As noted, I've been thinking about ways to add magic to the Leverage system. It's actually quite easy, especially if you're going to be adding it from scratch. It's a bit rougher when you are trying to emulate a specific magic system, but that's always the way of it.
This has lead to a bit of mechanical experimentation, and one particular bit of sausage-making struck me as something useful to talk about. There's a particular approach to fiddling around with Cortex Plus that is intellectually very appealing, but which produces the kind of results that look cool on paper, but end up leaden at the table.
Consider, if you would, Button Men. For the unfamiliar, button men is a simple dice fighting game from Cheapass games. The basic mechanics are very simple and easily adaptable to many situations (I've run entire wargames using them), and at their heart they boil down to "Each side rolls a bunch of dice. On your turn, you can remove another players dice if one of yours shows a higher value, or if you can add up dice to equal a value they show. After a 'capture', reroll any dice involved in the capture" Dice are scored based on their size, so you run the gamut from d4 to d30, and you can really throw down with almost any 5 dice. As time went on, special dice were introduced into the mix, with extra rules, and that's where things get curious.
Button Men is interestingly informative for Cortex Plus design because the dice tricks it includes are designed for a range of die values rolling against a competing pool. It seems very natural to bring those ideas over into a Cortex Plus roll by introducing ideas like "Knock a die out of the opposing pool", "Force an opponent to reroll a die" or "Reroll one of your dice". You can build some pretty cool stuff with this, stuff that allows for some sophisticated interplay between opposing rolls.
But whatever you do, don't.
This is one of those cases where the fact that you can do something doesn't mean you should. At present, resolving Cortex rolls is pretty quick - determining success is very fast, and spotting any complications or opportunities is only slightly slower. There may be some delay _before_ the roll, as decisions get made regarding what dice get brought in, but by and large that is a fun kind of delay because, to sound a little wonky, players are usually engaging the fiction (which is in turn represented by dice). The inertia of play keeps moving (or can keep moving - it's possible to bog down if people get too bonus-obsessed, but Leverage doesn't particularly reward that).
In contrast, playing dice games after the roll is a total show stopper. Yes, some players can glance at the dice and near-instantly make all the decisions necessary to move forward, but they are very much the exception. Play is more likely to stop as players consider the smartest option. This speaks directly to Linneaus's first principle of dice game design: Downtime is the enemy. No one want sot be left sitting there at the table while a player decides which of two options is marginally superior.
In broad game terms, this is something you avoid by making choices simple or obvious. This seems counter-intuitive, at least in part because we don't want to neuter the player's choices, so let me unpack this a bit. When we're talking pure game decisions, we want to minimize uncertainty. To use 4e as an example, players have a lot of tactical choices, but the only real uncertainty is whether they'll hit or not, and that applies equally to all options. A player may need to make a decision between burning a daily or encounter power in a given attack, and while this may lead to some indecision, that indecision is not rooted the player not understanding the outcome.
In contrast, let's imagine a Leverage rule that let's you force a reroll in an opposing die. If that die came up showing it's max, then it's no real choice, but what if it just rolled ok? What if forcing this reroll means you also reroll one of your dice? On a close roll (especially if there is other uncertainty, such as what tricks the GM might pull) you can utterly freeze up.
Anyway, the bottom line here is that it is possible to introduce all sorts of dice tricks into Cortex Plus, specifically Leverage, but it's not a good match.
1 - This, right here, is one reason I prefer GM transparency. When the GM has mystery powers to throw into the mix in response to player actions, it invites paralysis and paranoia.