Dan you didn't say which locos they are, or how many cars you're trying to pull. If the loco is a 1655, 1666, or 2026, some of the difference could be due to the greater friction from sliding shoes vs. roller pickups. If the tension isn't correct the shoes can actually lift the driving wheels off the track!! Another possibility is the tender. Just like a real railroad, you have to think in terms of NET pulling power at the tender coupler. A lot of post-1950 tenders were lighter than their 1940s counterparts, leaving less reserve pulling power for the train.
Adding weight will help. Lionel did exactly that with 726 Berkshire beginning with the 1948 model. The prewar 763E had a hefty weight above the gearbox. The 773 got Magne-traction instead. Don't overdo it. Weight adds wear to the bearings, and you still want the driving wheels to spin. A little wheelspin when starting is prototypical, and if the loco stalls when overloaded you can burn out the motor.
I believe that sintered iron wheels get a better grip on the track than the steel tires used on 1940s locos. If the wheels aren't worn smooth, a non-magnetized loco from the 1950s may still pull more than its 1940s counterpart. A much fairer test would be a 2037 vs. a 2018, or a 2046 vs. a 2056 with identical tenders in each case. I would bet the added pull from Magne-traction is at most one or two freight cars. It mostly depends on how clean the wheels are, and how worn they are too.
I'm a big fan of MPC and newer cars with fast angle wheels. I had an O27 layout for years. Pretty much any die-cast, 6-drivered loco I tested could pull enough MPC cars to chase its own tail around the 4x8. What are you trying to pull? When was the last time you lubricated your car axles? And if you have a 110 trestle set, all bets are off. Nothing realistic about it, that's roller-coaster territory!!
