[Fildy]

Given a standard tight and aggressive game in the first few levels, your success at SnG play is largely a matter of how skilled you are at stealing the blinds in the later rounds.
Since the online sites make their money by collecting your tournament fees, they want the tournaments to be over as fast as possible so that you will buy into another one. This is why they arrange the blind structure to rise as fast as they think they can get away with.
Some people say that SnGs are a crapshoot because of this. My bankroll disagrees. It's simply a different skill set than other forms of poker, and mastering blind stealing is the place to start. The fact that SnG play is different from other forms of poker is what makes it so profitable for the specialist. Armed with the proper knowledge, a SnG specialist will trounce any player who is not accustomed to the fast 'n furious SnG high blind structure.
Let's see how we can use the Preflop All-In calculator to look at some typical blind stealing scenarios. We are only going to scratch the surface of this topic in this tutorial, but it will give you a feel for what you can do with the calculator, and why it's important.
Let's start with a specific situation just to get the ball rolling. Click the links as we go along to choose the options in the calculator. I'm choosing a $109 buy-in here just to put some dollar figures on the expected values.
Say there's 5 players left and the stacks are even so far. Blinds are up to 150/300. You're playing in a typical 50/30/20 payout structure. You can see this by going to Edit->Structure and looking in the right hand columns.
It's folded to you on the small blind and you've picked up QTo. There's one player left to act behind you. You haven't been playing wildly or pushing much yet, so you expect some credit for an all-in here.
You might think "All-in? Why not mini-raise to 600 here. I want survival value. Good players don't risk all their chips unless they have to." And to that I say: you are thinking like the guys I crush daily in these things. Snap out of it. If you already understand why a mini-raise here is usually a bad play, you can skip down to the next section. Otherwise, read on.
Think about what happens if you raise to 600. You are giving your opponent 3:1 pot odds to call your bet. There will be 1200 in the pot, and you'll each have 1400 left, so the only real option left on the flop is all-in (or at least a bet which leaves you completely pot committed). If you don't pair up (which you won't about 2/3 of the time) and check, you're likely to get his whole stack shoved in your face. You'll be left in a tough spot, and will probably have to consider folding. Your opponent has just risked 300 chips to win 900. That's a pretty good deal for him, and you're not in the business of handing out good deals, you're in the business of taking 'em. If this was a steal from anywhere other than the small blind, you'll also give him the option of moving in on the flop, again, leaving you in a bad spot if you miss the flop.
When you consider the variables here, you generally end up with what's commonly known as the "10xBB rule." That is, if you have less than 10 big blinds in your stack, and you want to raise, you should either raise all-in, or raise with the expectation of calling a re-raise or a flop bet. In other words, once you raise, all your money's going in, so most of the time you want to apply the pressure yourself, and put your opponent all-in before the flop. Some people like to adjust this a little to say 8-9x, but somewhere in there you have to draw the line and start applying the pressure. Right now you've got a little more than 6x the big blind, so all-in is the best preflop raise amount the vast majority of the time.
Analyzing the all-in
The main ingredients in this situation are:

How much you pick up in blinds if you steal successfully.

How often you're going to get called.

How your hand fares against the range that calls you.

How much you stand to gain if you get called, and win.

How much you stand to lose if you get called, and lose.
Got all that? No Problem. Most players sort of understand these things instinctively, but they just don't know how to put it all together to make the right decisions. They're just kind of guessing. Sometimes their guesses seem ok and sometimes they just can't tell the difference between running bad and making bad decisions. Sound familiar? Let's keep going...
Say the player in the big blind seems to be a reasonably careful player who you expect might call you only if he picks up 77+,A7+ if you put him all-in. That is to say, a pair of sevens, eights, etc., or an ace with a seven or higher. Ok, so what should you do? Hit compute and some numbers appear.
Let's take them one at a time.

% Hold is how often BB picks up a hand in his calling range

% Call is how often he calls you. Since it's only him acting, this is the same as % Hold.

% No Call is how often you pick up the blinds.

% Win is how often you win the pot when you are called.

