It would take some doing to bend a crank-pin without other damage, so that seems unlikely.
Baker Gear on the 'Tich', I seem to recall – I'm not familiar with it but I know it's similar to Walschaerts.
I'm not sure if a displaced return crank would produce those effects but I suppose it's possible, by putting the valve-timing out enough to widen the lead or delay the exhaust enough to create the limping. Nor am I sure how the part can move, if it's on a square as conventional.
I would suggest, not running the engine on air; but instead moving it by hand on a couple of feet of rails, for testing problems like these.
Move the loco with the drain-cocks open, and see what odd happens in mid and full gears. There may be varying resistances as the pistons compress and rarefy the air in the cylinders, but these should be fairly easy to feel, and should be symmetrical.
If you can establish which is the faulty side, disconnect the eccentric rod from the return-crank and try again. If that alters the effects then yes, I would look to see if the return-crank has moved or come loose. I don't know how it's fitted on that loco, but if it uses a split-clamp method that may have slackened slightly.
'
If you can't find anything amiss with the valve-gear then we need think more drastic – that a wheel has moved on the axle. Whether that's possible depends how the wheels are fitted and if they have keys, but that would certainly create binding effects.
To test for that, disconnect both eccentric-rods and connecting-rods so the pistons and valves have no influence on things, then again push the loco slowly back and forth..