Hey tejon, thank you for giving this a try, but that formula is contrary to many things I want. It is at its core very similar to my formula for lockpicking - in a way that I could almost simply combine a failed lockpicking attempt with breaking the lockpick, i.e. returning to vanilla Oblivion's system. (A system which I found not convincing at all.)
What I had in mind was a http://upload.wikimedia.org/wikipedia/de/thumb/e/e7/LogWachs1.svg/390px-LogWachs1.svg.png for the chance. For the first 10 uses, there's little chance that the lock breaks, but then it gradually rises, until at around 30 uses, it will almost be a certainty that the lock is about to break.
Edit: After some testing, this seems like a good formula:
Chance = 0,8 / ( 1 + exp(-0,4*(Uses – Buffer)) * 16 )
Buffer = 19 + Security/30 + Luck/50 + SkeletonKeyModifier
This is the formula for the chance which would be calculated for each time the lockpick is used. "Uses" is the number of times the lockpick has already been used, the SkeletonKeyModifier is 25 in case the player uses my mod's Skeleton Key. For example, the chance for a normal lockpick to break at its 25th use is 15% with Security and Luck at 50. However, since each single use already calculates a chance like this, the chance that the lockpick WILL be broken after the 25th use - no matter if it was actually broken at the 14th use, at the 20th use or at the 25th - is 54%!
The chance that exactly the 25th use will break the lockpick is 11% (this number gets lower from the 27th use one, as then it will be much more likely that one of the uses BEFORE will already have broken the lockpick).
Edit 2:
I made three linear approximations from the curve that resulted from the formula above; basically, it is now a linear function that is capped at the values 0 and 80.
Chance = 4 * Uses - 55 - 0.2 * ( Security + Luck )
If this value is below 0 or above 80, it is capped to the respective value. For the probabilities, little has changed, but I was able to avoid the exponential component of the old formula, and keep everything rather simple. If we can agree on this calculation for the chance of breaking your lockpick, then I should be able to upload a new version tomorrow (need to do some testing first).

Edit 3:
http://img534.imageshack.us/img534/738/chancea.jpg (to make this less confusing: e.g. after 20 uses you roughly have a chance of 68% of NOT having broken your lockpick.)
(This diagram was made with (Luck + Security) = 50. With values that make this sum, you have an average lifespan somewhere in the area of 23 uses.)