Early firearms are less accurate than modern firearms. The main contributor to their inaccuracy is that the ball exits the smoothbore barrel with a random spin (which depends on exactly where and when the ball bounces on its journey along the barrel to freedom). A modern rifled gun, on the other hand, fires bullet after bullet with the same spin. Further, since the ammunition in early firearms varied a lot more, the weight of balls varied from ball to ball, and the powder charge varied from shot to shot. Modern ammunition is much more standard (but for highly accurate shooting, ammunition specially made for minimal variation from shot to shot is preferred).
How badly does this affect accuracy? At very short ranges, very little. At longer ranges, the random spin will move the ball in random directions, up or down, left or right. For a typical arquebus or musket, this gives approximately a 50% chance of missing a human-sized target at 100m, for a gun fixed in place so as to hit the centre of the target. In the tests by Krenn (1995), the spread of shot at 100m varied from about 30cm to over 1m (for both the vertical and horizontal spreads). Past 100m accuracy gets worse.
Thus, the inherent inaccuracy of early smoothbore firearms limits their effective range to about 100m (for, in principle, 50% accuracy), even though the guns have ranges of approximately 1000m. Up to distances of about 30m, the guns are accurate, and will hit human-sized targets they point at when fired.
For battlefield use, early firearms have an additional disadvantage over modern guns: there is a longer delay between pulling the trigger and the gun firing (the powder has to burn through the touchhole to ignite the main charge, and this takes time).
However, accuracies achieved in battle are far lower than the accuracy that tests with careful aim (e.g., 18th and 19th century tests discussed by Krenn) or tests with fixed guns (Krenn's tests). While Krenn's results show that close to 100% hits might be possible at 30m, battlefield results suggest that even at such point-blank range, hit rates might be as low as 20%, and rather than the technically-possible 50% at 100m, as low as a few %.
However, low levels of accuracy in battle are not solely achieved with early firearms, but have been readily duplicated with modern guns. E.g., Rostker (2008) reports hit rates by police (with modern guns) of 30% against targets not returning fire, and below 20% against targets returning fire. At 7 yards or less, 37% accuracy was achieved.
Thus, while the accuracy of early firearms was limited by technical considerations, the effect of the stress of battle on the soldiers probably had a bigger impact on accuracy in battle.
References
Peter Krenn, Paul Kalaus, and Bert Hall, "Material Culture and Military History: Test-Firing Early Modern Small Arms", Material Culture Review / Revue de la culture matérielle, [S.l.], June 1995. https://journals.lib.unb.ca/index.php/MCR/article/view/17669/22312
Bernard D. Rostker et al., Evaluation of the New York City Police Department Firearm Training and Firearm-Discharge Review Process, RAND Corporation, 2008. http://www.nyc.gov/html/nypd/downloads/pdf/public_information/RAND_FirearmEvaluation.pdf