First, the point Jasmin was trying to make which you failed to see. I, as a player can easily create a warrior type that can basically shut down a spellcaster. You cannot, as a mage type or any other type, basically shut down a warrior. Sure you sacrifice some damage and hit points for save versus spell/paralysis/mind, but its very feasible.

Now take the above, and put in a 5 to 10% failure rate. No thanks. Most power from mages and communers comes from their prayers/spells. If they are not casting, they are pretty much nothing. This is true for all the classes except, arguably, conjurers or a well zombied out necromancer and shapeshifters (whose skills by the way do fall under warrior chance of failure). A warrior who fails a stun is still going to be dishing out seriously good damage with his normal melee attacks.

For example, invoker a comes upon warrior a. Invoker a tries to cast a spell and fails it, lets say cone of cold, so therefore he is doing nothing but absorbing damage straight out for at least 2 rounds. If warrior a is smart, he can stretch that to 4 rounds with any number of lagging techs. That change drastically changes the game balance, although you see it as just a tiny cosmetic change, its not.

Your argument about how people always make mistakes in real life holds no water what so ever. The mistake aspect of this game still exists, in the form of a player making decisions. And even if that can be argued against, there is still this argument, CF is not real life.

Next, there is Nepenthe's answer. Which you also didn't address in your reply. His basic answer to your post is this. Spells and skills at 100% work 100% of the time except for certain factors which take away from that 100%. For spells its spell save, for skills, its environmental factors, trying to stun an enlarged individual, save versus paralysis, size comparisons, speed comparisons, dexterity comparisons and checks, strength checks and comparisons, race oddities, intelligence checks, etc. You may not SEE it but its there. When you see, "You fail to stun so in so" you think to yourself, "####ing stupid, its at 100% it should work" but what you should see, which you don't, is this:

Vershelt skill at stun 100% = base chance to work 100%
Vershelt is size 2, Target is size 2 = -0% chance to work.
Vershelt has strength 24, Target has strength 23 = +5% chance to work.
Vershelt has dexterity of 16, Targer has dexterity of 21 = -20% chance to work.
Target has save versus paralysis of -16 = -8% chance to work.
Final net chance to work: 77% chance of success
RNG output: 84
Result: FAILURE

And yet again, as this ties in with the other points, you are still outputting damage.

For spells its more like:
Necromancer casts weaken on Vershelt who doesn't have the head.
Necromancer % with weaken = 100%
Necromancer % with spellcraft = 84%
RNG output on spellcraft: 58
Level of necromancer weaken spell: 51+5
Level of weaken spell cast on Vershelt: 56
Level of Vershelt: 51
applied negative spell save to Vershelt for difference in level: +10
Vershelt save versus spell: -67
Base chance of weaken working: 100%
As adjusted for save versus spell: 43%
RNG output: 56
Result: FAILURE

At this point the mage has done no affect what so ever on you. While you are still beating on him despite failing your stun.

The chance of failure is in there. And if you decided to add on an additional 5 to 10% failure for spells, it leaves mages dramatically weakened when you apply the statistics. You would have to raise the percentage chance of non damage offensive spells working by 25 to 30%. Thus throwing the whole save versus spell situation into termoil like back in the beginning, when -20 save versus spell made you immune, to when -90 save versus spell did nothing.

All your posts recently have been directly from the "battlerager berserker warrior" perspective. And each post you have thrown out have a detrimental affect on all those you fought, and thus make "battlerager berserker warrior" better. Your objectivity is utterly lost. With the change you are asking, you would lower the percentage chance of a mage getting a spell off AND affecting you by 25% or more. This does not even take into account spellbane. Next time, weigh the pros and cons of the point you are trying to make. I would think you would do that every day being an attorney.