Do they ever miss on their designs?
Do they ever overbalance?
It depends on what you mean by "miss". Miss is is a good word to use because crankshaft design is like the Whack-a-mole game, there are many more issues to deal with besides counterweight design in crankshaft design. Take for example the 2.0 main cranks, the trade-off for reduced bearing drag is a crank that is very flexible. Flexibility brings lots of new problems such as torsional vibration, and distortion because a crank distorts in contorted ways due to it's offset rod pins. One of the problems that comes with distortion is the snout wobbles out of position which causes problems for the cam drive.
There are lot's of things you can do to try to reduce the problem by stiffening the crank but they are all trade-off compromises. For example you can add stiffness by making the rod throw arms stronger, but there are an infinite number of variations for how to do that and just simply making them wider or thicker won't get you competitive anymore.
Making the compromise to have stronger but heavier rod throws will bring new problems, for example it might change the frequency of torsional vibration so that it coincides with the frequency of a force in the engine and now the problems are worse than they were before strengthening the crank. When you run into this problem you have to get weight off the crank or move it to other locations to change the natural frequency. That means another compromise like using counterweight designs that are not what you would choose for bearing loads but they make the crank lighter. As you can see, it is all involved with trade-off compromises, you can't just design a crank that is strong or light and say it is the best design.
So given all those variables and the situation that the team may be able to simulate some of the dynamics but may not have all of the data or technical skills or software or time/priority required to simulate everything, at some point they have to do experiments to determine the consequence of a change. Sometimes those experiments are made without expectation that the problem will be cured, they just need to have data about the consequence of changing some design parameter.
So after many cycles of improvement inevitably a change is made elsewhere in the engine that for example results in faster combustion that introduces a new frequency for the crank to deal with and avoid vibration. It's a game of whack-a-mole, there is no such thing as a final perfect design.
As an engine reaches higher performance levels it becomes impossible to make the required shapes from economical crankshaft forgings.
As you can probably see from this process, in all of this compromise and relocating of counterweights to solve a specific problem, balancing to some particular bob-weight that could result in any infinite number of non-specific results becomes meaningless.
Evaluating a counterweight design and balance by something as simple as bob-weight and over/under-balance is like trying to determine if your meal is nutritious by weighing it, you need more specific information, the weight doesn't tell you much that is useful to determine if the meal is nutritious or not.
