Let's put aside for a moment the optics of a school teacher having their class play Poker (even though obviously they wouldn't be using money). I can imagine a lesson plan going something like:
1. Have kids play some individual games with each other.
2. Stop the games, and go over some of the actual math concepts behind Poker. At this point, the kids are engaged and will want to learn better strategies.
3. Run through a game as a group, with the teacher asking the class what they should do each turn. Ask students to explain why they think one move is better than another, get brief discussions going where applicable, and write probabilities on the board.
4. Let the students play another round of individual games to apply what they've learned.
"Stop the games", "put the game in context", etc. mostly kills this feeling for many kids. Only the most skilled teachers are able to do this without making the whole thing unfun.
Sure, if the "kids" at hand are university students, you could use poker. But I find it extremely dangerous to teach kids poker theory, who as theorized by the Dunning-Kruger effect might start believing themselves to be really good at it, and then proceed to spend a lot of time and money on a pursuit that might not lead them to happiness. You can teach statistics without making any new gambling addicts in this world.
I know the majority of the students wouldn't be affected and the effect could be even positive, but for some the drive to gamble is so strong that they really go off if that predisposition is nurtured. Although perhaps they are ticking time bombs either way, so no amount of protection will save them. But yeah, you got to be careful teaching kids positive feedback loops on things that might have serious drawbacks.
Btw, the kids who are worse at it would probably get their ass handed to them in the class games already anyway. Even after some training. So that's a good corrective and immediate feedback.
