I wonder if a model like the following might be better:
1. Player A takes action.
2. DM introduces complication.
3. Player B takes action.
4. DM introduces complication.
The DM complications are independent of the player's action, and force a reaction from a player. If the player fails the reaction, the DM's side scores a success...
What you are describing is for action to action, micro-level stuff. But I find it interesting that this is very similar to the structure that Mouse Guard uses at the opposite extreme, going from one "adventure" to the next. Basically, after the players have taking a few shots at solving whatever they are trying to handle (win or lose), the DM has a "turn" to introduce complications. If they had a lot of success during the "player turn", they can negate and/or trade some of that in for stuff they want, but if they had a rough time, they may have needed to avoid injury by giving the DM the right to hit them harder on his "turn."
Basically, the structure is that you can try anything you want, but eventually you run out of "attempts". Running out of attempts lets the DM hit you mildly. But failing badly lets the DM hit you harder, while great success lets you avoid some of the later hits.
In something like a 4E skill challenge, drifted to this, I'd set up a couple of options for the players:
1. You can go for something "neutral" on your turn, which is basically a low-risk, low-reward option. You have to beat the normal DC by 5 to achieve a success in the challenge, but you have to fail it by 10 to get a failure. Usually, this doesn't accomplish anything, but it is better odds than sitting out your turn if nothing good presents itself. Crucially, if you make the normal check, you get to mitigate/cancel a mild complication. If you fail the normal check, the DM hits you with an extra one. (This would be exclusive with the success/fail results at the ends.)
2. If you go for a more normal check, ti's succeed on normal DC, fail on DC -5, and minor complication in the middle.
After everyone has given it their best shot in a round of checks, the DM gets to unleash the complications.
I may have the math off (talk about needing playtesting), but the idea here is that you can "succeed" on the main challenge but still pile up a bunch of complication that may affect you downstream. Or you can "fail" the challenge outright, of course but his is a lot less likely than in 4E standard. Or you can "fail" by building up some many complications that the players decide to abandon the challenge as more trouble than its worth or too hot to now handle. The normal SC fail state can now be "hard fail" with serious consequences, because it is relatively rare. The usual failure is now that party works through the challenge, getting the XP, but now has to deal with the side effects. In this case, I would not give XP for a hard fail.