501.9a To "scry N" means to look at the top N cards of your library, put any number of them on the bottom of your library in any order, and put the rest on top of your library in any order.
Obviously, the larger N is, the more cards you get to see and the better the card becomes. The inverse is, of course, true. There are nine cards with scry in Future Sight, and almost all of them are playable.
Putrid Cyclops is the worst of the cycle. Most scry cards enable you to positively manipulate your library to ensure you are at least happier with your next draw or two than you otherwise would have been. This guy, however, has a measly Scry 1. He will often stay alive as just over 50% of your deck will cost two or less, and you get two shots at him living; but even then these odds are not great. You have approximately a 25% chance he will die from his own effect, not even factoring what might happen if your opponent has an active pinger on the board to reduce your odds further.
What do you do when you look at the top card and it is the card you most want to draw next turn which, unsurprisingly, costs more than two? You leave it on top and put the Cyclops in the bin. What if you have to play him late in the game or when you are mana flooded? Almost all of the cards, at least 75% of them, that keep him alive are land, so the only way he will survive is to effectively render your next draw impotent. I am surprised this card only had scry 1, when he is already poor, especially when comparable to Nessian Courser, as he is the only scry card which does not come with the side effect of actual improved card selection. I can only suppose that a higher scry would improve his chances of being played in a constructed Suicide deck, an archetype which needs all the help it can get!
The other four in the common cycle are all very powerful. Riddle of Lightning and Judge Unworthy are both quality removal spells. They are, however, slightly conditional; we will all have awkward moments when we are forced to point a Riddle at a Spectral Force in hope of flipping Bogardan Hellkite. Both of these cards demand that when you build your deck, you work out the average casting cost so you know how much you can expect it to hit for. It should be somewhere around the 3.5 mark. However, further attention is needed. You need to pay attention to how many cards you have on each slot, so you know the extremities of the range. This way, during the game, you can readjust the expected range, as it might be the case that you have drawn all of your late or early game cards.
Llanowar Empath is more my kind of card. This guy will almost always read "two for one." Sometimes, there won't be a creature in the top two, putting you in a predicament. You can gamble on the third card down being a creature (probably around 40% likely), or you can give up the hope of drawing a card and scry as per usual. This will all depend on the quality of the cards scryed into. Sometimes, there will be the land you need along with a powerful spell for a turn to come. In this scenario, it is often more important for you to hit you next land drop than for you to draw a guy. It will often depend on just how important one of the cards you scryed into is, or how important a land is. For example, if you scry into Swamp and Damnation, it will often be in your interest to miss with the Empath, so as to wreck your opponent with the Damnation.
Sometimes, even though you need a land, it will still be worth gambling on the creature, if the land is not absolutely crucial to your game plan and the cards you can currently see are of no import. In this case, you have either a 35% chance of hitting your creature and drawing it and, if you miss, then there is only a slim chance it will not be a land (given that your average deck is comprised of approximately 35% creatures, 43% land and 22% other stuff). These are the kind of decision that you should be thinking about, even though you will normally happily put a creature on top.
Foresee is one of the trickier scry cards. Without scry this would just be an expensive Counsel of the Soratami. For the additional mana, we get to scry 4. Four is a lot. In the late game, this will mean that we definitely draw two spells, and earlier on we will find the right ratio that we need. If we are specifically digging for something, be it our bomb or our out to the given situation, we get to see the top six cards. Don't like the top four, ship them to the bottom and shoot blind! This might well be the greediest of the scry cards. Let's say it's late in the game and you scry into two Islands, Ashcoat Bear and a Primal Plasma. The board position is fairly even at this point and we have more than enough land. I would always ship everything but the Plasma to the bottom. The Bears just aren't good enough to waste a draw on. Sure, I might draw a land instead, but I would have drawn that land a turn later anyhow. By not taking the Bears, you effectively cycle them away for free. It will often be the case that you will ship all four cards away, because you have to find an answer to the Dragon that will kill you in a couple of turns—do not be afraid to dig into the dark.
New Benalia is a card that I think most people overrate. Sure, it's a land with a free ability, but coming into play tapped is a huge downside. To boot, scry 1 is pretty negligible. I would only play this card under a few special circumstances: either a deck with lots of land, one where it was empty on a curve slot, one with no early suspends (unlikely given white has three common one-mana suspend cards), in a deck which can only really win through its bombs, or a deck with no early drops. It is bad in an aggressive deck that will probably only feature seventeen land and will want to hit its curve every turn, where topdecking it as your land feels like getting hit in the gut.
