Could be that it doesn't work at all too so fingers crossed! This setup only models for two coefficients but that should be enough that exactly leveling my Dobsonian will no-longer be necessary. a known-good mount modeling system in this mode. So with that work done I turned my attention to the geometric Alt/Az mode and that also appears to be working though it's harder to tell for sure since I can't compare the results vs. So I fixed that in OnStep (3.) also, and for the first time I see what looks like OnStep's internal GTA working in simulation. Naturally method (3.) internal to OnStep is a port of SP's GTA and so had this problem/bug also. Once corrected I could, for the first time, do an align in OnStep (2.) and pull the data into SP and visualize the alignment points with the mount model applied on the map. This "bug" didn't keep SP GTA from working correctly nor was there a problem with OnStep's modeling it's just a matter of the two not working in the same exact way. I then hunted down and fixed that problem.
I tested that code first on the well-proven method (2.) and noticed a coefficient "polarity" issue. In preparation for this work I added the ability to pull the alignment points (mount/star coords) and index offsets from OnStep and drop them into SP's GTA to enable better visualization of what was going on with Alt/Az aligns. Those tools being Sky Planetarium (SP) and Goto Assist (GTA) integration with OnStep. :) The upside is my simulation tools are better now than during earlier testing. Things look to be working in simulation though history has shown that this doesn't necessarily mean anything. Instead of continuing down that path I choose a more unified approach and adapted my geometric (Eq) mount modeling code to Alt/Az mounts. It worked in a basic sense, as good as a one-star align but using two stars, then beyond that additional work was needed to derive the coefficients that model the mount to improve performance. The Taki derived code in my testing didn't work well not because of a fundamental problem, but due to it's being incomplete. (3.) Geometric align based on any number of points (but limited to 9 inside OnStep, 6 in my Apps) for Eq mounts "Goto Assist". (2.) My geometric align based on 3 points for Eq mounts. After the change: (1.) My geometric align based on 3 points for Alt/Az mounts. I've made some changes to the alignment code in OnStep (unreleased.) The biggest change is a switch to geometric modeling for Alt/Az mounts in place of using the Taki method.