The Xbox Series uses Display Stream Compression (DSC), a lossless way of obtaining its HDMI 40gbps datarate. PlayStation 5 achieves an HDMI datarate of 32gbs, which is achieved by using a Fixed Link Rate (FRL). The Xbox achieves a better fidelity image (full 4:4:4 vs 4:2:2 Chroma Subsampling on the PS5), but it seems 2020 AVR systems have a hardware incompatibility and are unable to resolve the signal from devices that output a full DSC signal, in this case the Xbox Series).
In my experience, using ARC introduces audio latency. You can use optical output from your TV and send it to your AVR. It won't support the lossless surround formats that a direct HDMI connection to your AVR can provide but it will have the same capabilities that ARC has (with no audio latency to boot).
Yep. To elaborate: ARC (not eARC) is practically identical to optical in regards to audio support (uncompressed stereo or compressed surround sound), albeit with the additional downside of introducing extra audio latency.
eARC, while compatible with additional audio formats, still suffers from audio latency issues as you mentioned. Plus, not all TVs are able to export all audio formats via eARC either (e.g. the LG CX is unable to output DTS:X via eARC; this is due to the limited EDIF size, the metadata that accompanies an HDMI signal).
So to conclude: I would usually recommend connecting your console directly to your AVR system... But with the current state of 2020 AVR systems, you will have to either sacrifice something in the process whether you connect the system directly to your display or to the AVR system. So unfortunately, there is no correct answer (other than "wait to see what 2021 offers").