I used to think that
ssh -X me@some.box
"just bloody worked". However this might not work - ssh must play ball on both sides of the link. On the remote (ssh server, X client) sshd must sit behind some port, tell Xlib to send X11 requests to it and then forward them back to you the X server (where the ssh client is). If the remote box is locked down to prevent this, you will get:
X11 forwarding request failed on channel 0
as part of an otherwise working login. As it happens, I am the admin of the remote box in question, so I followed the ArchWiki and went to /etc/ssh/sshd_config and uncommented
X11Forwarding yes
X11DisplayOffset 10
X11UseLocalhost yes
This guy, got the same error message for a completely different reason, so you might want to look into that too.
The more likely cause of trouble is that some clients will try to do things that X thinks are dangerous and end up crashing because of BadRequest. Now is the time to sit on your hands. Do you really want to use that program? Do you really trust the box that it is running on? If the answer to both of those questions is "yes", then the best thing is to use
ssh -Y me@trusted.box
But only on those occasions when you need it. Otherwise, keep using -X or (better!) nothing a all. The rule here is:
ssh -- a remote trojan can find exploit some hole your ssh client and take over your local account.
ssh -X -- as above, but it can also exploit your X server and thus root your box.
ssh -Y -- as above, but it can also (without needing to find exploits), take screenshots and keylog you.
Have fun.