Bearing in mind that I've never done this by hand, I'm just reporting the results of other people's work on the computer:
"1> Overall the front and back throws will counteract the center throws.
Balancing the throws will reduce the tendency of the crank to deflect, and reduce bearing pressures. "
"2> The front and back pistons going down counteract the two pistons going up. If an attempt is made to balance the pistons with counter weights it will reduce bearing forces vertically but add to the forces horizontally."
Yes, vertically, I think so horizontally.
"3> However, added counter weights will be effective in counteracting the mass of the lower end of the connecting rods. "
"The counterweight may be further increased to counteract the mass of the big end of the rod. "
The big end is the lower end?
"Past that amount, a small amount of additional counter weight may be added to counteract the effective lateral inertia of the connecting rod shank."
...and some of the piston mass as well.
"Basically, with a symetrical crankshaft, and matched components, the engine will be self balancing overall, but the crank shaft will tend to deflect in the direction of the offset mass at each throw. This will add to bearing pressure. Balancing the throws and the big ends will reduce the tendancy of the crank to deflect and reduce balance induced bearing pressures."
Crank deflections are more complex than that, that is why it is important to do a speed sweep. Most 4 cylinders I have worked on have crank resonances in the running range of the engine when excited by second order (L/r) forces.
An inefficient alternative is to split the engine up into 4 coupled single cylinder engines, work out the counterbalancing for each, and then glue it all back together.
That tends to be a heavy solution as you aren't taking advantage of some complex interactions, such as the #1 cylinder being able to 'see' some of the #3 counterweight's inertia.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.