Partition actually has two rates; one used for training that is the usual ratio of count to total, and another that is slightly biased away from zero. By never having attributed probabilities of zero, this allows logs of probabilities to be calculated on validation or excluded sets of data, used in Entropy R-Square.
https://www.jmp.com/support/help/en/17.0/#page/jmp/statistical-details-for-the-splitting-criterion.s...
In the Partition platform, for categorical responses, Rate is the same as "Prob" in the Distribution Platform, and "Prob" is a slightly biased weight.
This is definately confusing.
JMP Systems Engineer, Health and Life Sciences (Pharma)