I have been reading multiple tutorials on lift and gain charts here,here,here and here.
While I understand how the curve is drawn based on our model, but I don't know how the dotted black lines for random model or dotted blue lines for best model is drawn.
I used the Python kds library to draw this line (as I do not know how to plot this manually)
For ex:
Can any of you help me with a toy data to explain why the lines are plotted the way it is from random and best/wizard model?
I have never seen these kinds of plots before, but here's the idea.
The "random" line is representing the results you would get if you did completely random guessing - so this is the "baseline" result if you used NO model. The farther your line is away from this, the more value it's providing. The closer your model's line is, the closer your model is to just being valueless random guessing.
The "wizard" line represents the best possible result attainable by a model (as far I can tell by this article). So, the closer your line is to this, the better. The "wizard" line is there just to give you a reference point about how good a model could possibly be (basically, how much uncaptured accuracy there is). It doesn't mean that the result is necessarily practically obtainable (it probably isn't).
Basically, the distance between these lines represents the range between the worst possible and theoretically best possible outcomes/results.