$EV Win is your equity if you get called, and win

$EV Lose is your equity if you get called, and lose

$EV Call is your average equity if you get called

$EV No Call is your equity if you successfully steal

$EV Fold is your equity if you fold the hand

$EV Push is your equity if you move all-in, considering all other numbers
The bottom line is that folding leaves you with about $188 of equity and pushing boosts it to $210, so this figures for a clear push under these conditions. Now's where it gets interesting. Given all these same assumptions, what happens if you picked up a measly 86s?. Hit compute again and watch the $EV Push number. It only went down by a mere $2 out of over $200, and is still a clear push. This is an important concept that I will expand on in later lessons.
The dominant factor here is that you are not particularly likely to be called, and so any hand which is behind the range 77+,A7+ is really not too different, whether it's QTo or 86s. Just for fun, try it with 32o, and compute. Still a clear push!
Why it works
Now, before you start moving all-in with your 32o every time in your next SnG, let's look at why this move works. First,

The 450 chips you pick up helps your stack quite a bit.
Take a closer look at two of the numbers in the calculator: your equity if you fold on the lower left, and your equity if you pick up the blinds, in the upper right ($EV No Call). Picking up the blinds helps our stack quite a bit, raising your equity a little less than 20%.
If instead the blinds are at 50/100 in this spot, and you try it again (hit ok when warned about your strategic options), now you see that this is a losing play, and you should definitely fold that trash. Of course, if blinds were at 50/100, you'd have more strategic options that going all-in available to you anyway. Although you'd still definitely fold your 32o!
The second thing is:

Your opponent respects your all-in.
Since you make so much equity when you pick up the blinds, every hand that he folds makes you money.
Consider a different scenario now. Imagine that you'd picked up several good hands in a row in this spot, and you'd moved in each time and stolen the blinds. The guy on the big blind is sick of you, and he thinks you're a bluff king, so you expect he's going to go to war with any pair, any ace, any king, and queens with a 9 or better, i.e, 22+,A2+,K2+,Q9+. Let's go back to your original QTo and blinds 150/300. What happens now?
Now it's super close. When it is this close, less than 0.5% of the prize pool, I advocate folding and waiting for a better spot. This illustrates the importance of considering your "image" when you are attempting a steal. This is pretty basic stuff in concept, but being able to put actual numbers on it really helps you to understand how to apply these usually vague ideas more completely and with more confidence. Once you can distill situations down to "what might this guy call me with?" you will be way ahead of your opponents who are just guessing whether their pushes are any good or not. Their guesswork is your advantage.
Stack sizes: punishing the short stack
Let's make the situation a little more interesting now by considering what happens when the stacks aren't even. This introduces new things to think about, and new opportunities to capitalize on. Knowledge is power. Let's begin with the very important case of pushing all-in over stacks that are shorter than yours.
Let's stick with the same general scenario, except now you've built a bit of a stack to 3800 and the guy on the big blind is hurting at 1100 because you've been taking his blinds, and you figure him for the same loose calling range because he's tired of getting pushed around.
Now your very same QTo hand figures very differently. Push it, baby!
Why is that? It's pretty simple. With even stacks when you got called and lost, you busted out of the tournament. Now if you get called and lose, it doesn't hurt you nearly as badly; you still have some chips to play with. Look at the equity in the $EV Lose columns for the two cases to see what I mean.
Now let's do something which is pretty interesting for this scenario. Let's start analyzing some calling hand ranges for BB, trying to see how he could hurt you by choosing the right calling hands.

TT+,AQ+. Try it.

22+,A2+,K2+,QJ+. Try it.

Any two cards. Try it.
He can't hurt you by tightening up, that's for sure. Even if he magically chooses only exactly the cards which are ahead of your QTo, he still can't hurt you here. Pushing all-in here is mathematically unexploitable. There is absolutely nothing the big blind can do to prevent you from taking advantage here, even if he doesn't care about winning and is ready to sacrifice himself to hurt you. Obviously, recognizing and capitalizing on these situations is very important!
Ok, with that small revelation I will conclude the first installment of blind stealing 101. There's a lot more to this topic, including:

stealing from other positions

pushing with a short stack

stealing strategy on the bubble

stealing strategy ITM
and SnG Power Tools will help you to examine each one.
To register, visit the registration page at sitngo-analyzer.com.
Good luck at the tables!

This is from SNG power tools.
Site in bold