OK, let's talk about wmajik's article

[Irieguy]

First off, brilliant article. Very nicely presented. Others have mentioned that it ignores +chip EV/- $EV considerations (which it does)... but you need to be able to digest the information in the proper context. MJ alludes to the fact that he is only alluding to the facts. His point was not to present a formula for shorthanded play.

When I first read it I thought "yeah, I know all that... but nicely presented." But then I noticed something when I played my first session tonight. I started to notice that as the field and my stack got smaller, there were several opportunities to push where I otherwise would not have considered it. I did not expect this, because I have recently begun to feel that I'm playing a little overaggressively in bubblish situations with small-medium stacks. But these situations kept occuring, so I started making some notes. Here's what I found (initial thougths only, so take with a grain):

I was playing 4 tables at once, and made it down to 6 players in all 4 with less than 600 in chips (PP, $33 SNG). This is when I noticed that my thinking had changed a little bit as a result of reading MJ's article. I counted 16 situations where pushing would be advocated in the article when I would have otherwise folded. Sixteen! That doesn't count all the times I was pushing anyways. I was playing like a complete maniac. It was insane.

I never beat a better hand heads-up, not once. I stole the blinds at least 70% of the time. I got 3 firsts and a third. Interestingly, my third occured when I woke up with Kings on the button 3 handed and pushed. Both blinds called me with Q-9, and Q-6! They were sick of getting pushed around and decided to gamble. Q-9 made a flush, and I was the short stack, so I got 3rd... but I was poised to win all 4, which is something I haven't been fortunate enough to do yet in close to a thousand SNGs 4-tabling at multiple levels.

But my results are irrelevant. The point is that the applied concepts in the article influenced a very aggressive player to become even more aggressive. Here are my thoughts:

1. If applied indiscriminately, these concepts will probably not help your game, and will introduce so much variance that it may take months of regular play before you even know what's going on.

2. If applied discriminately, these concepts will improve your ROI without a doubt... but it will take several months of regular play to realize it.

3. If any significant portion of the SNG players apply any significant portion of these concepts to their game, SNGs will become very difficult to beat.

This should spark an interesting discussion.

Irieguy

PS- I would suggest that nobody provide the link or clarify what I'm talking about if somebody doesn't already know. In the spirit of MJ's article... you have to work at it to figure this one out. No spoon feeding.

[AleoMagus]

Yes, I think that while it does suffer from flaws of incompleteness and could easily be misinterpreted, it is a brilliant article. I also think that in a strange way, the only way to apply some of the concepts suggested properly, you almost already need to have some kind of foreknowledge of these concepts with which to fully understand what is being suggested.

This is not even to say that I fully 'get' what is being suggested, just that as I re-read that article I am starting to get a better and better idea of opportunities to be hyper aggressive with nothing. I have also been playing with these ideas lately and it has been fun if nothing else.

I think two things which are really noticeable to me are:

1) The amount that you lose when called is not always that great. Just as we assume that we should 'always' win when we push with AK and get called by Q9, it is tempting to think that when you push with 6J and get caught by AK you should 'always' lose. This just isn't so. I have had more than a few SB pushes caught by AK or AQ only to suckout and win anyways. It's actually quite a bit of fun. I think this equity in these pots where you get caught is really important. This is obvious I know, but it's just interesting to see it's effects. Beratement about my stupid lucky play seems to be popular.

2) If you have been pushing once or twice each orbit (which you will sometimes) and you are starting to sense that players are ready to make a stand, you really need to slow down. On this same topic though, catching a high pair becomes very valuable because rather than finding KK and raising 3xBB and stealing with it, pushing suddenly becomes a good option. A couple times now, I've been in this situation and though it hasn't been a lot of tourneys, players really seemed willing to get into a pot with me where they might otherwise have let me steal with my monster.

So, used right, these ideas are really big. The trouble is, that is hard to do and I'm not there yet, I'm sure of it. I don't find my results are sufferring, but I don't feel I'm necessarily going to see a big ROI jump yet either. Still, my presence at these games has been very big, and every player was aware of me, that is for sure. The guys on my immediate left in these games have been in for some interesting times. In this respect, SNG poker has been a lot of fun today.

As far as the article's flaws with respect to chip EV and $EV, I agree with Irieguy. I think that the real point is only being alluded to and if you think this flaw makes the article 'useless', then you might want to read and think about it some more. I have no doubt many will disagree and not because they are uninformed either. Dissenters might be correct as I think the jury is still out on this one.

There is some truth though, as I see it, and I am looking forward to the next few days playing with these ideas.

Any other thoughts?

Regards
Brad S

[Mike28a]

I see your point about not spoon feeding, but it only took me 13 seconds to get the article once I finished reading your post! Survival of the fittest, I say /images/graemlins/smirk.gif On to the read! Thanks

[Zelcious]

You've all probably seen this before but lets say that there are four players on the bubble with equal stacks. One player goes all in and another knows he's a 60% favorite if he calls. According to ICM he would make a misstake if he called, but chip wise it would be a great move. Then you can think what happens if he's just 50-50 or even an underdog. This is why I think the article is dangerous. You could make a huge misstake if you only look to chips and not realizeing that it's a tournament and not a cashgame.

[KJ o]

True, and there is also the case that if you have more than one player to act behind you, you have to be more careful.

The case he describes, with four players remaining, hero in SB and UTG and Button folds, doesn't happen very often at all, even on Party $10. I'll check the figures when I get home, but my guess is it goes fold-fold to me in less than 25% of hands.

HU his advice is great, but HU with huge blinds is a crapshot anyway. I don't think perfect HU play can increase your ROI by the suggested 8%. (Or can it?)

[wjmooner]

I have been out of town for a couple of days, so I missed the thread with wmajik's article. For those who may have missed it, since it is buried in a Strassa thread, here is the link.

http://teamfu.freeshell.org/tournament/theorem_blind_stealing.html

Very interesting. I think this may be help my 3rd/1st ratio.

