The percent grade of a track is: 100x[rise]/[run]. The number of cars a locomotive can pull decreases with increasing grade and at 25% no locomotiveprototype or model (without traction tires)can pull even itself up the grade. Grades on prototype railroad main lines rarely exceed 2.2% but logging and mining railroads may have grades up to 10%. The G gauge Wasatch Mountain Railway has a maximum grade of 7.2% which is close to the 7.5% of the prototype Uintah Railway which inspired it.
The number of cars a locomotive can pull is proportional to the weight on its drivers. Any weight not supported by the drivers is dead weight which will limit the number of cars it can pull up a grade.
The Shay is ideal for the steep grades on the Wasatch Mountain because its entire weight, including its internal battery, is supported by the drivers. However, operating time with the internal battery is limited and it is usually supplemented by an auxiliary battery in a trailing car which decreases the number of cars it can pull.
The Mallet battery and electronics are in a separate car pulled by the locomotive and it is quite heavy. Initially, the Mallet could not pull the car up the grades! Both of the unpowered leading and trailing trucks of the Mallet contained heavy lead weights and spring-loaded bullets intended to keep the trucks on the rails. The bullets pushed down on the trucks, lifting the drivers, and the lead was just dead weight that had to be pulled up the grade. I removed the bullets and relocated most of the lead to inside the boiler, which increased the weight on the drivers. The trucks tracked just fine.
SHAY: Weighs 9.9 pounds (4.5 kg); all weight on drivers; no traction tires; can pull at least 11 cars. The auxiliary battery, contained in a separate car, reduces the number of cars that can be pulled.
MALLET: Weighs 9.1 pounds (4.1 kg); battery/electronics car weighs 5.2 pounds (2.344kg); locomotive has a traction tire on one wheel; can pull maximum 11 cars plus the battery car.
All cars are the same weight as purchased but "loads" are kept as light as possible: "logs" for the flat car are light-weight foam and coal loads are balsa wood with a thin layer of coal glued to the top.
The guests watched and wondered as the Mallet slowed down and stopped on the grade; the driving wheels spinning. I took off a car and the train pulled the hill but the next time around the layout it stalled again. This time the guests had questions but I didnt have answersmaybe some lubrication from my new locomotive had leaked onto the rails. I took off another car, bringing the train down to five cars but by the time the guests left the Mallet was struggling again.
That wouldnt do for my next open house, coming in a week, so I cleaned the rails and all the locomotive and car wheels with lacquer thinner and lightly polished the rails with fine grit sandpaper to remove any traces of solvent. Then the test and an unpleasant surprisenow the Mallet could pull only four cars and after one loop around the layout it stalled on the grade, wheels spinning! My wife says: What do the real railroads do about slipping wheels? Sand! I replied. So I found some powdered limestone in the shed and sprinkled it on the rails, brushing off the excess so there was only a very thin dust layer on the rails. I was back up to seven cars with no sign of slip.
An obvious question was: Why did the Mallet pull seven cars initially, then slowly lose adhesion as the day wore on and after the track was thoroughly cleaned? I believe the answer is that there was sand on the rails to begin with and it slowly wore off as the train ran around the track. The initial sand was limestone dust that accumulated on the rails as I re-ballasted my track, mud that splashed onto the track during rain storms and atmospheric dust that had settled on the track over a couple of weeks. Yup, sanding the rails really does work, even on model railroads.