A trial would be from a run of strikings that were used to test out a particular feature. An example of this would be my trial halfcrown struck from halfpenny dies. See below. However, during a conversation with Graham Dyer a few months ago, I learned that they were not interested in either halfpennies or halfcrowns, but rather the metal (steel). Therefore, trials can be unrelated to the die size or design.
An alternative would be a striking of a single die in lead or tin in order to assess the state of work in progress, an example of which would be the Moore penny uniface below. This is probably made from the P2135 obverse die, which developed a major crack and would have been reworked. This die was not to be found when Shorthouse visited Moore's workshop following his demise, but combined with the known die breaks and a hint of undertype on this obverse, is likely to mean this was the same piece of diestock. It is struck in >90% tin, with the balance mostly lead.
Anomalies could be things such as coins struck on the wrong blank, or the wrong coin struck on a particular blank depending on viewpoint.