WJ

[AleoMagus]

Yes, I think for many, better and more aggressive HU play could easily yeild an 8% ROI boost (especially if your ROI is low already). The two places where ROI will get the biggest boosts are in bubble play and HU. This is because the jump from 1-3rd and 2-1st are the biggest jumps that you will take.

I think that with the tightness I see on the bubble in a lot of SNGs, the SB gets folded to at very least 25% of the time. Some SNGs, this is more like 50% of the time. Also, I think that the conclusions which will spring from that article apply (to a lesser extent, of course) on the button and UTG also, and even in less shorthanded situations.

And yes, you are right that caution must be exercised with more players to act, but I still think that just adds a caveat to the points that the article is trying to make.

As far as Zelicious and all the ICM stuff, I understand what you are trying to say here. And I understand that for these reasons, you will not literally be pushing with any two every time you can take a shot at the BB on the bubble. The point is just that there are more +EV opportunities than we often think. That's $EV I'm talking about.

I guess what I'd say is this. Take the numbers in the article and analyze them with ICM to see just how profitable these moves are. If you or someone else doesn't do it, I will eventually anyways. The results of such an analysis will no doubt diminish just how often this can be done, but I think we'll still find the move is profitable more than one might initially suppose. I can't say I'm positive, but we'll see sooner or later.

This actually reminds me of another play I make a lot. I think Jason brought it up a while back (about an opponent who always did it if I remember) and after experimenting with it a bit, I have really integrated it into my regular play (and quite profitably too, I think)

Anytime it is four or five handed and I am up against opponents who are not too tricky, if some, or all of them limp to me in the BB, and we all have stacks of about comparable size 8XBB or more... I push. Any two.

Maybe I'll start doing it sometimes in the SB also now.

...

I've been thinking a ton in the past day or so about variations on this kind of thing.

Just one example: A really good spot to do this in, is when all the stacks are roughly equivalent, but one guy is really small and already out of the hand.

Ask yourself, if I pushed to you in the BB on the bubble and the guy currently UTG has less than the Blind he's about to move into, what would it take for you to call me?

Same thing with my push against the limpers. If one guy is already really short and out of the hand, I never get a call. I mean that literally, I never have. The only way I forsee it happening is when a tricky player limps to me with a high pair. It'll happen sooner or later, but not yet.

I'm sure this is all obvious to many, and in a way it is to me too, but I still feel like this is prompting a SNG breakthrough of sorts for me.

Keep it coming, I'm crazy about this thread.

Regards
Brad S

[rachelwxm]

Wow, you already started using it while I havenot finished reading! To be more specific, could you provide some of the examples that you would not push previously and are doing it now? All 16 of them would be great. /images/graemlins/laugh.gif

Also, my understanding is that ChipEV is close to EV when there are 5 or more pepple playing, distorted from bubble on. And the gap created by bubble only help this pushing, like no one would call you for 60/40 situation since it's -EV for them so it benefit the aggressor, right?

I am playing more at ps recently to fine tune my skills to use more implied odds and bluffing, representing, HP etc. But I certainly see great value of this article. /images/graemlins/cool.gif

[eastbay]

I think this theory has already been extended, corrected, and completed in various discussions on this board to any position with any set of stacks, with a reasonable approach to $EV. Open-raising all-in only, that is.

That I won't spell it out should be an indication of the value I think is there.

However, you just can't rely on open stealing all-in alone to crush these games. It's an important tool in the bag, but by no means enough, by itself, to be a big winner at the higher stakes.

For example, there's games where you'll never get to open from late position, or at least, not often enough to stay alive even if you pushed every time you had the chance. What then?

eastbay

[AleoMagus]

An interesting and mysterious post which generated some controversy at the time.

HERE (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Board=singletable&Number=103 5130&Forum=f22&Words=&Searchpage=0&Limit=25&Main=1 033781&Search=true&where=bodysub&Name=229&daterang e=1&newerval=1&newertype=y&olderval=&oldertype=&bo dyprev=#Post1035130)

Hmmmmmmm....

Regards
Brad S

[EnderW27]

Can this concept truly be applied in all analogus situations?

Imagine a 5 or 6 handed game where blinds are 25/50 and you're in the SB with 500. BB also has 500. It's folded to you.

Now, according to the math shown, it would be right to push with any two cards. But is this really the case? Chances are, you'll never be called except by premium hands and so your EV on the play itself has to be positive.

But...there are a few key issues I'm trying to think through.
First, when you are called, you're going to lose most of the time.
Second, even when you double up, you still don't assure yourself a payday, merely a greater chance at getting there.
Third, you're risking your entire stack to gain 50T, which is such a small amount compared to the chips in play.

I haven't done math on this, obviously. But it seems to me that this type of play would not be +$EV, even if it could be shown to be +EV.

[LinusKS]

[ QUOTE ]

That I won't spell it out should be an indication of the value I think is there.


[/ QUOTE ]

That depends on whether you're taking the Irie/Sucker line that if information is really valuable, you should only hint at it.

[ilya]

Irieguy, you are saying there were SIXTEEN times in 4 tournaments when it was folded to you in the SB and you pushed where before you would have folded?
You are including post-bubble play, then?

[ilya]

Obviously Sucker is talking about raising every single hand on the bubble from any position. Almost.

[mason55]

[ QUOTE ]
Can this concept truly be applied in all analogus situations?

Imagine a 5 or 6 handed game where blinds are 25/50 and you're in the SB with 500. BB also has 500. It's folded to you.

Now, according to the math shown, it would be right to push with any two cards. But is this really the case? Chances are, you'll never be called except by premium hands and so your EV on the play itself has to be positive.

But...there are a few key issues I'm trying to think through.
First, when you are called, you're going to lose most of the time.
Second, even when you double up, you still don't assure yourself a payday, merely a greater chance at getting there.
Third, you're risking your entire stack to gain 50T, which is such a small amount compared to the chips in play.

I haven't done math on this, obviously. But it seems to me that this type of play would not be +$EV, even if it could be shown to be +EV.

