[mc.pitchshift~] not responding as expected

I'm trying to incorporate an [mc.~]-wrapped pitchshifter into a patch. I've worked with the old [pitchshift~] plenty of times and gotten good results and I want to take advantage of the deviate function of the mc-wrapper for pitch shifting. However, after much poking and prodding, as well as research online and in various help resources, I have been unable to find a) instructions on how properly to patch this thing so it works as I want it to (if the mistake is mine) or b) other persons having similar trouble (if it's a bug).

Specifically, I want to apply a deviate message to -- for example -- [mc.pitchshift~ @chans 10] so as to produce a distribution of shifted pitches, e.g., between + 20 cents and - 20 cents. I'd prefer to shift in cents (as per the pitchshiftcents message), but I'd make do with shifting by a factor. The error I get when I apply deviate (in the [deviate $1 0] format) is 'pitchshift~: doesn't understand "float".' I'm interpreting this to mean that the second numerical message (in this case "0") is unexpected by the [pitchshift~] objects embedded in the [mc.~] wrapper. Conversely, if I send a pitchshift message (format: [pitchshiftcent $1]), I get no error, but then all channels shift identically. Here's an example:

It seems reasonable that either a) the [mc.~] wrapped object is supposed to work as the unwrapped object does, or (perhaps more likely?) b) the wrapped object requires that the pitchshiftcent messages be deviated and targeted before they are sent into the wrapper. If the former, perhaps a person in authority might know of and/or identify the bug and I can come back to it after it's fixed. If the latter, I'm hoping a sagely person on this forum might be able to guide my thinking as to an efficient way of doing it, since all my attempts at solutions have essentially been non-mc.~ workarounds (and therefore defeat my purpose).

Thanks!