You're doing a math that i can't condone:

*1 harmful skill every 2s, with 10% chance to fire a proc on each use, means that after 20s i'm guaranteed to have the proc up. *
If only things were that easy:

In real life you have 50% to win a head-tails game, but you're not guaranteed to get at least one heads after 2 trys just because your chance is 50% in each try. Playing heads-tails works like this; you get heads 50% of the time on the first try and 50% of the time on the second try:

1st: 50% heads

2nd: 50% of the 1st try + (50% of 50% on the second try)= 75%

For a 10% proc scenario this means:

1st use:10% chance firing proc, 90% fail

2nd : add another 9% for a total of 19% chance=10% of first try + 10% of 90%

3rd: add another ~8.1%, total 27.1%

4th: add another ~7.3%, total 35.4%

5th: add another ~6.6%, total 42%

6th: add another ~5.8%, total 48%

7th: add another ~5.2%, total 53%

8th: add another ~4.7%%, total 58%

9th: add another ~4.3%, total 62%

10th add another ~3.8%, total ~65%

Sources:

http://en.wikipedia.org/wiki/Bayes'_theorem

http://en.wikipedia.org/wiki/Bayes%27_rule
Someone more versed may help us with a tree diagram, but i decided to crunch the exact numbers for the Gold bracelet:

1st:15% to fire up proc

2nd: 15% + (15% of 85%)= 27.5%

3rd: 27.5% + (15% of 72.5%)= 38,375%

4th: 38.375% + (15% of 61.625%)= 47.61875%

5th:47.61875% + (15% of 52.38125)= 55.4759375

Having half more chance to fire the proc implies that you gonna need ~half as many tries to activate it, but by now we see were this is going:

6th: 55.4759375 + (15% of 44,5240625%)= 62,154546875%

7th: 62.154546875 + (15% of 37,845453125%) = 67,83136484375%

8th: 67,83136484375 + (15% of 32.16863515625%)= 72,6566601171875%

9th: 72,6566601171875% + (15% of 27.3433398828125%)= 76.75816109960938%

10th: 76.75816109960938% + ( 15% of 23.24183890039062%) = 80.24443693466797%

The Gold Bracelet with a 15% proc chance keeps the buff up ~80% in a "

*1 harmful skills every 2s*" scenario. Please note that while increasing the skill frequency would increase the % of time the proc is up, we would still be a long way from the dreamed "proc up all the time" situation you dreamed about. Making it 2 skills every 3s would end up with ~ 88% proc up on average.

Turbine did a great job adding proc chances to itemization, adding another tool to stratify math gaming skills among players.

Looking back at numbers and items, the defensive proc with + 1837 tact and phys mitigation looks yummy, not only because you're gambling in 4 stat slices instead of 2 and a half, but mainly because the more you are getting hit, more likely will the proc fire up, and we can get hit at a frequency much higher than once every 2s.