I'm not one to criticize someone's approach, but I'm also big on informed decisions, understanding what you actually get, not what you think you're getting.
There's a lot of that in this thread, mainly due to a single poor assumption - that speaker impedence is constant with frequency.
You're using a textbook crossover that assumes you're connecitng 8 ohm resistors to the bass, mediant, treble circuits. The resulting filters are flat up to a threshold, the "knee" which signals the transition to a 12dB/octave slope. I've never seen a speaker designed based on actual driver perfolrmance that looks like that. Let's see what might be happening...Tweeters
, like many speakers, have a tall peak in impedence, falling to a local minima before a long slow rise as frequency increases. Assuming the textbook frequency is above the minima, the rising trend just rolls off response at high frequencies, which you may have wanted to do anyway. Issues start with actual driver impedence vs. assumed impedence, leading to a level mismatch as well as a shift in filter freqeuncy from the calculated value. A 4 ohm tweeter in a circuit designed of an 8 ohm load will have a one octave shift upward in knee frequency and will be 3dB louder.Woofers
are not as simple. The impedence curve has a similar shape, but now we're going to use the region with the tall peak, so the impedence may vary by 10-20X. In addition, you have a TMWW, and dual woofers in parallel have half the impedence. Unless you have 16 ohm woofers, all the same issues arise with shifts in the filter knee, and level mismatches.
Your mid is just another woofer, but as a single, it really is an 8 ohm speaker, and so will perform the closest to what you expect... if it's got really flat frequency response.
To summarize...woofer and tweeter are likely to be crossing higher than expected, while the midrange driver is somewhere close to where it should be. The result is a broad peak from 450-900Hz, and a hole between 3KHz and 6KHz, with weak midrange due to higher output from 4 ohm woofer/tweeter sections. In addition, if the woofer has breakup modes at higher frequencies, like the one I linked, that distortion is still audible. You will have no BSC, as noted, so the low bass will be anemic if the speaker is out in the room. And any non-linearities in driver response are preserved if not exacerbated unless you're very lucky, or have a good ear and did a lot of tweaking.
I know this comes off negative; I don't see an alternative that informs your decisions. There are significant downsides to your approach, and DIY "the right way" requires a non-trivial investment in measurement gear to characterize the drivers. Ironically, the important part - XO design simulation - is a freeware spreadsheet.
And that is the genesis of "proven designs," all the savings of a clean-sheet DIY without the up front investment in measurement equipment and learning curve But you've got to solder....