[/ QUOTE ]

I would think in this situation you would want to back off just a bit. You can still push with almost everything but once you either steal a couple of times, catch a good hand when you're not on a steal, or if you're lucky enough to double up then you can back off and wait until you're starting to get blinded away again, or the blinds go up, at which point they will be worth stealing. If you manage to steal a few times or get caught with bad starting hand but double up anyways, this will help you the rest of the tournament because you'll be "the maniac" and, like stated in the article, can probably get some looser calls or someone to reraise into your monsters.

[LinusKS]

Brad, here's what I've come up with.

If you want to check this, and/or add to it, I'd appreciate it.

First of all, there's some mistakes in magic's chart.

The one I'm working off is Table #4, the last row for table type "loose."

In that row, magic has 46% for "Fold%". The number should be 36%. For "Fold Equity" he has 0.8 BB. The correct number is .54BB. The others are 64%, -.325, -.208, and .332.

The final result isn't much different - .332BB, instead of 0.5BB, but I wanted to mention it, because it affects my own calculations.

What I did was to assume a four-player game, with approximately equal chips, and a standard 20/30/50 split.

I'm also assuming the total prize pool is worth $100.

To make the math easy, I gave player 1 T2375, player 2 T2250, the sb (Hero) T2500, and the bb 2500.

The small blind is 125 and the big blind is 250, which have already been paid (they're already in the pot).

In this scenario, the Hero's chips are worth $25.07, as are the Villain's.

It looks like this:

_____FOLD__WIN BLINDS___DIF___WIN ALL__DIF___LOSE ALL__DIF
EV%__.2507__.2731_______.0224__.3936___.1429___0__ ___-.2507
$____$25.07___$25.31____$2.24__$39.36___$14.29_0__ ___-$25.07
Chips_2500____2875______375____5375_____2875___0__ ___2500

I'm just converting to $ values for ease of comparison, but you can already see the differences -

The 375 chips in the blinds you're gambling to win are only worth $2.24 in real money, and the 2500 chips you're hoping to win from the villain are worth only $14 to you if you succeed.

If you plug these numbers in (and I haven't checked this part of my math yet, so if someone would i'd appreciate it) you get:

Blinds: $2.24
Fold%: 36%
Fold Eq: $0.81
Caught %: 64%
Caught EQ: -$7.36
Caught Result: -$4.71
Total EQ: -$3.90.

In other words, if the math is right, you should expect to lose $3.90 every time you did this in a tournament.

[ilya]

I'm no mathematician but your calculations look good to me, Linus.

However...who the heck is actually going to call all-in for 10xbb with 64% of their hands? I think even at the $10s it's extremely rare to find that loose a caller on the bubble. If one of you is down to, say, 5xBB, then maybe -- but by that point, winning just the blinds is much more profitable proportionately, no?

[wmajik]

As a general reply to what has been said so far:

I do indeed understand independant chip models and why chips are not on a linear correlation with $EV. There has been disagreement with what I assume are my remarks of a 10% late game improvement causing an increase in ROI. I can agree with that - more chips don't mean more $EV. But, I also happen to think that most (not all) ICM theory goes out the window in STT games once you're on the bubble. You should always be shooting for 1st, so this strategy is directly pertinent to the all-or-nothing mentality.

I _do_ believe ICM is a very strong concept and is fundamental to deep stack games like MTTs or early level STT where blind:stack:table ratios are all important. Late game STT, not so much. All that a STT really becomes late game is a steal and coin-toss circus, seeing who can pull off the most stunts before exiting the stage. There's not a lot of edges to exploit and no time to go around waiting for a clearly advantageous call or double-up situation. You have to hold a monster or see a flop for that to happen first. With a short stack, that's just not going to happen.

In addition to just stealing more chips, there are secondary effects that result from this strategy that I only skimmed over. So, I am also talking about an overall game improvement, rather than just quantifiable 10% chips or something of that nature. This strategy is just as much about control in critical situations as it is chip building. This is why correct timing and application of this strategy is directly relevant to ROI.

Also, this article was not and is not a comprehensive STT guide. I can't imagine how it can be construed as one. I do have a real STT guide on my site and have had a limit STT guide that has been sitting there since October of last year. In addition, I never remotely mentioned anywhere in my article applying this strategy from any position outside of the SB/BB - that's a whole different ballpark.

As for myself, even I don't play like a wild gunman that some may imagine - but it's getting quite close. This month as part as a rapid STT tournament on my site (http://www.tightpoker.com/forum/yabb/YaBB.cgi?board=tourney;action=display;num=10966043 90), I 6-tabled $10+1 games for 4 hours. Since I am not a STT terminator like ZeeJustin or Jason Strassa, 6-tabling was still difficult for me. But applying this strategy in many situations (enabling me to go on auto-pilot), I am still able to pull off solid ITM and ROI numbers. Long term, who knows if this is still valid, but I tend to think so.

Lastly, it's poker we're talking about here. When is any rule a rule when the best answer for most questions is "it depends"? If someone reads my article and decides to raise 23o to steal 50 chips with a 700 stack in Level III because they saw a +0.2BB expectation on the play, that's their prerogative. Is it mathematically correct? Yes. It is +$EV correct? No. It's not meant to be a hand-holding article. I wrote it, because I think the concept is important and can enhance the game of those players who already know what they are doing.

I realize most of my reply has been a rebuttal so far, so I apologize. For those who are looking for some more bones to chew on, consider a reverse situation, where you are sitting in the BB instead and the SB limps in.

Say you put the SB's limp on a 'Loose' hand; you can see that you are good to push, given apropriate stack counts. Most likely this person then folds and you win the blinds. How often do you think they want to limp in on you when you've demonstrated you're not afraid to shove your chips into the pot? And what do you think it means when they finally do limp again? The answers are pretty clear. This is the type of control I'm talking about that isn't just +chip EV.

[Irieguy]

Yes, from 5 players down to 2.

