Any estimate of the number of accidental gun deaths caused by arming teachers can only be a positive number.
Ok, but exactly the same can be said to the contrary... the estimate of how many gunmen will get shot by teachers can only be a positive number.

What do you suppose is more frequent? Gun accidents (relating to trained people whose job it is to protect children)? Or school shootings?
Further, which do you suppose has the higher average death count?

It should be obvious that the lives saved by arming teachers will be significantly higher than any deaths caused by accidents, which will probably be something like one death a decade.

My sample size is not small. Ok there's only a handful of security guards at the Whitehouse, but you can take that further, you don't have to assume literally just the Whitehouse.

How many trained armed personnel accidentally shoot and kill people? Give me a yearly figure.