The most beneficial play to training consistency is the first play since you have no memory of what is in the map. Each successive play becomes worse and worse for training consistency. Given a countably infinite set of maps of equal difficulty (whatever the fuck this actually means) and 0 repetition across all of the maps and music, the most efficient way to train consistency is to play each map once.
It is obvious that such an infinite map set does not exist and even if it were finite it probably would not either, nor is such an example particularly helpful since if there was any sort of repetition (note that repetition is not well defined but should include things like geometric patterns, correspondences between rhythmic patterns and hit circle mapping, rhythmic patterns, etc.) then playing maps that have the most in common with other maps that you want to be consistent at is more effective than maps that don't.
Also, if we have maps of different difficulty (however you want to define this will work), then there is a possibility that the nth play of one map is more beneficial than the first play of another map. An extreme example of this might be to consider a 4 hr long marathon map with an extreme variety of mapping and an extreme difficulty (never dipping 7* on a 30 second interval perhaps) versus an easy TV Size (perhaps 5* average but only really 4.4* in most areas with one 15 second section that is like 5.6* difficulty perhaps). You could probably play the marathon a lot before it becomes better to move onto the tv size.
Since there are too many factors, we can't really compare the nth play of one map versus the 1st play of another. Assuming that the maps you play are close enough to a situation where we can compare them and that there are enough of them, then the "greedy" (Comp Sci term) answer is to just play each map once. This is the accepted answer and since nothing is defined well enough there is no way to give an actual proof that one answer is better than the other.
(Note that this also means there is no "best" answer and this applies to a good 90% of the questions that are asked here. Any answer claiming to be absolute is usually wrong though, unless it's in response to an absolute question such as "I'm doing x,y,z when I play. Will this make it impossible for me to improve" and the answer is no)
It is obvious that such an infinite map set does not exist and even if it were finite it probably would not either, nor is such an example particularly helpful since if there was any sort of repetition (note that repetition is not well defined but should include things like geometric patterns, correspondences between rhythmic patterns and hit circle mapping, rhythmic patterns, etc.) then playing maps that have the most in common with other maps that you want to be consistent at is more effective than maps that don't.
Also, if we have maps of different difficulty (however you want to define this will work), then there is a possibility that the nth play of one map is more beneficial than the first play of another map. An extreme example of this might be to consider a 4 hr long marathon map with an extreme variety of mapping and an extreme difficulty (never dipping 7* on a 30 second interval perhaps) versus an easy TV Size (perhaps 5* average but only really 4.4* in most areas with one 15 second section that is like 5.6* difficulty perhaps). You could probably play the marathon a lot before it becomes better to move onto the tv size.
Since there are too many factors, we can't really compare the nth play of one map versus the 1st play of another. Assuming that the maps you play are close enough to a situation where we can compare them and that there are enough of them, then the "greedy" (Comp Sci term) answer is to just play each map once. This is the accepted answer and since nothing is defined well enough there is no way to give an actual proof that one answer is better than the other.
(Note that this also means there is no "best" answer and this applies to a good 90% of the questions that are asked here. Any answer claiming to be absolute is usually wrong though, unless it's in response to an absolute question such as "I'm doing x,y,z when I play. Will this make it impossible for me to improve" and the answer is no)