Re: How positively buoyant are the seafrogs housings?
there are too many variables for a useful answer here without someone using the seafrog with a compact mirorrless and the same strobe arm arrangement as you.
generally speaking, the compact housings with cameras inside are negative, and more so when you add on the strobes. That's why there are so many float and fat strobe arm options, and float rings for the macro port. These rigs are commonly paired with a fisheye and a 4" dome for wide angle, which doesn't make up for the weight.
That said, if a manufacturer uses a larger plastic shell to cover a wide range of models, rather than bespoke molds with little dead space, they will be lighter. And if that goes to an extreme, the weight of the strobe/arms isn't enough by itself as it creates a rig that wants to turtle. Ideally you want to screw in ballast to the bottom of the housing to encourage the upright position.
But I'd be surprised if this is actually needed for the M50.
My M4/3 system (GH4) required a lot of floatation. My SLR one (R5) needed virtually none - one float at the top of the strobes to stop the pivot. I think I could do it by just making neoprene covers for the strobes, which would be good for them anyhow. But there is a lot of experimentation to do and it applies to every lens config. Salinity variation means you won't be dead on, but that is a tiny amount. Ideally, IMO, you're ever slightly positive so that for macro, you're pushing/pulling it down to steady, rather than having to apply a continuous lifting force. The latter results in a sore elbow after the dive, and false DCS fears.
One advantage to negative that occasionally comes up is the ability to put the rig down briefly. Once in a bull moon I've done this. I think there are more cases for it being floaty, most notably when your coil breaks during a deco stop in Truk Lagoon over a 300' sand bottom. (don't trust plastic coil systems to last forever - they will eventually fail)