[Irieguy]

Yes, Aleo, I have been doing the same: developing an intuitive feel for when the general concepts may be applicable to other situations, namely when you get a limper involved. My initial thoughts are that including a limper when you have any reason at all to think that he may add some folding equity makes the move even more enticing.

In the past, if I was in the SB or BB on the bubble with absolute garbage and was able to see the flop for free or cheap, I would do it (generally speaking) if I felt that pushing would be too -$EV. The implied odds are good, risk is low, and there is some residual chance of being able to play poker post flop.

My current thought is that I am missing some opportunities to push with garbage in this circumstance.

Irieguy

[AleoMagus]

Thanks for that work linus. Interesting.

Initially, I almost was ready to concede I was completely wrong until I realized that that you were doing this analysis based upon wmajik's 'loose' player.

What you have shown is as much as we all expected, raising with any two when opponents will call you 64% of the time isn't going to work. This softens wmajik's argument in the sense that it cannot be applied indiscriminately.

Lets consider the argument vs tight players instead, as I don't know of too many players who will call in that situation with hands like J6, Q3 or 54. Even the semi tight player seems too loose to me here. I just don't expect calls from KT or QJ either in this spot unless I've been stealing a lot.

Fold equity = Blinds*Fold%
=$2.24*72%
=$1.61

Caught Equity = (Win% * Pot) - Total Cost to Play
= (38% * $39.36) - $25.07
= $14.95-$25.07
= $-10.11

Caught Result = Caught% * Caught Equity
= 15% * -$10.11
= -$1.52

Total Equity = Fold Equity + Caught Profit
= $1.61 - $1.52
= $0.09

So, you'd make about a dime each time you tried this play against the sort of player who'd only call with AK-A7, AA-44, and KQ

I actually think that's a reasonable calling standard for most in that spot with the possible exception of hands like A9,A8,A7 and the inclusion of hands like KJ.

So it's not super great and you could not get away with this profitably more than once or twice as the calling standard would be reduced enough to make the play foolish.

Many will argure that for a dime, it's not worth it based upon the tight assumption. They might be right, becasue as soon as an opponent crosses the 16% mark for hands they will call with, it is unprofitable. I am still intrigued though, especially in situations where I can be very confident of the <16% call. This would mean at least a basic observation of play is necessary, and that there may be other factors which will encourage a fold (like a really short stack at the table). I also think there is something to be said here about the push in the BB vs 1 limper who you think will call <15%ish.

(Math note: I am actually unsure how you get the caught EQ of -$7.36 here, but most likely, it is me that is getting it wrong. Hopefully it's me, because I'm getting a much worse figure than you did. -$10.11)

An 8% ROI increase. Perhaps not, Not just on the bubble anyways.

Any thoughts?
Brad S

[Irieguy]

I have a couple of comments about this.

First, the $EV values change tremendously as the 4 players' stacks become more disparate... particularly if the shortest stack has folded. This is the type of situation where MJ's concepts can be applied most dramatically.

Second, with all equal stacks, it's usually correct to assume that all 4 players will tighten their calling standards at least 20%. I think it would be the rare player in this scenario that wouldn't fall into the tightest category. This makes the move clearly correct.

Interestingly, though, how would you react if you felt that one of the players was a maniac? It would suddenly become $EV positive to tighten up even more and watch the fireworks. Well... if you are open-raising all the time, your opponents will think you are a maniac. They will incorrectly think your calling standards are too low as well... and they will in turn make it even more correct for you to open-raise. Here's my pearl for this thread...


**If you open-raise often, opponents will call less than usual on the bubble when there are 4 equal stacks**

If your opponents are sufficiently tightened-up in these cases, I think you have to throw the ICM out the window.

Again, just some initial thoughts. I'm working this all out on the fly.

Irieguy

[AleoMagus]

Glad you came to post in this thread

[ QUOTE ]
I do indeed understand independant chip models and why chips are not on a linear correlation with $EV. There has been disagreement with what I assume are my remarks of a 10% late game improvement causing an increase in ROI. I can agree with that - more chips don't mean more $EV. But, I also happen to think that most (not all) ICM theory goes out the window in STT games once you're on the bubble. You should always be shooting for 1st, so this strategy is directly pertinent to the all-or-nothing mentality.

I _do_ believe ICM is a very strong concept and is fundamental to deep stack games like MTTs or early level STT where blind:stack:table ratios are all important. Late game STT, not so much. All that a STT really becomes late game is a steal and coin-toss circus, seeing who can pull off the most stunts before exiting the stage. There's not a lot of edges to exploit and no time to go around waiting for a clearly advantageous call or double-up situation. You have to hold a monster or see a flop for that to happen first. With a short stack, that's just not going to happen

[/ QUOTE ]

I have to disagree with what you have said about ICM and SNGs. I think that the jump from 4th to 3rd is very important also, and I think that CEV vs $EV considerations are FAR more important on the SNG bubble than are early. I'm much more likely to pass up a small edge to avoid a big confrontation on the bubble than I am in the early stages.

[ QUOTE ]
Also, this article was not and is not a comprehensive STT guide. I can't imagine how it can be construed as one. I do have a real STT guide on my site and have had a limit STT guide that has been sitting there since October of last year. In addition, I never remotely mentioned anywhere in my article applying this strategy from any position outside of the SB/BB - that's a whole different ballpark.

[/ QUOTE ]

Yes, I can see that this is just another play to add to the arsenal. I think that is obvious. I do think that some of these ideas can be extended though. Perhaps not much, but some. Your limper suggestion is one example, and I think it would apply even if the limper was outside the SB in many cases.

I like your ideas about control, and I also think that these moves become profitable in more ways than the math alone indicates. I'm still not completely decided, but this has been some interesting stuff to look at.

Regards
Brad S

[LinusKS]

Well, by my calculations, the Villain costs himself $3.38 by calling here. Unfortunately, his mistake hurts both of you.

[ QUOTE ]
I'm no mathematician but your calculations look good to me, Linus.