When you peek at the first card, you have to decide whether it's good enough to end up as one of your next few draws, or whether you want to gamble by putting it in the bottom, hoping that the next two will hold what you want. In general, unless the first card you see is one of the best cards in your deck, you will want to ship it. If you are desperately searching for a land, you should still probably put a land to the bottom if it's the first card, because odds are, you will always see another. Assuming that the first card you saw was not Damnation, then the next two cards come in. You have a little more choice now, but the guiding principle will still be to put them to the bottom if they're not hot. You should be cautious though, because the best Annelids are the ones where you end up with two or three quality cards on top. The final scry is easy as there is nothing special left about it. I am very aggressive with the first two scrys because I want this card to dig me to something good, not something ok. The next card you draw should always be spectacular, whether it's a removal for their best guy or one of your best threats.
One of the cards that I've been thinking about the most in Future Sight has been Mystic Speculation. It is a card with a lot of potential. It reminds me a little of how much Sensei's Divining Top was talked about in Limited. For three mana a turn, this card is effectively a Top activation coupled with a shuffle whenever you want to. Unfortunately, I have yet to play this card, so my opinion here is purely theoretical. Maybe it's a powerhouse in disguise, but I think not.
It seems like it will fit in two particular decks. It will find a home in the blue train-wreck, where everything went wrong in the draft and you will either be light on playables, or light on truly significant spells. Here, you will need the card to dig to your real spells in order to give you enough game to survive. The other deck that will find a home for Speculation is the white-blue never ever die deck. Here, it will be better than Whispers of the Muse. You will stabilize the board and will simply need to find your win condition, or enough flyers to overwhelm your opponent. Speculation is the card I most want to fit into a deck right now, just to get a real opinion of it. At the very worst, it might entitle you to keep a dodgy hand as a pseudo-mulligan, where you can burn it on the first turn to make sure you find the needed land.
That's it for the actual cards with scry, but I want to talk a little more about various aspects that come up concerning the ability in general. Firstly, every time you scry, take a moment back to analyze what's going on. Have your hand in one hand and the scryed cards in the other, and just look at them. What you are doing is working out what will happen in the next few turns. You have almost all the information in front of you right now, so make the most of it. If you're mana-screwed, work out whether you need to leave one or two mana on top. It would seem obvious that two is correct, but it might pan out that it is not.
You have to evaluate the true impact of the cards, not the impact they would have now, but what they will do turns from now when you cast them. A Hill Giant two turns from now might seem good, but when you draw it, you might have fallen too behind in tempo or board position such that you would rather have gambled your draw on the next card down in the hope of it being what you needed. You normally have enough information in front of you to make these decisions—sure, your opponent will do and draw things in that time, but you have to take this into account as well. When you scry, you will often be doing the thinking for three turns. Make sure you appreciate that and think accordingly.
Sometimes, you will be holding two of your six drops in your hand and will only have five land. You scry into the sixth land, instinct tells you that you want it, so you leave it on top. A turn later you draw your seventh land and realize you have no use for it. Sometimes, when you scry, you will be able to put yourself in a win-win situation. If you draw a land, then you can play the spells in hand, if you draw a spell, you can play that instead. Sometimes you will need that sixth land right there and then, but more often than not the correct play ill be to scry it to the bottom. Be careful not to condemn a future draw as dead if you can help it.
There is one more situation that you will face time and time again: whether to burn a scry spell just to scry or not. You've been holding Judge Unworthy for a few turns now, and you've been attacked by an Ashcoat Bears all the while. And all the while, you've drawn nothing but land. Each turn, you will want to cast the Judge Unworthy more and more. In this case, you must resist. You will draw an answer to the Bears, and you will Judge a far better card Unworthy, and then you will benefit from the scry just as much as you would now. But sometimes it is not as clear-cut.
Sometimes, you will be ahead in tempo but out of gas. It will often be better to Riddle a random guy, or even dome your opponent, just to ensure that your next few draws will continue to put the pressure on enough so you will seal the game. One of the biggest skills in a Limited mage's arsenal is the ability to reevaluate cards. Sometimes, it is better to Lightning Axe your opponent's Goblin token, sometimes it will be better to chump block with your Chronozoa the turn you cast it. The same applies to scry spells. Sometimes the fact that you are scrying will be more powerful than the other effect of the spell. Only though continual reassessment will you be able to make the correct decision. There is no easy rule as to when to burn your scry spell. It is normally correct not to do so, but stay on your toes.