Moving on to the most popular empirical method - Rifleman's rule.
The idea is very simple:
1. Measure inclination α;
2. Measure distance to target D (along line of sight);
3. Calculate equivalent horizontal distance De = D * cos(α);
4. Look up compensation as if shooting horizontally at distance De.
Rifleman's rule has another nice property besides simplicity: it does not depend on the rifle, caliber, zero distance, etc. - just inclination and distance. This allows implementing it in universal laser rangefinders.
Sadly it is based on the assumption that bullet travels in vacuum (space marines, nota bene), and its use on Earth has significant limitation.
GP11 vertical miss distance, cm
Bottom line: rifleman's rule is better than nothing and better than NDS "method". Yet 600 m at 30º is well in the red, explaining longevity of paper plate mentioned at the beginning of this article.
But why is everybody using rifleman's rule then? The answer is mostly hidden just outside of the table. Rifleman's rule works fine at distances shorter than 300 m, and is acceptable up to 400 m, which covers the majority of hunting applications. Army sniping relies on hunting skills, both historically (as the name implies) and currently (hunting experience is still a plus for an army sniper). But the table shows that what works for hunting in the mountains does not work well for game that shoots back, requiring you to keep the distance.
Rifleman's rule consistently overestimates the effect of inclination for the three calibers that I looked at. After using this method, impacts are below the target. Predictability is good, one can come up with an empirical rule like "for inclinations above x degrees at y meters, aim at shoulder level". Yet for reasons explained later in this article I decided not to bother.
Domain: 30º, 500 m. Maybe a little more if holding a correction.
Complexity: very simple -- 1 table lookup, 1 mathematical operation, often automated.
Pros: beautiful simplicity, independence from rifle and caliber, no need to amend or extend existing ballistic tables.
Cons: surprisingly limited applicability.
But not to despair, there's hope. Read on.