My numbers were with Patience, as I said. If you don't like the attack speed slowdown that's fine but when discussing crit options it exists. Your numbers above ignore it, which is why you don't arrive at 11%.
I showed patience numbers in two spots in that post.
If you include patience rate reduction, multiplying by 0.90 isn't correct, but it's close, for 6.2% and 5.4%.
Here's the other spot.
Kukri+DL = DPS+18%. Kukri+Patience = DPS+5.4%.
If you want to compare Kukri+DL to Kukri+DL+Patience, that's 3.2%, not 11%. It's such a small primary bonus that almost any imbue/sneak/procs make patience a bad deal. Here's the breakdown:
15x4x2=41 15x5x2=47 47/41=14.6% @ 90% rate =3.2%
Unfortunately, I can only show how to calc the exact scalars. It's very simple, just adding up hits. You can approach the counting slightly differently for the same result. This might make more sense:
rolls 2->14=13 are regular + rolls 15->18=4 are x4 hits=16 + rolls 19-20=2 are x6=12
add them up for 13+16+12=41
the reason this works is that you're essentially adding up the chance to get each hit roll and multiplying each one by it's critical multiplier; eg.
1 + 1 + 1 .. + 1 (13 times for rolls 2->14)
4 + 4 + 4 + 4 for rolls 15 to 18
6 + 6 for rolls 19 to 20
I find it quicker to think of it slightly differently, but the results are identical either way:
19 regular hits, each from rolling 1~20
18=6x3 critical hits for 15~20 @ x4 (but already counted 1 in the 19, so multiply by 4-1=3)
04=2x2 fat crits for 19~20 @ x6 (but already counted 4 above, so multiply by 6-4=2)
My guess would be using 19x1 = 10% is where your errors started; that only applies to a 20x1 weapon. You need to consider each profile in full.