There are often a lot of discussion about "breakpoints" in skills, and this always puzzles me. There were past attemtps to "smooth" the weapon skills by evading the current (and past) min delay breakpoints, and I did not understand the purpose.
1.) There are usually an optimum where skills give diminishing results per XP - the question is often not how sharp this breakpoint is, but how well the results the skill gives are communicated. For example, armour skill decreases the spellcasting penalties of your armour, but nobody knows how much, so there are questions when it is worth to train armour for this purpose.
[ The answer I think is that it is rarely worth to train armour skill for this purpose unless your armour skill is much lower than your spell skills for some strange reason and you are in a heavy armour, as training spell skills decrease your failure rate much faster. Usually you train armour skill to gain AC, not to reduce spellcasting penalties, as the former matters more (in terms of the skill level). ]
I do not see how it matters that much whether at some point you gain absolutely no more reduction of your spellcasting penalties, or there are just diminishing results. The difference is whether you know about these and know where you want to stop training. Hard breakpoints are easier to communicate, and allow you to set up skill training goals in advnace. I think it would be nice at least if you could directly see in % how much of your armour skill reduces your spellcasting penalties. That would get rid of such questions.
2.)
Do you think it would make sense to change this rule, so that you would get rid of any penality after a certain skill / strength level?
Note that the first and second part of this question does not ask for exactly the same change. You could make armour skill stopped contributing to reduction of spellcasting penalties at a certain level AND not reducing the penalties to 0. For example, the maximum reduction could always be 50% reached at the skill level = armour ER.
3.) I argue against removing breakpoints, because breakpoints are usually connected to equipment and spell selection deceisions. The main difference between weapons is the damage they deal, but the second most important difference is the progress of this damage by skill. Lajatangs deal less damage than great maces - but they reach most of their potential at a much lower skill level, therefore being much cheaper. This is directly connected to the progress of the damage by the skill, and the optimum of this progression: the "breakpoint". The fact that this breakpoint is fixed unlike the similar breakpoint caused by the casting success for spells does not matter that much IMHO, athough knowing the breakpoint *in advance* would be very nice.
Note that I understand that certain items has their "breakpoint" at level 27 - there are no diminishing returns for them. This is true for some weapons, armours and spells as well. Otherwise the complete skill would be useless over a certain level! But this does not mean that there should be no breakpoints for other items/spells, as exactly this differentiate them.