With three grits, I'd personally place emphasis on the 1st and last stages. At the 1st stage, a universally effective coarse grit for setting bevels on any steel (a coarse diamond would handle that). And the 'right' combination of abrasive type, grit & backing substrate for stropping at the last step is of huge importance; it can really alter the cutting character of the finished edge in all the right or wrong ways. Whatever comes 'in between' is more up to individual preference, such as how toothy or polished the finished edge needs to be, according to preference, prior to stropping.
The first stage chosen is what will make it easy or difficult to quickly repair or re-bevel a damaged or blunt edge profile; if that can't be done effectively or in a timely manner, following stages won't matter much anyway. The last stage (stropping) will greatly affect & ideally enhance the edge characteristics created in the intermediate stage. I'd say the last stage is the trickiest to figure out, and should be built around selecting a stropping grit & substrate combination that will both be effective in burr removal, and hard enough to work on high-wear steels (if applicable), but not so aggressive as to completely erase or obliterate the best attributes of the edge created from the stones.
Having said all the above, I tend to think anything from 320 thru 2000 grit (ANSI or FEPA-P standards) works well for most tasks & uses. Anything in that range would be what I'd likely choose for the intermediate step. Stay to the lower 'toothy' half of the range for draw-cutting tasks in coarse/fibrous materials, or to the higher end if fine slicing or chopping uses are the main priority. Either way, I've found anything within that range will work well enough for either use, most of the time.
What David Martin said, about grit type and binder/backing is important. Even at equivalent 'grit' size ratings, different abrasive types (diamond, silicon carbide, aluminum oxide, natural/Arkansas) will perform in vastly different ways on the same given steel. Some won't be as effective on high-wear steels. And differently again, depending on the binders used and/or overall quality of the stone. The abrasive type needs to be chosen first, appropriate to the steel(s) that will be sharpened. Something like diamond for very high-wear steels with heavy vanadium carbide content (S30V, etc.), or perhaps natural/Arkansas stones if only simple low-alloy carbon steels are the only ones used. For the most part, silicon carbide or aluminum oxide can handle most steels; the Norton stones suggested by DM are good examples, in both abrasive type and overall quality.
Very hard to recommend any 'best' trio of grits/stones, as there are too many variables to consider in narrowing down to only three stones for everything that may be sharpened. More details about steel types and desired uses of the blades would help in choosing what might work best.
David