That answers my question. According to the BIOS report the German criteria is "To qualify as a successful penetration the projectile was required to penetrate the plate in a condition fit to burst and not to deform in any manner which would prevent detonation." So a different criteria like 50% of the mass behind the armor or some witness plate behind armor penetrated would produce quite a different relationship between angle and armor resistance graph.
Indeed. However, if You look into Navy and Army APCBC penetration graphs, the performance against homogenious armour is often labelled with:
"gegen Homogenpanzer bleibt das Panzergscho� in der Regel heil, Grenz- und Heilbleibkurven fallen zusamenn"
Which means that the seperately drawn curves for intact projectile penetration and projectile penetration in either broken condition or intact but embedded in plate for very hard armour (incl, face hardened) are usually falling together for homogenious armour, which means that the projectile, even while stuck in the plate -where damage done to the projectile is typically most severe- would still be in effective bursting condition (whole or scratched, some other damage was also accepted: nose breakage and projectile set up as long as it didn�t reach the cavity). Normal 7.5 and 8.8cm projectiles would require 5 out of 5 projectiles to pass the plate in effective bursting condition at stated velocity. But note that in the trials, only blind filled APCBC were used, no life projectiles! Assertion whether or not the projectile received sufficient damage was kind of subjective based upon examination of the recovered projectile body and particularely, whether or not projectile damage reached the explosive cavity or deformed the base fuse plug.
At high oblqiuity different thing may happen. The highest obliquities systematically tested for in ww2 Germany were 60 deg and 75 deg impact. At these obliquities, base damage was a frequent occurance (no cap could protect from it), and nose breakage was normal. Eventually, the acceptance trials were thus carried out to 45 deg, where base damage didn�t regualely set in for proofing and manufacturer acceptance purposes.
But remember, what You call "armour resistence graph" is actually only a graph for the G-grenz penetration at fixed striking velcoity (simulating 100m distance in the example above), comparable more or less to US NBL definition at obliquities larger than 45 deg ( US Navy EEF definition at obliquities lower than 45 deg). Resistence of armour is our interpretation by comparing the respective penetration in very much hardened plates (150kg/mm^2 tensile strength) with those of construction grade steels (50kg/mm^2 tensile strength).