$\lambda$ refers to
Eligibility traces provides a way of choosing between MC and TD methods
Use Eligibility traces if a task is partially non Markov, but you want to use TD for their advantages
By adjusting lambda the e-trace can be placed anywhere from MC to one step TD.
Methods using e-traces require more computation than one-step, but offer faster learning.
Makes sense when data is scarce
The eligibility trace is a fairly simple mechanism which just keeps track of previous states visited by the learning method. T