However...who the heck is actually going to call all-in for 10xbb with 64% of their hands? I think even at the $10s it's extremely rare to find that loose a caller on the bubble. If one of you is down to, say, 5xBB, then maybe -- but by that point, winning just the blinds is much more profitable proportionately, no?

[/ QUOTE ]

At the 5xBB level your expected loss should be just $3.06.

[LinusKS]

Aleo, thanks for the post.

It looks like you made a mistake in the third line of your calculations. For fold equity and caught equity you're using the numbers for the "semi-tight" player, but for caught result you switched to the numbers for the "tight" player.

For caught result you should have

<font color="blue"> 28%</font>* -$10.11
= <font color="blue"> -$2.83 </font>

The final result is -$2.83 + $1.61 (fold equity) = -$1.22.


I am showing a $0.20 net gain from using this play against Magik's "tight" player.



[ QUOTE ]
Thanks for that work linus. Interesting.

Initially, I almost was ready to concede I was completely wrong until I realized that that you were doing this analysis based upon wmajik's 'loose' player.

What you have shown is as much as we all expected, raising with any two when opponents will call you 64% of the time isn't going to work. This softens wmajik's argument in the sense that it cannot be applied indiscriminately.

Lets consider the argument vs tight players instead, as I don't know of too many players who will call in that situation with hands like J6, Q3 or 54. Even the semi tight player seems too loose to me here. I just don't expect calls from KT or QJ either in this spot unless I've been stealing a lot.

Fold equity = Blinds*Fold%
=$2.24*72%
=$1.61

Caught Equity = (Win% * Pot) - Total Cost to Play
= (38% * $39.36) - $25.07
= $14.95-$25.07
= $-10.11

Caught Result = Caught% * Caught Equity
= 15% * -$10.11
= -$1.52

Total Equity = Fold Equity + Caught Profit
= $1.61 - $1.52
= $0.09

So, you'd make about a dime each time you tried this play against the sort of player who'd only call with AK-A7, AA-44, and KQ

I actually think that's a reasonable calling standard for most in that spot with the possible exception of hands like A9,A8,A7 and the inclusion of hands like KJ.

So it's not super great and you could not get away with this profitably more than once or twice as the calling standard would be reduced enough to make the play foolish.

Many will argure that for a dime, it's not worth it based upon the tight assumption. They might be right, becasue as soon as an opponent crosses the 16% mark for hands they will call with, it is unprofitable. I am still intrigued though, especially in situations where I can be very confident of the &lt;16% call. This would mean at least a basic observation of play is necessary, and that there may be other factors which will encourage a fold (like a really short stack at the table). I also think there is something to be said here about the push in the BB vs 1 limper who you think will call &lt;15%ish.

(Math note: I am actually unsure how you get the caught EQ of -$7.36 here, but most likely, it is me that is getting it wrong. Hopefully it's me, because I'm getting a much worse figure than you did. -$10.11)

An 8% ROI increase. Perhaps not, Not just on the bubble anyways.

Any thoughts?
Brad S

[/ QUOTE ]

[AleoMagus]

You are correct, I did make a mistake, though not in the third line, as I was actually trying to solve this assuming MJ's 'tight' player. My mistake was in the first line.

It should be:

Fold equity = Blinds*Fold%
=$2.24*85%
=$1.90

Caught Equity = (Win% * Pot) - Total Cost to Play
= (38% * $39.36) - $25.07
= $14.95-$25.07
= $-10.11

Caught Result = Caught% * Caught Equity
= 15% * -$10.11
= -$1.52

Total Equity = Fold Equity + Caught Profit
= $1.90-$1.52
= $0.38

Again, this assumes that in this situation, an opponent will call you with AA-44, AK-A7, and KQ. I think this kind of range is very reasonable to assume, even for a gambling player. More likely, the range of calling hands will be much tighter (it sure would be for me).

Oddly, it looks like we are still not getting the same numbers, as I am getting almost double your $0.20 results. Am I getting this wrong somehow? If I'm not, this play is looking better all the time. Just once/tourney would mean an extra $0.30ish/tourney (that's an immediate 2-3% ROI boost). I say only $0.30 instead of higher becasue sometimes you would be playing those hands anyways.

Regards
Brad S

[ZeeJustin]

Every calculation this guy does gives you a random hand in the SB. Just because a play is profitable with a random hand does NOT mean it is profitable with any 2 cards.

If you and the BB both have 10x BB, and you shove in the SB w/ 32o, and the BB calls you with the 64% range of hands that the author says a loose player will call with, you will lose about 1.572 BB / hand. This isn't even close to profitable.

Note that 32o will win 31% of the time when called against the loose player range of hands.
The loose player range of hands accounts for about 64% of hands.
When you steal, you will win 1.5x BB.
When you get called and win, you will win 10.5x BB.
When you get called and lose, you will lose 9.5x BB.

STEAL:(.36) x 1.5 .54
WIN: (.64 x .31) x 10.5 2.0832
LOSE: (.64 x.69) x 9.5 4.1952

4.1952 - 2.0832 - .54 = 1.572

[eastbay]

[ QUOTE ]
[ QUOTE ]

That I won't spell it out should be an indication of the value I think is there.


[/ QUOTE ]

That depends on whether you're taking the Irie/Sucker line that if information is really valuable, you should only hint at it.

[/ QUOTE ]

Indeed I am, as anyone interested in +$EV would be.

eastbay

[eastbay]

[ QUOTE ]
As a general reply to what has been said so far:

I do indeed understand independant chip models and why chips are not on a linear correlation with $EV. There has been disagreement with what I assume are my remarks of a 10% late game improvement causing an increase in ROI. I can agree with that - more chips don't mean more $EV. But, I also happen to think that most (not all) ICM theory goes out the window in STT games once you're on the bubble. You should always be shooting for 1st, so this strategy is directly pertinent to the all-or-nothing mentality.

I _do_ believe ICM is a very strong concept and is fundamental to deep stack games like MTTs or early level STT

[/ QUOTE ]

I think you have that completely backwards.

eastbay

[Zelcious]

A very good point !
The $EV will be even worse I think according to ICM.

