... folded to me on the Btn, holding K2s

[fretelöo]

Ok, one thing about blindstealing. I’m new to 6max, so, naturally, that’s one aspect I have to spend time on. Now, since I’m a systematic guy, I just pulled up that pf-chart that’s floating around 2+2, and 2tabled along, strictly by the chart. That chart recommends raising 67s from the Btn, among others. Well, my FR self would regard this as a pretty bold steal…

One table in particular is great. To my immediate left sits a 6/6/15 nit, and next to him a 98/2/0.3 clickadonk. And I start to wonder…

(Of course, that’s exaggeration. But for the purpose of what follows, let’s just assume that SB always folds to my stealraise. Of course, that’s not true, but this will result in the ranges to be adjusted, not the general point being less valid. Let’s further assume that both me and BB are just awesome postflop players. Neither gains an advantage over the other and in the long run, our equities will match those that ‘stove comes up with. This will probably be among the most important issues though: The sick RIOs you face…)

So, I wondered: How strong should my hand have to be, if SB always folds and BB always calls? I’m wagering 2SB to win 2.5SB (1/0.5 blind structure). That’s 2:2.5 = 4:5.
In other words, my equity in such a HU situation has to be 4/9 vs. BB’s 5/9. Firing up ‘stove, we come up with 96o. Ho hum… Not the prettiest of hands, for sure.

If BB calls every single time (and ONLY calls – we come to 3bets later), and if we’re just brilliant postflop players, able to not give up any equity there is on later streets, we’d be able to steal with 96o and break even. In plain figures:

equity win tie pots won pots tied
Hand 0: 55.509% 53.12% 02.39% 13370942340 601068318.00 { random }
Hand 1: 44.491% 42.10% 02.39% 10597789824 601068318.00 { 96o }

0.44491*2.5-.55509*2 = 0.002095 Steal 500 times with 96o and win a SB. [img]/images/graemlins/tongue.gif[/img]

Fine. But of course BB won’t call every time. He’s smart enough to dump his 72o, right?

So, let’s assume he just calls 80% of the time. Well, things just get better for us:

equity win tie pots won pots tied
Hand 0: 59.594% 57.52% 02.07% 11570815596 417136818.00 { 22+, A2s+, K2s+, Q2s+, J2s+, T2s+, 92s+, 82s+, 73s+, 62s+, 52s+, 43s, A2o+, K2o+, Q2o+, J3o+, T5o+, 95o+, 85o+, 75o+, 65o, 54o }
Hand 1: 40.406% 38.33% 02.07% 7711058160 417136818.00 { 96o }


0.2*1.5+0.8*(.40406*2.5-.59594*2) = 0.154616 Steal 10 times and win a SB.

And so on. If he folds even further (50%), we keep making $$$.


equity win tie pots won pots tied
Hand 0: 65.010% 63.88% 01.13% 8046609564 141901938.00 { 33+, A2s+, K2s+, Q2s+, J4s+, T6s+, 96s+, 86s+, 76s, 65s, A2o+, K5o+, Q7o+, J7o+, T8o+, 98o }
Hand 1: 34.990% 33.86% 01.13% 4265294784 141901938.00 { 96o }

0.5*1.5+ 0.5*(.34990*2.5-.65010*2) = 0.537275 Steal two times and win a SB.

Now, let’s assume he starts realizing what we’re doing and 3bets us liberally. Let’s give him a very liberal 3betting range of 30%. So, knowing that we’re suicidal enough to steal with 96o, he 3bets his top 30%, calls his top 70% and mucks the rest.

So, that leaves us with a bit more complicated calculation (and I leave the ‘stoves out – just trust me on the percentages – yes, I left the top 30% out of his calling range)

0.3* (0.32011*3.5- 0.67989*3) + 0.4*(0.43087*2.5-0.56913*2)+0.3*1.5 = 0.1487805 Steal 10 times again.

Now, let’s try figuring SB back into this. Things get a bit more interesting. First of all, SB will never be able to call/3bet us as lightly as BB could previously because he has to take into account that BB is yet to act and might have AA. Also, his share in the pot is smaller. While BB has 1SB already invested and stands to win 3.5SB for a further 1SB investment, SB has to invest 1.5 SB to win the same amount – while still being OOP to (potentially) two players.
Thus, while I have no idea how to figure this out mathematically, I guess it’s relatively obvious that SB is probably best off to enter with a 3bet himself if he choses to enter the pot at all. He hopes to push out BB to be HU with someone who steals with any two rags.

Combining those two considerations, he should probably tighten up a compared to the 3betting range of BB because he still has to fear BB joining the fun. He has to make sure that he has a REAL hand against the two of us – if it comes to that – and he also has to make enough the times he enters the pot to make the increased investment worth it. Something like a 20% range looks about right. { 66+, A7s+, K8s+, Q9s+, JTs, A7o+, K9o+, QTo+ }

Now, things get iffy. BB will have a monster every now and then. Seeing that he knows we steal with almost any two, and that SB will therefore relatively lightly iso-raise, his capping range is likely relatively broad. Let’s say 10%? { 88+, A9s+, KTs+, QTs+, AJo+, KQo }

Calculations skyrocket now… Just to wrap our head around the numbers once more: We steal with 96o. SB folds 80% and 3bets 20%. Of the 80% he folds, (of those 100%) BB will 3bet us 30%, call 70% and muck 30%. If SB 3bets, BB will cap 10% and muck otherwise.
We already had the first part of that. We just have to make sure that we count this only as 80% now.

