I cant open what ever that is you're referring to, but YES Sorry I got it all backwards as soon as you raised it! Mixed up in my head between the two systems - Piston valve addmission is centre, slide valve is side. swap what I just said, bottom line, both systems are opposite, so the eccentrics will be opposite if the radius rods are to set the same way on both chassis'.
I could easily concieve how an engine like this could have 3 eccentrics facing one way, and only one facing the other, it would entirely depend on which way the cams pivoted relative to the lifting rods...dono why you would, but yes can be done. Its possible....and I'm guessing here, but the actuator for the valve gear is under the engineer's side tank, just in front of the cab. It pushes and pulls the reversing rod down one side of the loco to the lifting rods via some pivoting 'L' shaped cams...you can see them there on the model...when all the cams rock in the same direction with the valve change, then the eccentrics will always face the same way relative to main rod (on an all piston valve loco), but as seen in the photos, when one set of cams rocks down (lowering the radius rod), while the rear set rocks upward , raising the radius rod, then in order for both engine units to run in the same direction, the eccentric on both engine units will be opposite. Now on this crazy 3 matching, 1 different thing, maybe only one side has the cams rocking in opposite directions due to some space constraint where the actualtor takes up space where the cam would otherwise rock...this would mean the radius rod on one side of the loco lowers, while on the opposite side of the same chassis its lifted to the exact same amount, and the eccentric on that both sides would need to face in opposite directions. I cant believe someone would engineer that, but I dont know modern locos that well, who knows...check the photos I guess to see if the radius rods on both side of the loco lowered or lifted the same on all sides or not.
The one way to really know what should happen with this model, is to look at the connection of the lifting rods to the radius rods on the real engine, and see how that lifting rod is connected to that pivoting 'L' shaped cam. If you can see that the cams can pivot in the exact same way on both chassis, on both sides of the loco, then all the eccentrics should face the same way (forward of the main rod crank). If the cam look like they would pivot in opposite directions on the two chassis relative to each other when the Johnsonbar is moved forward etc, then reverse the eccentrics on one chassis - have the forward chassis with eccentrics facing forward, and the rear chassis eccentrics facing aft, which seems to be visible on several of those WA Mallets I can see. If you can check some photos of the real locos in motion (ie with radius rod out of neutral), you will see the radius rod on the front chassis down, and the radius rod on the rear chassis up...if you could see both sides of the loco from the same period in motion, see if the same condition is on both sides of the loco, then you'll know whether a 3-1 thing was done.
I'd just set the eccentrics to make sense with the lifting rods on this model and you'll be logical.