Pace is one of the most important factors in any game of dominion. The player whose deck is better suited to the pace of the game is in incredibly good shape to win, and her opponent must catch a good bit of luck to beat her. The question is, then, how do you get that?
There are two general concerns here. The first is controlling the pace. There are a few different ways to do this. If you're trying to speed it up, often the best way is to just build the fastest engine to get to 50% of the VP. But there are other options as well, such as gainers which can help you run three piles out, and trash-for-benefit cards (I'd like to call them sacrificers or saccers for short), i.e. using a salvager to trash a province and then buy a province.
There are also several ways to slow the pace down. The most obvious of these is attacks. Almost every attack will slow the pace down. Curse-givers will make every engine, even a big money engine, sputter. Handsize reducers do the same. The Spy attack speeds you slightly whilst slowing your opponent slightly (though because of opportunity cost, spy may not actually speed you much at all). But there are other ways of slowing the game down a little bit as well. If your opponent is going for provinces, you can slow the game down considerably by not buying any. This usually more than doubles the amount of time it takes them to run the pile out, as their deck is probably better suited to grabbing 4-5 provinces than it is in getting all 8.
Besides controlling the pace, the other major factor is knowing (and controlling) the pace your deck wants to run at. Now, as always, every deck is a little bit different, but by and large, most cards can be separated into groups based on the pace they want to play at. The essential short term cards are victory cards. The player with the VP lead has, in the vast majority of circumstances, the has the short term advantage - if the game ends quickly, they're likely to win. However, there are a number of other short-term advantages as well - having a bigger hand (usually due to hitting them with a handsize-reduction attack) or a better deck (due to chancellor or some kind of sifting), having a pin (like the infamous deck-killing masq pins), and even just having stronger cards in your hand right now are all short-term advantages.
There are also forms of long term advantage. One of the most obvious of these is VP-token generation. Every time you play one of these cards, you gain points with no cost in deck clogging, ergo the longer the game is, the more you can play them, the better it is, therefore they play well in the long run. Deck thinning, usually through trashing, is another common, key form of long-term advantage. Engines of just about any kind are long term advantages, even big money - the caveat is that if your opponent has a better engine than you do, they're going to have the long term advantage, not you.
But the different engines you can build have different speeds. In general, there's trade-off between being faster and being better suited for the long haul. This produces a sort of characteristic curve for that strategy, and its the combination of these curves between different decks that determines the contour of the game.
So let's look at some concrete different cards and the paces they play at. Philosopher's Stone is perhaps the longest of the long term ccards. It takes a long time for it to pay off, since it requires an investment of going for potions, and as the game goes longer, it only gets stronger and stronger. Apart from basic victories, something like salvager is close to the quintessential short-term card; it gives up your long-term assets for tangible short-term gains. But most cards are more complicated than these.
Chapel is a great long-term card because it trashes efficiently, which makes you a little weak in the short term. However, if you're just going chapel/money, you're going to be susceptible to clogging up (ironically) if you green too soon. So it's a fast-but-not-long card, generally. Bishop gives you some short term dividends - generally the lead when you play it, but it tends to accelerate your opponent even better than you, since they aren't wasting time buying bishops and can choose to not trash when they need the critical money. Then late game, its short-term benefits show up again, as the extra VP it gives become huge. Gardens is an interesting card in that it is very good in fast-paced games, as it's quick VP and can often be part of a quick 3-pile, classic gardens. But it can also be good in the extreme long term, as you build up a huge mass of cards. However, it isn't so hot in the mid-rage, where it tends to be about a duchy, or often a little worse.
Just about every other card has these kinds of dynamics as well, some simple, some less simple, but its all about having the short term advantage when the game comes to an end. And if you ever find yourself at both long- and short-term disadvantage, that means it's time to roll the dice - for instance, by breaking the PPR.