Consider the TD prediction algorithm for policy evaluation and assume that the TD target is calculated using the second largest in an environment with at least $2$ actions. Is it possible that this process is part of an $"$on$-$policy" evaluation?

Consider the TD prediction algorithm for policy evaluation and assume that the TD target is calculated using the second largest in an environment with at least $2$ actions. Is it possible that this process is part of an $"$on$-$policy" evaluation?

True

False