What does it mean to be a good teacher?

It's back to school time. A little later in England than in the United States. A week into the new term I ask my 13-year-old daughter how things are going.

She says she mostly has good teachers this academic year. What's the difference between a good and a bad teacher, I ask.
 
Her reply goes something along the following lines. If you're in a class with some difficult students, then you want a teacher who's quite strict. But if you're in a class with students who all want to learn, you don’t want a strict teacher; you want one who will let you express yourself a little, and try new things out to let you learn.
 
And then she gives me a cameo of a teacher who is strict in a class full of good students, and its negative consequences. (She assumes I know what happens when a lax teacher is in charge of a group of troublesome students.)
 
She goes to a school whose students exhibit the typical range of behaviours. I know this because, as a favour, the Social Research Unit surveyed the entire school body using Goodman’s SDQ. (The school’s leaders didn’t like the results because they reckoned the students were happier and better behaved and more pro-social than they actually were, so the results never saw the light of day).
 
According to OFSTED’s very doubtful methods of assessment, the school is satisfactory (third on a grade that extends from 1 ‘outstanding’ to 4 ‘inadequate’, with a good capacity for sustained improvement. My layperson’s assessment – more dubious than OFSTED's – is that it's an average school that thinks it is better than it is, and that it serves kids whose parents, like me, think they are better than they are.
 
A more objective assessment is that it's a school I'm happy for my children to attend. It's adequate, and they can walk there and back in safety.
 
My 13-year-old daughter is not a prevention scientist. But if her hunch is right, then perhaps we could do more to manage classroom mix. I remain convinced that Tom Dishion should get a Nobel prize for his discovery that placing groups of anti-social children together is a very effective way to increase anti-social behaviour (probably the least acted upon, reliable finding in the history of children’s services) .
 
But what proportion of anti-social kids can a class support? And is there a way of finding an optimal mix between teacher competencies and the ratio of anti-social kids in a class?