Well.....
When I first ventured into the abyss that is garden railroading, I have had to do it with a garden that had a complex gradient but rose about 3ft at its peak at the back of the lawn/beds area.
I wanted big wide radius curves for the main line to facilitate the longer American locos and rolling stock that I wanted to run and the line would not be able to cut through the middle of the lawn...on penalty of death.... I also did not have the room to build a wide curve helix.
There was a two and a half foot wall that was between the patio and the lowest level of the lawn, so I could not use trestle bridges to bring the low level lawn to the same level as the apex, otherwise the conservatory would have had an over 5 foot barrier running across between it and the lawn!
I also did not construct any deep cuttings at the apex of the lawn due partly, to the fact that the roots of a large tree at the apex would have caused problems but also that it would have meant that the trains would have had to be running out of sight in a cutting, about 2 to 3 foot deep at maximum, for about 40 to 50 feet. Also the cutting would have been a huge collector of leaf little etc...
So.... I decided to just build the line with the gradients in place.........
Obviously I tested it as I built it...but I didn't test it very well... I could see that my newly bought steam 2-6-0 mogul LGB loco was going around the layout with maximum 1 in 10 or 1:10 gradients, in a perfectly respectable manner....... bad move... I didn't test it, at first, with any rolling stock attached...imagine my disappointment when I saw that the little loco could just about pull just one box car as well as its own tender!
So... I had to overcome the tortuous gradients with grunt. This came in three forms:
1. first the introduction of early era diesels to my, at first designation, only steam era layout...these had twin powerful motor blocks with traction tyres and where an immediate improvement.
2. secondly the use of multi heading and/or helper locos in a train... again a great improvement and it meant that even the steam outline locos could now haul longer trains without problems
3. fitting 'invisibly motorised' bogies into boxcars and passenger cars...this meant that one steam loco could look like it was overcoming the gradient hauling a train on its own, when in fact there would be two or three 'motorised' wagons/cars in the rake behind it.
4. finding the joy of geared locos like shays and climaxes... the shays just literally pulled their weight and did what they did in their prototype form.... they didn't need multi-heading except for the times when I wanted to see them 'lashed-up' behind each other!
One drawback was that, going up was fine with enough 'grunt' but coming down could have its own problems with the force of a heavy train on couplers etc sometimes forcing derailments.
This was largely overcome, when multi-heading, by having the slowest 'helper' loco at or near the back of a train to 'restrain' the train on the way down.
Also making sure that I standardised the couplers on a train helped. I used aristocraft knuckles for mainline freight and streamline passenger as they are strong and resist buckling well. Bachmann knuckles for my shorter 'local line' trains were also fine.
LGB 'hook and loop' were okay as long as there were hooks at both ends on each car/wagon.
Live stream came to the layout about 9 years after the initial start of the build, and I found that, again, geared locos were the way to go. Excellent performance and great plumes of steam as they toiled up the hill!
The layout does have the appearance of the Rockies and Utah about, it so it does look okay especially with the trains winding around tortuous cliffs etc.
So the upshot, in answer to the question, is that you can build layouts with heavy gradients but you then have to do things to enable trains to run on them and to overcome the problems that gradients pose going down as well as up.
Would I do it again without levelling the garden.... hmmm, I'll have to come back to you on that one!
EDIT
Yeah, I forgot to add, when I first wrote the post, about adding weight to locos.
It really works
BUT be careful with any that have traction tyres.
Locos without the tyres will, even with sensible weight added, slip their wheels when a gradient outweighs the tractive force ( yes, before I am corrected, I know that is not the correct scientific terminology)
BUT the axles with traction tyres will not slip until there is probably too much force going through the gearing system or connecting rods or both, so bad things can happen......
I haven't had it happen yet but there must be times on my layout when things have been close...
; it shouldn't work, but it does. Five years on I've probably re-laid 1/3rd of my 90ft line, smoothing curves and yes, reducing gradients. You learn as you go; get some track down, try to avoid the biggest errors and then see how you go. Get something running, and then try to get it running better. It's a hobby: enjoy the challenges!