Thoughts on the VH system

I've heard a lot of doubt in ZB recently, and I thought I could shed some light on the VH system, the "beginner" way to do ZB. It requires knowing 40 COLL algorithms and 16 VHF2L algorithms (which you can then mirror), though there are many mirrors or one move different algs for the system. This is what happens:

Cross (or X-Cross): basic. Get a cross on the bottom.

F2L minus one pair: insert all F2L pairs except the last one. Still basic.

VHF2L in the last pair: create the edge corner pair and perform one of 16 unique algs (I don't count mirrors). This will orient all edges on the last layer, giving you one of 7 OLL cases or an OLL skip.

COLL: using the appropriate COLL algorithm, permute and orient all corners while keeping the edges oriented. If your OLL skipped, go to normal PLL.

EPLL: Now all you have to do is permute your edges, most likely your fastest PLLs (except Z...). Easy to recognize, fast to perform.

So my thoughts? COLL is brilliant. The cases are fast (usually), easy to recognize, and they get you almost instantaneous recognition of your PLL, with a 1/12 chance of a skip to boot. A few cases I'm not happy with, but it doesn't matter too much. COLL + EPLL is about 15% faster for me than OLL + PLL. However...

VHF2L is rather iffy. I just do partial edge control instead of bothering with these algorithms. ZBF2L is rather iffy as well, especially if you use intuitive F2L over algorithmic. Regular F2L is about 50% faster than VHF2L for me.

So that means that the CFOP system is about equal to the VH-CFOP system (remember that the last F2L slot is REALLY fast with Fridrich, so 50% faster is about the same as a 15% slower last layer), though full ZB is a different story...