The problem with not attacking, of course, is that you give up two damage and that damage could make a difference. Or could it? There's nearly always a chance, since both players are going to have draw steps and you can't be certain you know the contents of his deck. There are plenty of scenarios where two damage matters a few turns down the line, but are they plausible enough to take the risk of being attacked?
This game, your goal is to win by attacking with Paladins and finishing him off with Jitte counters and/or Lightning Helix. As long as you have mana, you have that three damage in your pocket putting your opponent at an effective eight. By putting your opponent at an effective six, you plan to finish the game off safely two turns from now: Next turn you can attack him down to four, winning without needing to ever use the Jitte. That lets you leave up the mana for Lightning Helix or Devouring Light, which in turn means that next turn Wildfire is no longer a threat to you.
If you leave both Paladins back to block, that means you're going to have to do a little more to finish the job. You clearly won't be attacking with both Paladins at the same time unless you're confident you have the kill, as that would undermine the whole point of leaving both Paladins back to guard against a second Magnivore. Once you start down a road like this, you generally need to keep going down it. There's nothing worse than sacrificing to avoid losing to a card only to realize that to make up for what you've sacrificed you need to lose to the card anyway.
The key question is, can your opponent set up a plausible enough scenario where you lose the game because you didn't attack, and if so
If he casts a Magnivore on his turn, you're very happy that you held back so those scenarios are out. There are two categories of games where this move could backfire. Either you're hit by Wildfire, or you're hit by something because he had an extra turn. A Wildfire this turn puts you behind by two life points and that's potentially an extra turn that your opponent will have to answer your threats or kill you. Until you draw land you'll be essentially helpless, and your opponent will get to keep a minimum of two lands. That gives him mana for Eye of Nowhere, which alone is enough to turn the game in his favor. You don't want to be messing around in this spot, and this scenario is quite likely. It's easy to see you losing the game because you can't quite finish him off.
The other question is whether there's a realistic chance that he can draw a solution without starting with Wildfire and cause the two life to matter. The problem for him is that he can't start with Eye of Nowhere, Magnivore or Wildfire. What does he start with instead? If he starts with nothing, that indicates he has none of these cards and you are in great shape. As noted last week, these scenarios are extremely hard to lose and two life is unlikely to matter. Your opponents' deck is no longer all that dangerous to you if you play conservatively.
It seems like it now comes down to which of these concerns is more important. In the Wildfire scenario you change the size of the window by a turn and make the game significantly more difficult to win. In the Magnivore scenario, you can turn what would have been a precarious position (or in the nightmare scenario even an outright loss), into a game that probably favors you as long as your opponent doesn't have another big weapon in reserve, although you'll likely lose to any serious follow-up. Wildfire is more likely than Magnivore, but there seems to be a strong argument for hanging back. I certainly can't fault anyone for that decision.
I already touched on this problem when I noted how bad it is to play around a card only to find that playing around it made you more vulnerable. I've talked a lot about the decision of whether to play around cards, and how that decision can determine the entire course of the game. In this case, I essentially asked myself whether an aggressive deck like this one could afford to hold back more creatures to block than it needed for the attackers in play. Would I seriously be willing to leave my first creature back to block on the off chance that my opponent would play a Magnivore? Would I do this for an extended period? How much damage was I prepared to miss out on in this spot?
There's a dramatic shift that takes place when you can see the end of the game. If I couldn't see all the way to the end of the game, I wouldn't even consider holding the second Paladin en-Vec back. Once I do that, I'm dooming myself to a defensive game that is highly unlikely to succeed. That's not the way I built my deck, and I would never include Devouring Light in such a deck if I had the choice, even if I considered it a high enough quality card to warrant inclusion. It does not play into my game plan the way I would want it to.
In Magic, your assumptions are never safe.
How do you make sure your rules of thumb don't prejudice you and cause you to unfairly dismiss options? It starts by trying to keep an open mind and keeping an eye out for when the situation has changed. It also requires constantly questioning those rules of thumb. In Magic, your assumptions are never safe. Even if they were correct when you first made them, Magic is constantly changing. A lot of early Magic theory turned out to be wrong even at the time, but most of what held true then is not true now. The game has transformed itself from one in which card advantage was rigid and absolute into one in which advantages are far more fluid and complicated. It's a better game, but it's that much harder to figure out. To me, that's just another reason it's a better game.