In WW2, a 10% hit rate was very high. Sometimes, such accuracy was achieved. For example, in November 1942, off Guadalcanal, Washington fired 117 16" shells at Kirishima, sinking her with 16 (or 20, according to the Japanese, or 9), for a hit rate of 13.7%. This was at the very short range (for battleships to fight each other) of 5-8km, and at night.

Lower accuracy was usual:

• Tirpitz, in a gunnery test against a target ship, maneuvering under radio control (but not shooting back), at the long range of 25km, hit 9 times, with 120 shells fired (7.5%).

• Bismarck vs Hood and Prince of Wales: 7 hits from 93 shells (7.5%); Prince of Wales shooting at Bismarck: 3 hits from 59 shells (5.1%)

• Renown vs Gneisenau and Scharnhorst, April 1940: 3 hits from 230 shells (1.3%); Gneisenau shooting at Renown: 2 hits from 54 shells (3.7%)

• Duke of York vs Scharnhorst, December 1943: 13 hits from 446 shells (2.9%). This was impressively accurate shooting under the conditions (storm, night).

• Vittorio Veneto vs British cruisers, March 1941 (Cape Matapan): 0 hits from 94 shells (0%).

(For comparison, hit rates at Jutland in WW1 were approximately 3%.)

The problem isn't the accuracy of the gun itself, but that one is shooting from a rolling, pitching, maneuvering platform at a maneuvering target. At longer ranges, the target has over a minute to get out of the way before the shells arrive - at 30kts, the target will travel almost 1km in 1 minute. One must shoot at the predicted position of the target, not at where the target is when the shells are fired. Good shooting will straddle the target (put shells on both sides), which might or might not translate into hits, depending on luck.