I went with option 2. In retrospect this might not have been the best idea, since I'm actually thinking entirely different things about the game... but whatever.
Here's my stance. Skullgirls needs the average combo to be shorter and less damaging than it currently is in MDE. That's my sincere belief, and as far as this goes the current beta patch is a great step forward. It gives the defending player more chances compared to the two-touch kills running rampant in MDE.
But the problems of the game run a little deeper than that, and Undizzy is an opaque, tacked-on system that renders you unable to exactly gauge how far you can go after the reset. Now, there's two ways to tackle this issue - one, audio and/or visual cues that give you a rough estimate of how much Undizzy your enemy has left once you tag him again. Steps of 40 or 50 should be good enough for that.
Or, two, you axe Undizzy in its current form and try to cover what it covered through existing, perhaps also new systems. In short, you somehow find a way to limit or weaken conversion opportunities. Since, in Skullgirls, so many starters lead into full combos that the tools that don't become severely devalued, and because everything kind of leads to the same combo it also does roughly the same damage. There's no real reward for punishing blocked UmeUpdos, there's no real difference between high/low/crossup/noncrossup (throws are the exception to this right now and look at how many people are outraged by this...) You can - and should - read IsaVulpes' threads, he's written far more and far more eloquently about this topic.
--
So, whatcha gonna do? I'm not the most qualified person, but let me please say what I'd want. Warning, crazy idea ahead, consisting of three concepts aimed at creating a more interesting spread of combo damage.
One, a redone damage scaling system. Let's say the scaling goes roughly like this:
1st hit 130%
2nd hit 100%
3rd hit 75%
4th hit 55%
5th hit 45%
6th hit 35%
7th hit 30%
8th hit 25%
9+th hit 20%
With throw, assist, DHC scalings and such all kept at their current levels. Suddenly, the starter matters much more. In certain cases (Filia IAD j.HP cough cough), this could be abusive, but you could fix these cases individually. (j.HP hitting twice for the same damage in total, anyone? With basically no blockstop on the first hit?)
Second: In conjunction with that, adjust Undizzy to reduce combo length even more compared to the current beta build, so that you'd have a damage variance of 3.5~4k meterless off bad starters, like Throw, Assist or multi-hitting things to 5.5~6k for optimal starters, like a s.HP punish or whatever. However, Undizzy would instantly decay to 0 - it'd be no longer needed to enforce damage variance. You want to do more damage? Reset into something beefy!
Third: Buff the damage supers do by maybe 20% or so, to ensure you still can pack a punch if you spend the bar.
---
That's been it, thank you for reading.
Tl;dr: If you complain about your combo getting worse through the changes: That's the entire damn point of the beta. Your call to be angry or pleased at that, but let me give you two pieces of useful information: One. If Mike Z is being lenient this time around, I foresee this same discussion about combo damage about four months from now, like it already happened twice. Two. Most games mainly remembered for their combo game are kusoges like Hokuto no Ken. CVS2 had ridiculous A-Groove custom combos, KOF13 has HD mode bullshit - but that's not really the core gameplay of either game. An important element, sure, but not the central one like it's sorta become in Skullgirls. Or has been in Hokuto no Ken.
...
As Number 13 above stated, this is my ideal Skullgirls, or at least a lot closer. It doesn't have to be yours, and I (try to) respect that (since I really like figuring out beefy combos, as well...) - but please give it and my mindset, my reasoning a consideration. And please please tell me should I be really wrong.