Thus:

0.8* (0.3* (0.32011*3.5- 0.67989*3) + 0.4*(0.43087*2.5-0.56913*2)+0.3*1.5) + …

Now we have to figure out if we should even call the cap, if it comes to that. When the action comes to us, there’ll be 9SB in the pot, 2 for us to call and SB calling for sure as well. Thus, it’s two to win ten for us.

444,502,624,104 games 1474.406 secs 301,479,120 games/sec

Board:
Dead:

equity win tie pots won pots tied
Hand 0: 45.249% 43.87% 01.38% 194982396564 6150078826.00 { 88+, A9s+, KTs+, QTs+, AJo+, KQo }
Hand 1: 33.858% 32.39% 01.46% 143993143392 6507000766.00 { 66+, A4s+, K8s+, Q9s+, J9s+, T9s, A9o+, KTo+, QTo+, JTo }
Hand 2: 20.893% 20.62% 00.27% 91652559144 1217445412.00 { 96o }

Well, it seems it’s just barely break even for us to call this raise. So, for fun’s sake, let’s say we call. Now on to the 0.2-part

… + 0.2* (0.9*(0.30722*4-0.69278*3)+0.1*(0.20893*10-0.79107*-4))

Putting it all together:

0.8* (0.3* (0.32011*3.5- 0.67989*3) + 0.4*(0.43087*2.5-0.56913*2)+0.3*1.5) + 0.2* (0.9*(0.30722*4-0.69278*3)+0.1*(0.20893*10-0.79107*4)) = -0.054578

Ok, phew. So it seems I don’t have to steal with 96o after all. But is it really THAT close? Just for the fun of it, I’ve done the whole last calculation again for something somewhat better: J7s

0.8*(0.3*(0.37105*3.5-0.62895*3) + 0.4*(0.53125*2.5-0.46875*2) + 0.3*1.5) + 0.2*(0.9*(0.34605*4-0.65395*3) + 0.1(0.22880*10-0.7712*4)) = 0.218

Now I looked (of course, all of this is rather crude methodology) at ‘stoves internal hand ranking: J7s is among the 41.2% highest ranked hands. Before that come hands like T8o, J8o, Q4s, K2s, a lot of really marginal stuff, really. Of those, I’d say I’d dislike playing K2s a lot more than T8o. So, for a final calculation, I just checked the former:

0.8*(0.3*(0.41127*3.5-0.58873*3) + 0.4*(0.54270*2.5-0.45730*2) + 0.3*1.5) + 0.2*(0.9*(0.38496*4-0.61504*3) + 0.1(0.22725*10-0.77275*4)) = 0.294

Bottom line: Blindsteal like crazy! K2s and J7s turned out to be quite decent winners. I guess I just wanted to see that for myself. Ok, move on, nothing to see here... [img]/images/graemlins/crazy.gif[/img]

[BigBadBabar]

lol

[fretelöo]

[ QUOTE ]
lol

[/ QUOTE ]

If that's all you have to say, I guess I have to work on my punch lines... [img]/images/graemlins/wink.gif[/img]

[Watkins]

Fritos,

You should read the Stox book, blind stealing and defence are covered in depth. For example, he covers where stealing with K2 is preferable to 8T and visa versa.

Looking at these things on your own will really help to solidify stuff in your own mind but he will put you further down the road at a quicker pace.

[Xhad]

fretloo, a lot of this equity stuff is pointless because most hands will not go to showdown. I've posted hands in the past where I 3bet a possible steal-raise with 98s, where I'd never dream of doing such a thing with A4o.

[BigBadBabar]

sorry fret, here's a more reasoned response:

[BigBadBabar]

just joking with you. i think what your math is showing is something that a lot of players feel either intuitively or with lots of experience; that certain sets of hands are easier/harder to play in certain positions against certain opponents. and you're also right that probably all of us can steal the blinds more because usually it's +ev to steal with any two or almost, and if you find the right guys who are too tight, raising with any two until played back at is the right play. i do feel more comfortable raising t8s from the co than raising q5o even though q5o is probably a better hand (just guessing here, didn't stove it), because it's easier to play postflop, which is what xhad touched on.

[ESKiMO-SiCKNE5S]

generally raising here depends on what happened the last orbit or 2

[fretelöo]

[ QUOTE ]
fretloo, a lot of this equity stuff is pointless because most hands will not go to showdown. I've posted hands in the past where I 3bet a possible steal-raise with 98s, where I'd never dream of doing such a thing with A4o.

[/ QUOTE ]

Xhad,

I agree. And to take this into consideration, I stipulated that both BB and me are awesome postflop players (of course, if he's that awesome, he wouldn't have the stats he has but anyway).

Of course, all of this doesn't take into considerations that BB might just get pissed and resteal with 72o or whatnot. Or that we might fold an ui Q5o to BBs ui Q4o etc.
But I think, that in the long run, all of that should equal out to some extend. There'll be spots when I fold the K-high winner, there'll be spots when he folds his 22 vs my K-high c/r on a AQJ flop. Ultimately, 'stove calcs can be translated into "connects good/well/better/ with a random board than..." and I think if you take them with that grain of salt, I think they're a reasonable guesstimate.

[marchron]

I will almost always openraise any Kx from the button, as it stands to be the best hand remaining vs. two blind hands.

If the blinds, especially the BB, will defend with anything and are generally passive, you can get away with openlimping a lot of hands, even things like 22 or Q8s, because the lack of implied odds these hands suffer since the pot won't be multiway is offset by the fact that you can usually see all five cards for free, and if the fish bets you know instantly where you stand.