Yes there are 3 criteria, it is not enough that a system is 'sensitive to initial conditions'
But warning there is no general theory for chaotic behaviour in n body systems each one needs to be analysed. You may like to start with the work of Poincare on the stability of the solar system and all the developments from that.