Do you need to do anything differently when you anneal short cases? The cases in question are .357mag pistol brass necked down to .17.

I tried two methods I had read about here. First, use a steel rod in the neck, position the torch to blow on it, and pull off when your fingers get hot. With that short case, I don't think it got the neck hot enough. It didn't begin to glow dull red anyway.

Other method was stand the brass in a pan of water. Torch, then knock 'em over. Again, I didn't see any red glow. Left the torch on for up to 20 seconds. Could hear the edge where the brass hit the water hissing. I think the water was close enough to the neck that it was heat-sinking the brass too well. At any rate, I got the brass hot enough that is hissed pretty well when it hit the water. But never any hint of dull red glow.

Do I need to lower the water? Use mitts if use the "rod in the neck" method? Or am I OK if I just get a good hiss when the brass hits the water?

Everyone says the case in water doesn't work well. I tried it and managed to ruin some vg sako brass. If you can get the neck red hot you've already oversoftened the body of the case.
I tried the melted lead method (dip the neck of case in powdered graphite and then dip it in melted lead until the colour change in the neck is the same as that seen on reputable brass, e.g., Lapua) and it seems to be ok for 308 to 243 (the necks crack if not annealed).
But for more radical case forming, a short 357 to 17, you may need to consider Ken Howell's method (Designing and Forming Custom Cartridges) which at least has the benefit of more control - turning the polished case mouth in a small flame or ring of flames with a heat sink at the case head end, using temperature-sensing crayon to indicate when enough is enough. Howell devotes about 5 pages to what is a reasonably fiddly process.

Check out www.reloadingpro.com for a truncated but clear version of Howell's annealing process.

I've spent a good deal of time and effort turning 218 Bee cases into Mashburn Bee cases and have finally come up with the following method. First, I stand a single case in a pie pan filled with water that comes up the case about .5 inch to keep from overheating the web area. Second, with the ambient lighting in the room turned down I spin the pie pan slowly twice around while playing hottest part of the flame of the propane torch at the neck area of the case. The case will change color and glow at the tip of the flame only. Upon completion of the second trip around I knock the case over into and fully covered by the water.
So far this technique has decreased shoulder failure during fireforming from about 40% to zero so, I'm assuming that it works.