[AleoMagus]

Ok, Good point.

I guess what this means then is that in the calculations

Fold equity = Blinds*Fold%
=$2.24*85%
=$1.90

Caught Equity = (Win% * Pot) - Total Cost to Play
= ( <font color="red"> 38% </font> * $39.36) - $25.07
= $14.95-$25.07
= $-10.11

Caught Result = Caught% * Caught Equity
= 15% * -$10.11
= -$1.52

Total Equity = Fold Equity + Caught Profit
= $1.90-$1.52
= $0.38

the 38% figure given for our win% in pot equity is oftn going to be a big exaggeration. I don't like defining my win rate against the 64% call. I'd prefer to define it against the 15-16% call. so, just to solve for exactly what win rate I need to make this profitable:

$1.90-(((X*$39.36)-$25.07)*15%) &gt; 0

-(((X*39.36)-25.07)*.15) &gt; -1.90

-((X*39.36)-25.07) &gt; -1.90/.15

-((X*39.36)-25.07) &gt; -12.66666

-(X*39.36) &gt; -12.6666+25.07

-(X*39.36) &gt; -12.403

(X*39.36) &gt; 12.403

X &gt; 12.403/39.36

X &gt; .3151

So, we need at least a 31.5% chance of winning a confrontation against the tight players range of hands in order to try this.

23o does not qualify
T4o does not qualify

89o does qualify
J7o does qualify

I will have more to say about this soon but need to run.

Any thoughts?

Is this reasoning correct?

Regards
Brad S

[LinusKS]

The point here is that you will never play a random hand.

You'll always be playing two specific cards.

The only way you could play a random hand is if you played your hand without looking at it.

If you look at your hand, you'll be better off folding hands like 23o, and (maybe) calling or min-raising hands like AA. The only hands Magik's analysis really applies to are "medium" hands - hands that will win about fifty percent of the time against a random hand.

[LinusKS]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]

That I won't spell it out should be an indication of the value I think is there.


[/ QUOTE ]

That depends on whether you're taking the Irie/Sucker line that if information is really valuable, you should only hint at it.

[/ QUOTE ]

Indeed I am, as anyone interested in +$EV would be.

eastbay

[/ QUOTE ]

Well, if you're really interested in +EV, you might consider intentionally misleading people. For example, telling people to raise all-in with crap on the bubble.

[Irieguy]

This is the difficult part about assimilating the information in MJ's article.

MJ is just saying "here's what happens if you go all-in from the SB every time it's passed to you and your opponent will call X% of the time." He is NOT saying "you should therefore go all-in with any two cards."

The only way to determine how you would fare if you pushed every time in these cases is to run the sim. with a random hand. That's an accurate simulation.

It's like somebody saying "if you go outside everyday in a city where it rains 25% of the time, you will get wet 25% of the time." You can't then argue "no, that's not true if you only go outside when it's sunny."

Irieguy

[KJ o]

[ QUOTE ]
So, we need at least a 31.5% chance of winning a confrontation against the tight players range of hands in order to try this.

23o does not qualify
T4o does not qualify

89o does qualify
J7o does qualify

[/ QUOTE ]

How dou you know that? Sounds like a link to a site I don't know about...

[ilya]

It's an oversight.

.....

Woohooo! I am an [censored]!!

[gergery]

[ QUOTE ]

It's like somebody saying "if you go outside everyday in a city where it rains 25% of the time, you will get wet 25% of the time."
Irieguy

[/ QUOTE ]

But the fact is once you go outside, there is now either a 100% chance you are wet, or a 100% chance you are dry. There is no 25% chance anymore.

Which goes directly to ZeeJustin's post, and where the biggest adjustment to wmajik's post needs to happen: when you are caught on the blind steal you are NOT a 1.2 to 1 underdog vs. loose players hands with many hands (same thing with the other types too).

in fact, if the loose player is going to call you with the best 67% of hands heads up ~(846 of 1326), then you are a 2:1 dog with about 15% of the hands you could be dealt, making your push overall EV- with those. Run some #s out on another 15% or so of hands and those come close to EV- as well. Right there you're down to pushing with only top 70% or so of hands

Overall, it's a great article. It's conclusion of "Be more aggressive cause when you run the math it pays off more than you think" is dead on. But it takes it too far.

--Greg

[AleoMagus]

This nifty little peice of software can do sims with a defined range of hands

Poker Calculator (http://koti.mbnet.fi/jraevaar/pokercalculator/)

Regards
Brad S

[wmajik]

I will accept responsibility and have noted that I need to reword my article to address the significant difference between a 'random hand' vs 'any two'. ZJ is right in that a random such as 23o, can be a long term loser. The only reason a 'any two cards' shows positive return, is because the average return of all hands (AA to 23o) eventually average out to a positive number.

A more realistic (and optimal) strategy dictates you lower or raise your push requirements based on your opponent. If your opponent is very tight however, I would still make an argument for pushing 23o when he's only going to call you 1 out of every 20 hands (5%). My definition of a 'tight' player only calls 1 out of 7 hands (15%) as well, so while excersing tighter push requirements probably isn't wrong, the "It's not what I hold, but what my opponent doesn't hold" principle is still strong. Trouble appears when you are faced with a 'semi-tight' or 'loose' player - and thus can require you to adapt your push requirements. More on that in a bit.

At first, this article started out on a napkin with me wondering if K2 was a marginal hand to stealing with in the SB/BB situation. My 'hmm..' response emerged when I found out exactly how often someone is actually willing to fold a hand, especially a realistically tight player. So I went on and did analysis of Q/x, J/x and so on, and found that averaged out over all hands (hence, the 'random' hand), there was a lot of value in raising, simply for fold equity. So, I was able to convince myself that K2 is a very reasonable hand to push in this situation, along with many other hands. At the very worst then, from reading this article, you should see that marginal hands can and should be pushed in certain situations. That alone, I believe, should improve the game of players who otherwise would not have taken advantage of these situations.

