I purchased one brand-new for big bucks in 1997. As others have said, mechanically they are very robust. Good casting and assembly quality. Detail equal to or better than Lionel's 1990 1-700e. Unfortunately like the 700e, there are a lot of "bright" surfaces for what was marketed as a scale model. If you're not a collector you could probably paint or chemically blacken the offending surfaces, but even doing that takes effort and skill. MTH toned this down on the PS2 versions a few years later.
The 20-3020-1 was a purely conventional loco. Its gear ratio is 16:1, which is pretty tall when combined with drivers that scale out to 78". Unlike Lionel, MTH used a single-threaded steel worm and a 16-tooth bronze worm wheel on the axle. The worm wheel is quite smaller in diameter compared to the one Lionel used in its 700-series Hudsons. The "pro" is that the MTH model doesn't have a huge gearbox sticking up above the middle driver. The "con" is that the smaller gear equates to a shorter lever arm for the worm, and lower starting torque at the wheels.
Although the gearbox is much smaller than Lionel's, when viewed from the side, the boiler shell is still pretty obviously U-shaped, so it looks best when viewed from above. This loco would have benefitted from a lower casting to round out the boiler, as MTH used on exactly one run of its Premier Berkshires circa 2000. [In truth there's a lot of daylight visible above the drivers of a Van Sweringen Berk!] Lionel used a two-piece casting to get a full, round boiler on their Milwaukee S3 and Legacy Reading T1s.
With the stock 9434 12V Pittman motor, my J-1e could exceed 100 scale MPH at full throttle. On a test loop of O72 tubular, my slowest minimum speed was about 9 scale mph with some slowing noted on O72 curves. Rubber tires on both sides of the rear axle, fairly wide gauging of the drive wheels, and one-piece side rods without bushings make this a "tight" loco. It says O42 on the box, but I doubt that it would run realistically into O42 circular curves, particularly when running light (i.e., NOT pulling a train.) IMO, if your curves are that sharp, you're better off with the metal-tired scale Hudsons from Lionel, or more "traditionally-sized" motive power.
After a year, I decided that I didn't need 100 mph of top speed, and I wanted better low-speed performance. With a cast-in gearbox there was no hope of changing the gear ratio, so I substituted a 24-volt version of the same Pittman motor. This is an easy-to-do modification that really mellowed out the loco's performance; if you're going to stay conventional I highly recommend it. I can still get about 65 mph of top speed with 18-19 volts on the rails. Current draw is greatly reduced with the 24-volt motor. Starting a long train with the slack stretched is smoother too. Unfortunately the slowest minimum speed didn't improve by much. After a lot of testing I concluded that none of these motors want to keep turning much below 800 rpm in this application, when being fed rectified DC from the Proto-1 electronics. I'm sure that some flavor of closed-loop speed control would give me a minimum speed of 3 mph or less.
Pulling power is excellent, although I did have two instances when the loco shed a rubber tire. In one case the tire got tangled in the side rod; thankfully I noticed it before serious damage was done.
My biggest source of dissatisfaction with the loco is the lack of coasting. For a long time I ran with others on public displays. Despite having a (small) flywheel, the MTH Hudson's tall gearing and self-locking worm drive made for a rod-graunching, grinding halt any time someone caused a short circuit, or accidentally turned off the track power. These things happen, and the "emergency stops" were scary to see and hear. Rubber tires are partially to blame for this too, because as 2-rail O scalers know, smooth wheels skid a little, even with the gears locked up.
A minor, related problem is that the Hudson has only two rollers on the loco and none on the tender. It would occasionally stop abruptly on switches, crossings, etc., because both rollers landed on dead spots, and there wasn't enough coasting, capacitance, etc., to reach powered rail. I looked into adding a roller to the tender, which I'm sure would help. It's doable, but would take some original engineering, trial-and-error. Shame on MTH for not offering an upgrade kit, or building it with 3 rollers in the first place!!
Re: coasting, the flywheel is nestled just above the front of the ashpan, and there isn't much room for a larger one. I thought about milling out the boiler shell, but it would be visible. I even tried having a custom flywheel "slugged" with tungsten rods inserted near its rim (think bullets in the cylinder of a revolver) by a local machine shop here in TX. After a few months of silence I checked back with them only to learn that they somehow "lost" my materials . Although mechanically identical, PS2 versions that appeared a few years later do exhibit some degree of coasting. This amazes me! (Unfortunately the PS3 versions don't coast, at least in conventional! )
Bottom line, I guess now these are a bargain, although it would take money and effort to add the features everyone expects in today's top-end locos.
I don't recommend it if you want coasting or smooth, single-digit speeds on curves tighter than O54. PS2 versions have blackened bright-work, coast better, and should crawl right out of the box--if they run. PS2 electronics are a different kind of crap-shoot. Personally, if mine were lost to fire, theft, etc. I would replace it with a Vision Line. Back-drivable gears and OEM Legacy speed control make it worth saving up for IMO. My $.02.