Stealing the blinds. How often to do it, and what does it pay? (Stats)

[fabadam]

In the view of the recent hype of posting special 100th posts, I want to make my 179th post special. OK, so I'm a bit late. Sue me.

Seriously, I recently did the micro session review for the first time. I noticed that onthebutton (my review buddy) didn't steal the blinds enough. So I was berating him in the review to steal a lot more.
Afterwards, I thought, well OK, I may be stealing a lot, but does it really bring in a lot of money?
So I did a little investigation on my PT database.

I started by just selecting all my limit games, including short-handed to extremely short-handed. I then noticed that this was skeweing the data, so I tried again, this time filtering for 7 or more players (thereby excluding all my 6-max play).

Here are the results of the Dutch jury (obscure Eurovision gag):

Total # hands with 7+ players: 15,141
(this is $0.5/$1 and $1/$2 on Party)
# hands from SB: 1656
# hands from Button: 1661
# hands from CO: 1561 (weird, but that's what it says)
Attempts to steal blinds: 43.90% (I'm a maniac in this area)


I then used the Filter of the general info tab of PT to determine what my steal results were, by filtering on "Chance to Steal & Raised":
Opportunities to steal where I raised: 367 hands
Winrate on these hands: 19 BB/100

If you calculate that down to the total winrate, I steal 2.4 times per 100 hands, winning 2.4 x 0.19 = 0.46 BB/100 of my winrate is due to blind stealing.
(This is a bit much, since it also includes "steals" with monster hands like AA)

This is all very nice, but how does that help in real play? Not much, except to learn that stealing can be profitable.

So I tried to determine what hand range is profitable for blind stealing. Unfortunately, sample size is rearing its ugly head at this point: there's only a few hands of data on each hand by this point (you'd need at least 300,000 hands to have statistically reliable data by my estimation).
However, if I look at the results by hand I notice that:
* Stealing with A-any is profitable for nearly all of them
* The same for K-any.
* Q-x becomes iffy below Q9
* pocket pairs and connectors do really well. Small sampe size though.

So my stealing recommendation for small stakes tables is:
* any pocket pair
* any 2 cards 9 or higher
* A-any and K-any
This boils down to a 40.9% Attempt to steal percentage.

Final note: my stealing attempts from the SB were much less successful than from button and CO. Position really is king.

[bozlax]

This is a good analysis of your play. You're missing an enormous element, however. Do you know what it is (hint: I gave you a hint in my first sentence)?

[fabadam]

[ QUOTE ]
This is a good analysis of your play. You're missing an enormous element, however. Do you know what it is (hint: I gave you a hint in my first sentence)?

[/ QUOTE ]

No, I didn't analyze my opponents, other than that they were at a Party .5/1 or 1/2 table.

Edit: my steals are often read dependent: I have Fold SB and BB to steal on my HUD display, and don't try to steal against obvious never-folders, except with good hands. This is a recent development though.

[kerowo]

What do YOU do on the turn when you steal. What do YOU do on the river?

[fabadam]

On the flop I basically auto-bet.

On the turn I nearly auto-bet, but I do slow down if I smell a rat. I'm not able to give strict guidelines for that, it depends on the board and the opponent.

On the river I'll bet if I think my opponents called all the way on a draw. I've often felt afraid to bet the river on crap, and been amazed at what folds I get. A LOT of villains will not fild until the river.

I find fold to bet by street a very useful stat: there are people who will never fold the flop, but always fold the turn UI. There are people who hardly fold flop and turn, but fold 80% on the river. The differences are stunning.

[plonker]

Just coming back to poker after a few months off (due to me needing my BR) starting from the very bottom again .5/.10 C tables. Blind stealing was and still is my weakest part of my game.

I've started looking for blind steal attempts (read dependentI.e Passive) with any broadway cards.

If anyone can comment/guide on this wold be really good.

Fabadam really good thread you've started here IMO

[Smurph64]

*grunch*

Great post. I have some comments about steals based on my observations. Steal is about timing more than anything for me and about being comfortable with a call against those players left in the tank when the steal fails.

It also means that I am ok folding my hand if I get reraised on my steal attempt or the one guy I don't want to call out of the 4 or 5 guys left calling. This may look weak tight and that can be a good thing too.

This isn't just related to who is in front of me in the blinds but what happened in the last ten hands and what I want to happen in the next ten hands.

If you look at your steal rate by session I think that is something that will be much more useful to work with than overall considering what cards are stealing cards.

I mean by definition a steal is a raise made with the intent of everyone else left to play folding. Your cards are irrelevant. If I win a steal attempt that goes to flop its a failed steal attempt pure and simple.

The card selection is important in that it teaches you how to play post flop with a less than premium hand but its still not a steal recommend or steal don't recommend criteria.

What makes a good steal attempt then if its not just cards? I usually try to set up steals when my cards are running bad, whether that means I am getting outflopped or just being card dead.

I make raises in middle position sometimes against certain players left with game theory hands with my sole intention being to beat the blinds.

