Good drafts and bad drafts have a few thing you want to look at:
(1) The skills of the people involved. Wisp might be a terrorist in the hands of a professional player, but no one I know can play him well, so if I drafted him in my stack that would be a bad move. Even if a hero or combo is theoretically good, if the people you are playing with can't hack it then you are just setting yourself up for failure. Same thing goes with banning. You have to weigh the likelihood of another player on the other team being able to use any particular hero to good effect. If you're not in a very high bracket, the chances are probably pretty slim that an opponent would be able to do more with Wisp than any other support. Now, heroes like Warlock and Death Prophet are a different can of potatoes. They can have huge impacts on a game and require relatively little skill to accomplish it. Even if you play horribly, a rock in your face or 25 spirits floating around is something to be concerned about. If you're in a stack, make sure that your best players are in roles that give them the most game impact (carry and mid, primarily).
(2) Win your lanes. Your wombo combo isn't going to do anything if you lose mid, your offlane feeds, and your safe lane gets disrupted. Whatever your other strategies are, its essential to have a shot at winning or at least breaking even during the laning stage. If your mid doesn't bottle crow like a pro, its really dangerous to put a hero like DK there against a Viper, Invoker, QoP, etc. If the enemy picked Dark Seer, don't leave a melee in the safe lane while you try to offensive tri. If you're planning on running a defensive tri, winning the lane means getting kills, so don't pick supports that have a hard time doing that.
(3) Give yourself leeway. If your relying entirely on one aspect to win the game (ganks, pushing, deathballing, etc.) and it fails, you automatically lose when it doesn't work completely and win you the game. I'm sure you've all been in games where it feels like you have to constantly do everything right to win, and as soon as you slip up once you're finished. You don't want to be in that position. A lot of times it happens because your team is really good at one thing, but but bad at everything else. You can push towers really well as a 5 man, but you can't actually control the map because individually you're not that scary. You can gank well, but have no pushing power so you get turtled to death. You have an amazing teamfight wombo combo, but do nothing once the enemy gets a few BKBs, that sort of thing. For an illustration, look at game 1 of the TI3 finals. Na'Vi was able to destroy all of Alliance's outer towers easily within the first 15-20 minutes, but they still ended up losing badly because their lineup was a complete gambit that was super strong in the first 20 minutes, but sucked afterwards. They had zero leeway in their composition. It was throne it by 20 minutes or lose the game.
(4) Psychology matters. If your teamlooksscary, there's a good chance it will be scary. If you've followed dota for any substantial period of time, I'm sure you've noticed that heroes go from zero to OP overnight without actually changing much or at all. A lot of that is just people becoming afraid and playing scared or with a defeatist attitude when they encounter the hero. If your team is stacked with "OP" heroes, and theirs isn't, dissension will set in immediately on the other team and the "inevitable" loss will be a self-fulfilling prophecy. Conversely, if you draft a bunch of dark horses, people will immediately be skeptical of what your doing, start flaming, and not play as well as they otherwise would if they believed they could win from the outset.
(5) Make sure you can initiate. If every time you try to do something the enemy team hard backs and you find yourself asking why you can't get anything done, it's probably because you can't initiate. Basically, you can initiate in two ways: (A) You can force fights. You can do this by having a good push lineup or by doing Roshan. In most situations the other team will feel compelled to respond and will be forced into a team fight. Often times they will jump on you, so make sure you have at least one hero that can stand back and counter initiate. The idea is that you choose when to fight, so hopefully you will pick a time that's advantageous to you. (B) You can pick fights. To do this you need heroes that can move quickly and stop fleeing people in their tracks. I'm talking blink dagger carriers and long ranged stunners. You use these assets to get the jump on people when they are at unawares or isolated from their team, allowing you to push down objectives at a numerical advantage or pick off people one-by-one as they come to help the target you jumped on. If you can't start fights through either of these methods, you have a serious problem and are probably going to lose.