Slate put out an article with the percentages of spare conversions of the "hardest" spares to convert. Surprisingly it's not the 7-10 split, but the 4,6,7,9,10 "Greek Church", the 4,6,7 and then the 7-10 split. I've actually never seen a "Greek Church" converted, but have seen a few 7-10 splits converted. It makes me wonder why the 7-10 split in perceived as the hardest to convert at times. 

http://www.slate.com/articles/sports/sports_nut/2015/02/hardest_shot_in_bowling_it_s_not_the_7_10_split_it_s_the_greek_church.html

I'd say the 2-8-10 for me personally or the 4-6-7. I've never been able to get a bounce in the back to get it and can't get enough angle being right handed to get the 4-6-7 without missing the 4

I've made the big 4 twice, once sliding the 6 over and once getting a bounce from the back. I got the bounce for the 7-10 once, and slid the 6 over to make the 4-6

The data is severely flawed. It is not the most difficult but the data shows that it is the least converted by percentage. It does not take into account that many times on a greek church in a PBA tournament a bowler is more apt to take the count and not risk a low count by going for it and picking the 7 or 10 off clean. The 7-10 would be converted more because going for count and trying to pick it up don't vary nearly as much as it does when approaching the greek church.

Seen the Greek church made twice. Once against my son in a tournament. Kid wasn't even trying to pick it up. Was trying for the three. Thought he missed everything but the 10.

My son made it earlier this year in his league. Has gotten 4 of the 5 two other times. He loves shooting at it. I just tell him I'd prefer he not get so much practice at making it.