All the built in filters are supposed to be butterworths (including the 1st orders), but I just measured them in response to your query and I can confirm your measurements. My measurements suggest the 6 dB/octave filters are -6dB at the crossover point not -3dB, which is incorrect. I'm not sure what to even call it; I don't think there's a named filter slope that's 6dB an octave and -6dB at the crossover point (no such thing as a 1st order linkwitz riley, etc.). In any event I measured -12dB at the crossover point by stacking two of them, which is close enough to your -13 to say we're likely seeing the same issue.
This made me worried about the other filters, but I re-measured the higher order filters (which I have been relying on for my own crossover) and I can confirm that they are correct butterworths (-3dB at the crossover point). That part is unsurprising because I'd measured the higher order filters many times in setting up my speakers, but obviously I wanted to reconfirm. The 6dB/octave filter was a fairly recent addition (within the last year or so) and I don't have speakers that could benefit from a slope that shallow; I was assuming they worked the same way as the other filters, but clearly they do not as currently implemented. Good catch.
Until this gets fixed, I can confirm that two of MC's 12dB filters stacked make a working 4th order Linkwitz Riley (which is what I've been using myself). If you really need a 2nd order response right now, you can use 2nd order butterworths plus a -3dB PEQ notch to get rid of the lump. The latter approach will produce a LR-2-like slope and phase response because MC's filters are all minimum phase.
Hopefully Matt can fix this though, so work arounds are not necessary. This is not expected behavior from a first order filter:
https://en.wikipedia.org/wiki/Low-pass_filter#Continuous-time_low-pass_filtershttps://en.wikipedia.org/wiki/File:Butterworth_response.svg.
