyneras' chart.

For attacks, the realistic implementation is to roll for every dimension in which you can miss -- that's usually height and width for thrusting and missile attacks, height and depth for swing attacks. This is quite inconvenient so I entirely understand why GURPS doesn't do it, but it turns out that what matters is mostly the smaller dimension -- if you're 10- to get the width right and 14- to get the height, you're 41.8% to get both, which is about 9-.

[Rupert]

Originally Posted by Anthony

For attacks, the realistic implementation is to roll for every dimension in which you can miss -- that's usually height and width for thrusting and missile attacks, height and depth for swing attacks. This is quite inconvenient so I entirely understand why GURPS doesn't do it, but it turns out that what matters is mostly the smaller dimension -- if you're 10- to get the width right and 14- to get the height, you're 41.8% to get both, which is about 9-.

Just to make matters more annoying, when it comes to ranged attacks it's harder for humans to get things right vertically than it is for horizontal accuracy. Partly this is because generally we're using ballistic attacks in a gravity well, so range estimation matters, but also most non-telescopic sights make sighting in the horizontal plane easier than in the vertical, and also for hand-held weapons it's just that bit easier to be precise in terms of left/right than in height. It's easier to shoot a post that's standing vertically than one lying on the ground with a gun, even when range estimation isn't an issue.

[Anthony]

Originally Posted by Rupert

Just to make matters more annoying, when it comes to ranged attacks it's harder for humans to get things right vertically than it is for horizontal accuracy.

Only against unmoving targets, accuracy drops off rather rapidly once you need to track and lead the target, and horizontal movement is far more likely than vertical.