Power is the probability that a test rejects the null hypothesis when the effect is real. With few samples it is low: a test that stays quiet may simply have seen too little data, not a healthy system.