Now I have no idea if this works at the higher levels but I do know that it works at these levels and I need as much cheap practice as possible.

I always go and play at least a few 100 hand sessions each week where I work on one aspect of my play. Blind defense, Steal attempts, bluff attempts after the flop, whatever.

I liken poker to golf. Practice is important. Focus practice is important. Its fun to blast a bucket of balls with your driver but how much will it help your game. Similarily its fun to blast away at a weak table but its much more practical to work on a game part that you have happen all the time.

Steals are excellent shots to practice. Sometimes they are the difference between a winning and losing session.

[fabadam]

Smurph, you make some good and/or interesting comments.

[ QUOTE ]

It also means that I am ok folding my hand if I get reraised on my steal attempt or the one guy I don't want to call out of the 4 or 5 guys left calling. This may look weak tight and that can be a good thing too.


[/ QUOTE ]

I don't think I've folded any preflop raises preflop yet, since I always convince myself the huge pot is giving me pot odds. I think I'm right there too -- if the pot is big enough, all you need to do is hit the flop, any odds.

[ QUOTE ]

This isn't just related to who is in front of me in the blinds but what happened in the last ten hands and what I want to happen in the next ten hands.


[/ QUOTE ]

I'll agree image is important. People here will often say the fish don't look at what you do, but that's not true.
For many of them that is actually ALL THEY LOOK AT. They have been looking at poker, and think it's all about psychology and reading your opponent's bluffs.
They understand bugger all about pot odds, but they do look at what you do and make an impression from there. For many of the mor observant fish, their actions are almost entirely determined by guessing at their opponents behaviour (more so at 6-max IME).

[ QUOTE ]

If you look at your steal rate by session I think that is something that will be much more useful to work with than overall considering what cards are stealing cards.


[/ QUOTE ]

I disagree. When I'm stealing, I make a statement: I think my hand is better than yours. Of course, I have no clue about the hands of the blinds. So they have any hand. So my hand is better than 2 "any hands" .
Reason for this: if they don't fold (pity), I can still play, because I have a reasonable hand.

It's not just about bluffing out the blinds. I'm also telling them: if you want to play, you'd better have a hand. If they don't want to listen to me, I'm fine with that.

[ QUOTE ]

I mean by definition a steal is a raise made with the intent of everyone else left to play folding. Your cards are irrelevant. If I win a steal attempt that goes to flop its a failed steal attempt pure and simple.


[/ QUOTE ]

So this is not true. I see the steal as a semi-bluff.

[ QUOTE ]

The card selection is important in that it teaches you how to play post flop with a less than premium hand but its still not a steal recommend or steal don't recommend criteria.

What makes a good steal attempt then if its not just cards? I usually try to set up steals when my cards are running bad, whether that means I am getting outflopped or just being card dead.


[/ QUOTE ]

I'll sometimes try a steal on 82o against two guys who've been folding their blinds to any pre-flop raise. My database indicates this just isn't a good idea, though the sample size is inconclusive. So I stand pretty much by my range ( though you could make a case for A4+, K7+ and Q8+ just as well -- it's the general idea of 40%).


Note: smurph, from your posts I'm guessing you're way older than the average poster here, just like me. I'm 43, where I think most guys here are in their 20s. How about you?

[Bilgefisher]

Nice reply. I agree that a steal is read dependent, but I disagree that the cards do not matter. I might have missed your point there but hear me out. I think when stealing its a combination of cards and reads.

A steal can be broken down into situations.
-Your opponent is likely to fold no matter the cards. This is against tight opponents who never defend their blinds. You can widen your stealing hands and situations.

-Against calling stations, you want to put your opponent in a lose lose situation. If he folds he loses, if he calls your hoping he calls with an inferior hand and thus makes a mistake. You need to be a bit more selective on your hands. Your making a value raise here. Your betting on the fact that your two cards are better then an avg blind hand he will call with or better yet, fold with. Your also betting on superior postflop play against him.

-against tricky opponents you have to be cautious. If he's willing to 3-bet with air then you need to be ready on flop turn and river to counter his play. He may also 3-bet with good cards. I narrow my steal selection against these opponents since my postflop play is not nearly good enough to exploit these situations.

I am sure there are more situations that come up, but this is the basic 3 I see. I need better then "any cards" to outplay my opponents postflop if they do decide to call. I agree that a call is a failed blind steal. When I fail, however, I want a backup plan to keep me from failing postflop as well.

[bozlax]

[ QUOTE ]
Your betting on the fact that your two cards are better then an avg blind hand he will call with or better yet, fold with.

[/ QUOTE ]

Why would it be better yet if your opponent folded with worse cards than you had?

OP - yeah, that's what I was getting at. Stealing, imo, is more dependent on your opponents than your cards, to the degree that against the right opponents (ones that never defend their blinds), or if I have a table-image of being aggressive AND making my hand, I'll steal with nearly anything, but against opponents that have read enough to recognize a steal situation and will defend with any two, meaning their 3-bet plus further aggression postflop doesn't tell me anything, I'm likely to not even raise otb with AJo. For instance.