The best thing to do is view the article as theory (hence, the 'theorem' title) and extrapolate on it. Originally, I had written it very minimalistic with numbers and concepts. My mistake, I now recognize, was trying to flesh it out while not completely following through.

As part of the theory, I made a chart for hands that varying opponents are willing to play for a 3BB raise. I re-used this chart even while converting to a 10BB push for consistency sake. Though this chart remained the same, I cannot state enough the significance that a push has on your opponent's calling requirements. A number of posters have pointed this out as well.

Realistically, I do not believe even a 'loose' player is calling 1000 chips with T9 to defend 100 in chips. Any reasonable player isn't calling with A2, KJ, 22 or other marginal hands either. On a bubble situation, reasonable players will tighten their requirements even more, for the precise reason that Linus and Eastbay advocate - ICM.

Though I appear to stand alone on this one (and thus will probably have to re-evaluate my position when I have time), my advocacy for increasing aggression on the bubble ironically enough stems from this exact concept. Against any tight opponent on the bubble, they are not only unwilling to call a push, but unwilling to risk their chips on the bubble.

Take for instance, if one of us were sitting in the BB on the bubble and a seemingly solid player pushes all-in from the SB, our play would be glaringly obvious - fold. If he pulls this off 4 times in a row, we'll start to become the wiser and hold our breath when anticipate the next crying call, but by then, the villain will be up 6BB from the last 4 encounters and likely no longer in the 10BB or less 'attack zone', so to speak. If the villain then starts folding on a semi-regular basis, you might just assume that he had a hot run of cards.

So against a tight player, I think this strategy works well from many angles. The strength comes from the fact that if he's not as tight as you want him to be for reasons of folding equity, your pushing should force him to play tight.

An example of a incorrect application of this strategy however, would be confronting a large stack or confronting a loose player. Especially bad would be a loose player on the bubble who has no concept of 'bubble'. As the article tries to illustrate, the equity is in your opponents folding, as opposed to calling (as you always lose EV when you get called).

The thing about loose players however, is that everyone already knows how to beat a loose player. It's straight up standard STT strategy - you wait for a good hand and hope your loose opponent calls with a dominated hand. You let them hang themselves with bad plays. If you both have big stacks, you see flops and out-poker them as well. Everyone talks about it here day in and day out.

I'm sure many of you, upon reaching the final 3 in a STT will groan if you have three solid players in the game. Why? Because you know it's going to be coin-flips and tough flops. If it's loose players (depending on their stack) it's usually better, because you know you've got a skill edge on them.

Thus, when I keep saying 'for those who know what they are doing' - what I'm really saying is that if you know how to beat loose players already, this strategy will help you beat tight players as well; and if they aren't tight, this can force them to play tight.

Anyhoo, like the election, everyone has an opinion already, so I'm probably just talking my way into oblivion by now. Thanks everyone for the input, I will be modifying my article as a result of the critiques I have heard.

And um, vote Kerry. Yeah... /images/graemlins/smile.gif

[AleoMagus]

Ok, I'm back...

In looking at this some more, I think there are some real conclusions that can be drawn about what kinds of hands can be used to steal, even when not entirely clear on what the BB (or limper if you are the BB) will call with. As long as you can narrow him down to a reasonable calling standard like:

Pocket Pairs, A7+, KJ+, Suited Broadway.

This is actually pretty loose in my opinion, and I think as long as you do not have know your opponent to be a total manic, and as long as you have not been robbing him blind,
you can probably assume he'll be this tight or tighter when facing elimination (10XBB) to save his blind.

Anyways, this turns out to be 17.9% of hands, so using the figures Linus came up with (ICM) that put a $ value on the cost of playing and the $EV of winning, we get:

Fold equity = Blinds*Fold%
=$2.24*82.1
=$1.83

Caught Equity(CEQ) = (Win% * Pot) - Total Cost to Play
= ( <font color="red"> X </font> * $39.36) - $25.07
where X is the % of the time we will still win if faced with a call

Caught Result = Caught% * Caught Equity
= 17.9% * (CEQ)

Total Equity = Fold Equity + Caught Profit
= $1.83-(CEQ)

So what this means is that CEQ needs to be greater than -$1.83 for our steal to be profitable.

Solving for the %Win we need when called gives us:

$1.83-(((X*$39.36)-$25.07)*17.1%) &gt; 0
-(((X*39.36)-25.07)*.171) &gt; -1.83
-((X*39.36)-25.07) &gt; -1.83/.171
-((X*39.36)-25.07) &gt; -10.701
-(X*39.36) &gt; 25.07-10.701
-(X*39.36) &gt; 14.369
X &gt; 14.369/39.36
X &gt; .365

Or in other words, we need to win 36.5% of the time when we are called in order for this to be profitable. This is where we can set exact stealing standards

According to Poker Calculator, against that range of hands, the following will win at least 36% of the time when called:


A3+ (A2 just misses the cut)
K4s+
KT+
Q6s
QT+
J8s+
JT+
T7s+
T9 just misses the cut, as do all other nonsuited connectors
97s, 98s
86s, 87s
76s, 65s

So, lots of hands (33%), but hardly any two. Many might be stealing with this many or more already.

A point to be made though which is VERY important to this discussion is that this is not any kind of 'SB pushing guideline'. I say this not because I think this will get you in trouble (though no doubt, it could) but more so becasuse I think there are many many situations where you could push with way more. Many many situations. If for example, we use MJ's original 'tight' profile which calls only 15% the range of playable hands increases dramatically. Surely all nonsuited connectors would play, as would many other hands which just miss the 36% mark.

Another time you could greatly expand this is a situation where there is a really short stack at the table and players will tighten up until he busts

If you were playing someone who only called with TT+, AT+ and KQ (8.2% of hands) then you could steal profitably with any two .

Interesting to note is that many times on this forum, it has been recommended that this 8.2% calling standard be employed for these kinds of all-ins precicely because of ICM's bubble implications. Looser calls mean a confrontation that will usually be unprofitable to both parties in the showdown because the players not in the hands are gaining $EV. This of course is based on the assumption that the pusher has a pocket pair, A7+, or a big king.

