I'd suggest a pre soak for all the sheets if using the shuffling method, shuffling in the pre soak bath just as you will during the rest of the process.
I'd also suggest using a larger tray than that which physically fits the film. This will give you plenty of developer to work with, plenty of room to shuffle, and will minimize the effects of any non-random flow patterns that develop in the tray. For instance, I'd use 5x7 trays for 4x5 film, and 8x10 trays for 5x7 film. If I shot 8x10, I'd use an 11x14 tray. (For 2x3 sheet film, and when I had a 3x4 camera, I'd use 4x5 trays, though it took me a while to track a set down, and they eventually ended up in my hands by chance. I never actually found a set for sale.)
This is a continuous agitation technique. I've had ruined film by letting it sit just a short amount of time. Be especially attentive to continuous agitation for the first half of the development time.
Because it is a different agitation technique than you use by doing one sheet at a time in a tray, you may find that you need to change your development times to get your intended contrast.
I try to vary the direction of the draining. I drain from opposite corners each time through the stack.
You can mark the first sheet submerged with a sharp blade on the edge (Making an extra notch near the notch code is one way to do it.), or by turning it 90 degrees from the rest of the stack. I slice a top corner or two on the first sheet, and also turn it 90 degrees.
Development time will be staggered with this method. I put in a new sheet every 30 seconds, and when it is time to stop the development, I pull out the first sheet after the normal developing time, the second after 30 additional seconds, and so on. Therefore, for each sheet beyond one, I increase the time I set on the timer by 30 seconds. For example, if the developing time is nine minutes, and I have four sheets, I set the timer to 10:30 (30 seconds additional for every sheet beyond the first). When nine minutes is up (clock hits 1:30), I take out the marked sheet, then the second sheet at 1:00, and so on. A slick way to do this to help avoid confusion is to rig two Gra-Lab timers together, so that the first timer hitting zero triggers the second timer to start counting. For example, with this method, the first timer would be set to 9 minutes, and the second to 1:30. When the first timer hits zero, you pull the first sheet, then pull the other sheets in order every 30 seconds on the other timer. This might be convenient for some, even though all it does is prevent you from having to subtract one time from another and remember the difference.
Because of the staggered stopping meaning that fingers will be going into and out of different solutions, and then repeating, I don't use a stop bath, to prevent contamination of the developer from my fingers. I use a tray of water, with a stream running into it if possible.
Make sure that the agitation is coming only from shuffling. IME, any agitation by rocking can easily scratch the emulsion (and can also turn the sheets, so you lose track of which is the first).
I have pretty much given up on the shuffle method personally. I was never able to get consistently scratch-free film, no matter how careful I was. However, by following the tips above, I was able to get nicely developed and even negatives.
