Alas, as I keep reminding you, that calculation is actually wrong.

So while you are saying 1.7% is good, and higher is better, it is actually quite a bit higher than that.

ckpool software does not store or display the information required to calculate that correctly, so his reported blocks usually claim much better luck than it is.
(not sure why anyone would be happy about that ... but anyway)

As I've explained a number of times, the problem is that shares/diff must be added up per difficulty change,
b ut using a total shares / current diff is far from correct.

To make it clear in a valid example, if you did 1% diff for 2 weeks and difficulty was 100T,
but later include that in a 110T, suddenly that 1% diff has dropped to 100/110 = 0.91% diff
The % amount of work done hasn't changed, but the calculation has ignored that.

To give an example that may explain it better, but not very realistic,
Image diff is 100T and you do 50T work during that 2 weeks, you have now done 50% of a block work.
Then say next diff, it jumps up to 125T, so at the start of that 125T diff, suddenly the amount of work you have done drops,
if you calculate it incorrectly using 50/125 = 40% work
But you haven't done any less work %, just the calculation is wrong.
The work % value of a share, is based on the difficulty when the share was submitted, not much later when difficulty has changed.

When finding a block lasts for a few weeks or months, that % difference, when calculated incorrectly, becomes very noticeable and makes the pool look like it is luckier than it is.