I think this is a bad rule; firstly because it is easily established that the 'saving throw' is now, in most instances, just a duration determinant. Secondly, because it is something that really should scale with level. Your thunderwave hits a peasant and an epic fighter, doing 2 points damage to each and knocking them off the cliff. Both have the same 55% chance of saving themselves from going over. Why? Doesn't make sense.
To be fair, they don't have the same chance of falling off the cliff - since the peasant is just about guaranteed to be hit by the Wizard, while the Fighter is likely to shrug off the Thunderwave. Even if they do get hit, the Fighter may well have a variety of magic items to help them if they do get knocked off the cliff, while the peasant will simply die.
But there is some rational behind making the roll to avoid falling identical all around - it isn't based on your spell, but on the ability of either fellow to grab ahold of the edge of the cliff. (Though this also seems like the epic level fighter should be better at.)
In the end, the save mechanic is a somewhat odd fit here - it seems like they simply grabbed a useful mechanic to fit this one specific circumstance. They wanted some way to prevent push effects from being instant-kills, without having to work out a complicated sub-system involving athletics and acrobatics (which would inevitably result in some characters always going over the edge, and others never doing so.)
I think, as far as game mechanics go, the system they have in place (for avoiding being pushed off ledges) works - but conceptually, it is a little weird. I think making small modifiers to it - bonuses for Athletics or Acrobatics, for example - is both reasonable and balanced. I think fully overhauling it with a different system would probably be a lot of work for very little gain, but I don't think it would unbalance 4E all that significantly.