Tips for solving with ZBF2L
How to solve quickly using ZBF2L:
Obviously I am just a beginner at using the ZB method, and a master of this method might not agree with me, but here are my tips for solving using ZBF2L based on what I have learned so far.
1) The biggest barrier to solving with ZBF2L is making sure you have all the algs memorized. If you are using an intuitive ZBF2L approach rather than memorizing all the algs, then it is essential that you know exactly how you would react to each situation in which you would attempt a ZBF2L solve. For those memorizing the algs, I highly recommend the following method for rehearsing the cases when you don't have a cube around to practice on. Picture the 8 cases for one of the standard 23 F2L cases in turn, and go through them in your head. In doing this notice two things, 1) How the pair is placed together, and 2) at what point the flipped edges are corrected. Make sure you can do this for each of the 8 cases within a standard F2L case. This usually takes 5 minutes or so, since it is harder to do these moves in your head without a cube. You can do this on the way to class, in line to pay for something, or when you are taking a break from homework or work.
2) Once you know all the algs, the biggest barrier will be recalling the algs when they come up (under stress, knowing the timer is running and that you better recall it quickly). If you use intuitive ZBF2L approaches I guess this wouldn't be as much of a problem, but if you have memorized the algs this is a very big problem. I am still dealing with this, but the next few tips below are methods I have found to help me deal with this problem.
How ZBF2L differs from F2L
The main thing I have noticed with using ZBF2L is that it is completely different from F2L in terms of solving pace. This may not seem like that big of a difference, but you may not realize how hardwired your F2L pace is in your solving. Try solving the F2L just slightly slower than your normally do. You may find that you either speed up again without realizing, or that you have a hard time maintaining your normal F2L flow at the different speed.
The pacing for ZBF2L is very different from F2L in that it is significantly slower. I don't claim to be a master of standard F2L, but I do get my best times when I slowly accelerate from the cross at 60%-70% speed to the last corner/edge pair at 100% speed. I find that starting slower and accelerating during the solve works for the ZBF2L as well, except that you have to start slower than you do for F2L.
On a ZBF2L solve I start off slightly slower for the cross than for the F2L, maybe 10%-20% slower. I have noticed a few advantages to this. Because I am solving the cross at around 40%-50% speed, I find it is much easier to follow the pieces for an extended cross solve. I find myself using an extended cross or 2x2x2 + other blocks approach for the cross much more often than I did with Fridrich F2L. I think this has definitely helped my times, and I hope to continue to use this strategy in the future.
The strategy for the ZBF2L is also different than for the F2L. For the F2L you are trying to solve the first two layers quickly, leading into your first "look" when you apply the correct OLL algorithm (or optional COLL if you have oriented edges). The strategy for the ZBF2L is to solve the cross and 3 corner/edge pairs quickly. This is shorter than a normal F2L solve, which means that your acceleration from cross through each pair must go faster than with standard F2L. Also, you can't go 100% speed on the 3rd corner/edge pair, since that makes it hard to identify the last corner/edge pair. Remember that the strategy for the ZBF2L is to solve the cross and 3 corner/edge pairs quickly, not the entire first two layers.
The last corner/edge pair
This is obviously different between the two methods, because in standard F2L it still falls under the "intuitive" portion of F2L, but in ZBF2L (with memorized algs, and not done intuitively) it becomes your first look. Also, I advocate learning the VH method (LL is done as COLL/PLL) variation where you use ZBF2L as your basis, so that means you will now be solving with a 3-look method, instead of a 2 look. You'd be surprised how acclimated your brain gets to only having to do 2 looks. Having to do three, when being used to two, can throw off your flow during a solve sometimes. It can also be confusing on the 1/27 chance that your corners come up oriented. This is technically a ZBLL solve, however your brain won't register it the same way as if you actually know the ZBLL case when some of the corners are flipped. Normally the only time you see oriented corners other than the 1/27 chance is if you mess up the ZBF2L alg and have to finish with OLL/PLL. Even though you are given an easy ZBLL case, you will still feel as if you messed up.
Don't practice your old method
This is by far the most frustrating part for me. Using Fridrich I can very consistently average under 18 seconds, however with ZBF2L I am only just now starting to average sub-20 consistently. It's frustrating for me to know that I know how to average sub-18, but that I am purposefully not doing it. Of course learning ZB is an experiment for me, and I don't know what the outcome will be, however I have a lot of faith in the usefulness of the method once you know all of it. Therefore if you are serious in learning ZB, don't practice with your old method - at all. Don't even practice it for fun or to maintain your regular level of skill in solving with that method. Learning ZB is such an undertaking that you have to devote your brain to it totally. The only time I practice anyting from Fridrich is to practice the OLL algs every now and then, since I would use an OLL/PLL approach on some very special cases. Also, the frustration of knowing I can't average my normal Fridrich times with ZBF2L is a big part of my motivation to continue trying to perfect my strategy.
Again, I'm no master of ZBF2L, so this may not be the best advice for learning it. All I can say is that thinking this way has helped me to narrow my ZBF2L averages from 30 seconds to consistent sub-20.