So basically, it's below the "Reliability*Utility" threshold (that's my term for the drastic dropoff in usefulness of a product/service/person when their reliability goes from 100% to 90% or lower) That's too bad. Maybe some unit tests (historically notoriously difficult to implement with GUI's) would help that dev.
Fleshing my concept out on a whim (since I am surprised I can't find another reference to this idea online!) with the help of ChatGPT (yes, it can do this now!), it looks like this equation would capture this concept:
U(r)=Umax ⋅e^(−k(100−r)^n)
where the total perceived Utility of a product/service/trusted person, given r=reliability percent (out of 100), is the maximum Utility times e to the (-k(100-r)^n), where k and n are tweakable based on the given product/service/person but which I want to assign values of 0.03 and 1.5 to, respectively, after plotting it.
> So basically, it's below the "Reliability*Utility" threshold (that's my term for the drastic dropoff in usefulness of a product/service/person when their reliability goes from 100% to 90% or lower)
Fleshing my concept out on a whim (since I am surprised I can't find another reference to this idea online!) with the help of ChatGPT (yes, it can do this now!), it looks like this equation would capture this concept:
U(r)=Umax ⋅e^(−k(100−r)^n)
where the total perceived Utility of a product/service/trusted person, given r=reliability percent (out of 100), is the maximum Utility times e to the (-k(100-r)^n), where k and n are tweakable based on the given product/service/person but which I want to assign values of 0.03 and 1.5 to, respectively, after plotting it.