Questions about parallel forces

Why is torque error likely to be large?

Can an error in one measurement be negated by error in another?

Torque error, or any error in a measurement can be due to losses (such as friction), inaccuracies in the measurement equipment, or improperly performed measurements.

Yes, an error in one measurement can be negated by another error. The errors can also be additive, which is why uncertainties must assume the worst-case values added together.