I have studied that according to Aufbau rule the energy of subshells is dependent on the sum of $ n $ and $l$ values. This would imply that the energy of subshells in a shell varies as 

$$ ns \lt np \lt nd \lt nf  $$

and this makes sense(intuitively from Bohr model) too as the closer the electron is to the nucleus the lesser energy it would have. 
And this idea corresponds to penetration power as well where the trend is 
$$ nf \lt nd \lt np \lt ns  $$
as farther the subshell is from nucleus lesser it experiences its force and hence lesser penetration power. 
Everything made sense until I saw the graphs of Radial Probability Distribution. [![enter image description here][1]][1]


  [1]: https://i.sstatic.net/eNJ1r.png
and also the expression for average radius of subshell
 $$ 1/2(n^2 + l(l+1)) $$
According to these both my conclusions about penetration power and energy levels should be wrong as for greater $l$ value average radius and even most probable radius both are smaller than corresponding values of smaller $l$ values. What is wrong with my reasoning? Kindly Guide.