This is a somewhat technical post, so apologies if someone gets bored to death while reading it :P :P
...
The figure that you see in Image 1 is called as a trapezoidal curve. this is a rather simplistic depiction of the velocity-time curve in traction, and will help me highlight a rather important point related to traction energy consumption
consider a train, going...
more...

from station A to B which are separated by 20 Km. Assume that the loco in question takes about 2 Km of distance to reach to full speed, and one Km to brake from MPS to zero speed at station. That means, for a distance of 17 Km (85% of the train's distance covered) the train is operating at MPS. assuming that frictional losses are negligible, you'll realize that for most of the journey, the train is in the part of the curve that is shown between T2 and T3 in the figure. since by Newton's first law we know that things in motion tend to stay in motion :P, if you ignore the frictional losses you realize that no energy is required by the train in time duration between T2 and T3. In reality we do have friction and train resistance loss, but they are quite small when compared to the energy required in time duration between T=0 and T1. Upto time T1 train has to first overcome inertia, accelerate from rest and then keep accelerating till it reaches MPS. this is where bulk of the energy required for locomotion is used, and beyond T1 and especially after T2, very less energy is required. in fact, in 3 phase locos, during the last phase i.e. braking, energy consumption is negative i.e. you return energy back to grid.
Now, not all trains have non stop runs of 20 Kms. passenger trains will have halts every 5-10 Kms or so, and suburban trains i.e. EMU's have halts every Km or so. As you might have realized, these trains require substantially higher power to move the same amount of load as they spend most of their time in 0 to T1 phase, and barely have time to coast as their schedule often don't permit them to go for lazy braking.
on other hand, fright trains, once in motion, can easily travel long distances without stopping. even when halted for OT/crossing, the time spent waiting is so much that they usually turn of their blowers after some time, thereby having very small losses when stationary. hence, even though quite heavy, freights actually require lesser amount of energy to move the same amount of load due to their motion characteristics
this whole "energy required to move a specific load per unit distance" is called SEC or Specific Energy Consumption, and is a very useful metric to see how energy consumption varies across pass and freight sector.
in the Image 2, you can clearly see how the SEC values for freight are almost one third of their corresponding values in passenger operation. in other words, you could say that over a given distance, a passenger train of 1500 tonnes (a full length ICF rake weighs this much) will take same amount of energy as a freight train of 4500 tonnes. many freight, like container or empty BOXN/BCN rakes weight much less than that, so you need less energy overall to operate them than you need to operate passenger trains, hence in newly electrified sections where some TSS are still under construction, preference is usually given to freight trains rather passenger ones in the beginning.
...
Source:
Image 1: click here
Image 2: click here
...
further reading (for those interested): click here (this is from BL Theraja's book A textbook on electrical technology)
...
tagged it under JBP, for its their RF's who are wondering about this for quite some time now :P if anyone has any objection to that, please tell me, I'll remove it.