I've seen (and quoted) the same point differential ... but the important part if how good a teacher/coach/mentor the person it. Rating differential means nothing if the person is not good at a) accurately identifying YOUR weaknesses, b) correctly building a training program to remediate YOUR weaknesses, and c) doing all this in a manner that is constructive and supportive of the student.
I've had coaching from players rated from 1900 through 2600 ... the 2600 was the worst.
Who cares about rating difference? What a novice really needs is a good teacher. Put teaching skill ahead of chess skill, and the effects will be quite dramatic. Then, play someone 400-600 points higher and apply the new theories to that game to test your skill.
Well, the very best players have coaches weaker than themselves... It's not how good a player you are, but how good a coach you are.
I think the important thing is that BOTH the teacher and student are comfortable with the rating difference and should use common sense. The student ultimately needs to feel comfortable with the teacher and see improvement in his game. If it's a beginning tourney player rated 800 or 1000 looking to get to a class C level, I think that working with a 1700+ player is fine. In fact, many titled players have difficulty working with beginning and class players in my view. On the other hand, a 1600 player looking to achieve an expert rating or beyond should probably study with at least an expert. In the end, it's the effectiveness of the teacher that is most important. I studied with an IM who was not very helpful (as have many), so don't equate a title with effective chess instruction. However, I do feel that anyone below the mid 1700s USCF should not be teaching - even beginners, only because they're still making too many many mistakes in their own games at that level.
For the "G" Group's mentor program (feel free to join here: http://www.chess.com/groups/forumview/quotgquot-mentoring-program; you may have to join the group first), we have the minimum rating difference at 300 between the coach and the student. It's been working just fine, so I'd say 300 points is the ideal minimal differential.
Of course, if the mentor is good at teaching, that 300 points should evaporate pretty quickly...
G
I think the important thing is that BOTH the teacher and student are comfortable with the rating difference and should use common sense. The student ultimately needs to feel comfortable with the teacher and see improvement in his game. If it's a beginning tourney player rated 800 or 1000 looking to get to a class C level, I think that working with a 1700+ player is fine. In fact, many titled players have difficulty working with beginning and class players in my view. On the other hand, a 1600 player looking to achieve an expert rating or beyond should probably study with at least an expert. In the end, it's the effectiveness of the teacher that is most important. I studied with an IM who was not very helpful (as have many), so don't equate a title with effective chess instruction. However, I do feel that anyone below the mid 1700s USCF should not be teaching - even beginners, only because they're still making too many many mistakes in their own games at that level.
I strongly disagree with the notion that weaker players should not try to teach chess. IMO, the "safe rating distance" is anything above 0 (i.e both players =1200) however, for this to work the very nature of the relationship should change, they both can learn from each other.
Moreover, I also think that the idea of being taught "wrong concepts" is also inaccurate, we learn chess through experience and discovery, we would soon learn how our teacher was wrong.
Well I like the differing views on the subject. I remember some of the good and bad teachers I've had (non-chess and chess) ... the really good ones didn't have too many accolades or titles behind them. The best math teacher I ever had certainly didn't do his PhD in the subject!
Though isn't it fair to atleast correlate differences in "chess knowledge" with differences in Federation-based Chess ratings? I should think so.
A bad teacher, like bad information, that is trusted can really screw you up. Some people seem to never disentangle themselves from such falsehoods. Here, in these forums, there are even those who have read or were told the wrong basic point values for the pieces, or who can't grasp that pieces are used in context and the basic point values can't simply be applied mechanically to any position to figure out if they are "winning."
Someone a couple hundred points superior to a student, someone who is level-headed and can transmit a lesson, could be a suitable teacher. If the student can get more points and teaching skill in a teacher, all the better of course.
When I say "safe"", I mean enough to NOT learn horribly incorrect things or even worse, pick up/emulate really bad habits.
I've searched around these forums and seen a lot of people say "300-400" points is safe enough to guarantee that the student will get something substantial from the experience.
Though I'm sure almost all of us on chess.com are qualified to teach a complete beginner enough to break into a "novice" level federation rating ( 1000 ELO/USCF for example), but what then?