Couldn’t be too much worse than some of the stuff GM put out- the '61 Pontiac 4, or that split-pin V6 crank. I guess it would have to be a long crank and engine, with a pin for each cylinder. Probably wouldn’t be worth trying, but I find weird engine configurations fun to think about.
I think you might be able to balance it with an inline-6 if you matched the cylinders that were in sync; say 1 and 6 are oversquare, 3 and 4 are undersquare.
I see other problems, though. At the top of their strokes the pistons would be at different distances from the crank. What would the head look like? Assuming you wanted the same compression ratio in each cylinder, you might need a separate head for each cylinder, or at the very least some odd configuration where the head gasket was not parallel to the axis of the crankshaft.
I’ve heard of some odd ones. Wasn’t there a Honda motorcycle engine some years ago that had oval cylinders? And BRM raced an H-16 engine in Formula 1 back in the '60s.
By the way in 1967 BRM eventually got 420hp @11,000rpm out of that unusual 3 liter engine. They expected close to 600. It was not a success though it won I think one race.
Yes, what you are describing about bore and stroke has been proven mathematically but as we all know in the real world things work out differently. 2 Cylinders with equal displacement regardless of their bores and strokes should apply an equal amount of force to the crankshaft. http://craig.backfire.ca/img/torque-bore-stroke.png