API, create experiment or add trees to experiment -- if an incorrect honest expert email is provided, the designation of that honest expert will fail silently

[brachbach]

trafficstars

(as opposed to if an incorrect malicious expert email is provided, in which case it will error out)

Note that my data on this from creating trees with 2 honest and 1 malicious experts -- so the difference in behavior between the honest and malicious cases may be due to 1 vs. >1 experts, as opposed to honest vs. malicious