Glitched derivative on recursive function that involves nCr

Something’s not working the way it seems like it should.

When I defined a recursive function f(a,z)={1=1:nCr(2a,a),f(a,z)}, as well as a constant z=1, and typed d/dz f(2,z)

I saw the output is 3.5

But I expected 0 as the output, because f(a,z) is not actually dependent on z, it’s identically equal to nCr(2a,a). When I replaced nCr(2a,a) with (2a)!/a!^2, it correctly output 0. Only nCr makes the derivative go wrong.

Here’s the link to the page: https://www.desmos.com/calculator/y3gmaqdcil

And here are some screenshots:

Thanks for reporting this. You can report graphing calculator bugs to support@desmos.com. The forum here is for Activity Builder. While you can use the graphing calculator within Activity Builder, there is a separate team that handles the calculator.

This really looks like a bug.
I recreated it:
https://www.desmos.com/calculator/pj6bjno2b8

We may however point out that the function f, although well definied, has a strange circular depency because its definition looks as if it depends on itself. That is: the definition should be equivalent to f(a,x) = nCr(2a, a), but apparently it is not in desmos.
There is a bug.

Maybe a precision error? Also, it “fixed” back to being both 0. Maybe just a one time glitch for you or maybe the Desmos devs saw this post, I’m not sure.