Hemi,
There are options to filter stars used for alignment in the manual see sections 7.1-7.3.
https://inter-static.skywatcher.com/...61208V1-EN.pdf
The next issue is whether you have set elevation limits - see section 6.2 of the manual - eg nothing below 15 degrees of the horizon. If so this will limit the stars available for alignment. You might like to disable this.
I normally do the mechanical alignment of the mount within using a calibrated Clinometer app to set the elevation, and using the dec circle to set the azimuth of the mount using a known bright star due east or west - set the dec axis and clamp it so it indicates the declination of the star, the adjust the azimuth of the mount and RA to align on the star - do NOT adjust the declination !
As reported at the end of the SynScan software alignment routine I’m regularly achieving this within 15 arc min of the pole.
During the SynScan alignment sequence it is important to align to the selected star accurately. The better you do this the better your GOTOs will be.
The errors in subsequent GOTOs will be no less than the inaccuracies you had in the alignment routine. So if you just eyeballed it with a finder, you can expect GOTO errors of a degree or more.
On mine I put a crosshair eyepiece in my main scope and align on the crosshairs to within 1-2 arc minutes. Subsequent GOTOs are good enough to put the scope within 15 arc min of the target - which I suspect is the resolution of the encoders in the mount. Important, because with a focal length of 3100 the field of view is only 30 arc min anyway.
Then last issue is cone error in the mount ie perpendicularity of the axes, and between the dec axis and the optical axis of the scope. If this is off by as much as a degree the GOTOs will be similarly inaccurate. It pays to measure and correct this, but that’s a whole story on its own.