What is most interesting to me of all about this steal strategy is it's effectiveness against good players who will not read you for bluffing. If you are using against other good players who read you as also being a good player (some of our 2+2 tourneys might be described this way), then I believe this stratgey could be employed to the fullest of it's implications (at least once or twice before it started getting suspicious).

I think back to my initial guide for beating the party 10+1 for example. In it I recommend calling big all-ins only with AK, AA, KK, QQ. That strategy will get absolutely raped on the bubble against this play. I personally, later softened to a flexible TT+, ATish+ call guideline but even that will lose to this play.

It is just too hard to defend against without sacrificing $EV to the folded players.

Regards
Brad S

[KJ o]

Great, thanks!

[RacersEdge]

I like the approach of this article - but the application is in question.

How easily does it extend beyond the case of being folded to in the small blind? What happens if you take the next most complex case - being folded to in the button? Now you have a "geometric effect" for your folding equity. And even if it held up there - how big a weighting can it have at least at the PP 10+1 games where even being folded to in the button can be rare?

Again, though, I do like the methodology - probably can be extended to other situations.

[AleoMagus]

[ QUOTE ]
How easily does it extend beyond the case of being folded to in the small blind? What happens if you take the next most complex case - being folded to in the button?

[/ QUOTE ]

While certainly more complicated, I think a similar methodology can be used to calculate which hands can be considered +EV to steal with on the button. Previous calculations really just weighed EV of three different possibilities

1) Stealing from the BB (based on the prob. that he will fold)
2) Getting caught and winning (based on prob. of being caught and prob. of winning hand if caught)
3) Getting caught and losing (based on prob. of getting caught and prob. of losing hand if caught)

Together, this will give a kind of equity threshold where we can say "if opponent in the BB will call Y% of hands, we need to only steal if we will win X% of the time against that range."
In an earlier post, I detailed exactly what I though those hands were if an opponent called 17.9% of the time (Pairs, Suited broadway cards, AK-AT, KQ)

So, with two opponents we need to consider more possibilities

1) we steal from both
2) we get caught by SB and lose
3) we get caught by SB and win
4) we get caught by BB and lose
5) we get caught by BB and win
6) we get caught by both and lose to either
7) we get caught and beat both

each of these possibilities will depend on exactly what range of hands we expect to get a call from, and each will also depend on new $EV figures that will come from the fact that we are stealing more blinds this way (1.5 instead of 1). We will also need to consider things like the case where we are caught by both. If we are caught by the SB, then the BB will probably tighten to the very tightest 5-10% at most.

All of this is possible to calculate, and just as I managed to give precise steal hands when assuming an opponent would call 17.9%, it should be possible to give precise steal hands when assuming either opponent will call this much (or with whatever frequency(s) we want)

For any situation, it should be possible to expand our calculations to assume even more opponents when considering the EV of a steal. For example, if we include even more possibilities, we should be able to figure this out even for the UTG player in a bubble situation.

What gets very artificial about all of this is that it assumes precise calling standards for your opponents. This is useful, but is hard to make very confident judgements about. There is also the fact that these calculations assume very close to equal stacks.

At any rate, doing such calculations would give at least a reasonable idea of what we should be stealing with against certain kinds of players.

Many will probably tell me I'm wasting my time, but I think I will probably go ahead and make these very calculations in the next while (at least for the button anyways). Don't get me wrong, many of us have played enough to have a really good idea of this anyways, and it definitely varies from SNG to SNG for all the reasons I have already described (different calling standards and stack sizes).
Still, I want to see the results, probably just beacause of my surprise about some of the results in the SB/BB scenario.

We'll see.

Regards
Brad S

Oh... one more thing which is actually very important. I brought this up in another post lately and I think it may be the real reason why these ideas about stealing may be ultimately flawed.

Just because something is +EV, that doesn't mean it is the best option.

I and MJ have argued about the fact that these kinds of plays are +EV ($EV and CEV), but that doesn't mean anything until we compare them with their alternatives.

[rbilabronze]

I have a question about how the ICM calculation is being computed in this thread. When I plug the numbers into the ICM calculator I am getting the following numbers:

Stack Size: ICM $EV Calc:
2375 0.24805509042797175
2250 0.23933185076560756
2500 0.2563065294032103
2500 0.2563065294032103

The way LinusKS is calcuating the ICM he is getting .2507 for the 2500 stacks. The only thing I can think of that is causing the difference is the treatment of the money in the blinds. The way I am calculating it I am ignoring the blinds which are already in the pot.

Am I using the calculator incorrectly? Any help would be appreciated.

[JoeTable]

does anyone have a link to this article?

[BigDave]

Blind Stealing Article (http://teamfu.freeshell.org/tournament/theorem_blind_stealing.html)

[BigDave]

I'll try again here.

Article (http://teamfu.freeshell.org/tournament/theorem_blind_stealing.html)

[JoeTable]

thanks

[AtticusFinch]

So here's a question about the general assumptions. How do you determine how "loose" a player is as defined in this article within the short space of time you have to get a read on them in a SNG?

A player who plays a lot of hands early may well (and often does) tighten up significantly on the bubble , and there may not be time to assess their "push calling looseness" (vs. preflop looseness) before a critical decision point comes.

Is semi-tight a reasonable default classification for players on which you don't have much data?

I'm looking at a particular example in a tourney I'm in right now. There's a player with whom i've played 200 hands. His VPIP is 35%, but his fold-to-steal percentage is 100% for both SB and BB. Last time I played this guy I took his BB 5 times in a row, 3 of those with absolute garbage. But if it were the first time I had played him, I might have seen him as "loose" and used the incorrect strategy.

On a related note, has anyone thought much about how to apply these same principles to your own calling standards? I've seen it alluded to, but a more comprehensive treatment would be interesting.

[AA suited]

so if are the BB w/less than 10BB, and someone raises, then you should re-raise all-in/call all-in if you have top 67% since you have the odds?