Mr. Bumpy, this is a massive "can of worms".
I suppose my summary is, don't mess with it unless you can MEASURE your results. Changing drivers and changing values will almost surely make matters worse. Fortunately, the cost of measurement equipment has declined. Here is a fair primer on measurment:https://www.audioxpress.com/article/testing-loudspeakers-which-measurements-matter-part-1
It is marginally possible to obtain tolerable sound or even acceptable sound when tuning a speaker by ear, but the most optimal results will always be obtained with MEASUREMENT and listening. Using on & off axis gated and in room responses combined with listening to music in your room is a very process oriented project. Additionally, a simple 2 way will be much easier than a 3 way due to the low frequency crossover room reflections and the difficulty in measuring these frequencies.
In general the function of crossover components as I consider them:
Capacitors - stop low frequency and pass high frequency. A larger capacitor will pass lower frequencies. A smaller capacitor will only pass higher frequencies.
Inductors - stop high frequency and pass low frequency. A larger inductor will pass less high frequency. A smaller inductor will pass more high frequency.
Additionally phase shift occurs across these components and is dependent on impedance, frequency and component value. This is a starting primer for inductance phase shift:https://www.allaboutcircuits.com/textbook/alternating-current/chpt-3/ac-inductor-circuits/
In my opinion, the traditional mathematical models created by Mr. Linkwitz and others were groundbreaking in their time. However, for today's designer, the idiosyncrasies present in ALL loudspeaker drivers and the ability to measure and create a custom crossover for these drivers will vastly surpass a mathematically applied model.
Designing loudspeaker crossover circuits can be super fun, but there IS a learning curve - pun intended