Blind Defense - short handed.

Jim Morgan · #1 11-03-2005, 10:19 PM

Lately I have been playing a fair amount of short handed O8 online. I think I am playing pretty well in general, but I really believe that I don't understand proper play in the blinds. In particular, I don't understand heads up play of "bad" hands.

Here is one such eye-opener...

5c 2c Ad 4h 0.474
7s 5s 7c 6d 0.526

I am really surprised just how much
Changing the 5c to the Kc still only makes the "good" hand a 52-48 favorite.

Using twodimss, I am seeing lots of strange things, but for now let me stick with one example big blind hand.

Suppose I am in the BB with 7s Tc 4d Th
Suppose all fold to the button who raises
Suppose SB folds.
Suppose his hand is a "good" O8 hand.

As 4c 9d 2h 0.486
7s Tc 4d Th 0.514

What the hell? This Piece of dog doo is a favoite?

Now here are some flops...

Flop 1: 8c Kd 3h

As 4c 9d 2h 0.452
7s Tc 4d Th 0.548

Or... Flop 2:
3c 3d 8s

As 4c 9d 2h 0.436
7s Tc 4d Th 0.564

Or... Flop 3
2d 2s 8s

As 4c 9d 2h 0.844
7s Tc 4d Th 0.156

What we have here is a really nasty situation.
When I don't flop a set with my TT, there are a LOT of
flops where I am in decent shape at this point and
yet my hand looks like it totally sucks. And in the
third example, it actually does really suck. If my
opponent had a K instead of a 9 on that first flop then
clearly I am also WAY behind

4c Ad Kh 2h 0.876
7s Tc 4d Th 0.124

As I am looking into this, I am running into a common theme I see in most if not all forms of poker. (and it is one that I still don't know how to handle). This hand seems like it is going to be faced with being a coin-flip on many flops, but that on other flops it will be significantly behind.

I have never understood how to handle such situations. It
seems like folding this hand preflop in order to avoid this nasty dilema is too extreme. If my pre-flop pot equity were typically much lower (say .0.35) then I might simply need to give up even though the pot is offereing me 3.5 to 1 (actually less due to rake).

It can get even nastier.... Suppose the flop comes like
this... 8c Kd Qh

4c Ad 9h 2h 0.289
7s Tc 4d Th 0.711

Now I am big favorite against A249 and a huge dog against
A24K. If faced with a bet, folding TT against the first of these 2 hands is a big mistake.

My instinct is to simply bet out in order to get some idea of where I stand. However, there is a problem. A lot of folks will call here with most if not all hands. Rightly or wrongly they want to see the turn. The problem here is that it is pretty tough to keep leading at the pot with TT74. The call by A249, though technically inferior, has now made my life miserable. If the turn is a three (or any other non-pairing low) my opponent will now have the proper pot odds to call. If he spikes an Ace, I am more or less dead.
And if no low card comes or the board pairs, I am WAY ahead, but I can hardly be expected to know it.

So.... all of the meandering analysis seems to support the idea that playing TT74 heads up from the BB is a real bitch when you miss (and you usually will)

However, with an EV that is often at or near even, simply mucking seems like a big mistake.

Is the combination of poor position and the frequent situation of being "coin flip or way behind" so strong that we must fold away this much equity, or are there some guidelines to help play these hands profitably?

In closing, here is another matchup that blows my mind...

Ts Jc 3d 2h 0.475
6s 4c 2d Ah 0.525

Personally I detest hands with 2 paints and 2 lows. This one is the best of the bunch, and I still hate it for heads up. And yet, a far nicer looking hand is barely ahead.

Crazy game....

Jim

Buzz · #2 11-03-2005, 11:59 PM

[ QUOTE ]
Crazy game....

[/ QUOTE ]

Jim - Agreed.
_
I originally wondered if 23TJn was a better starting hand than 23QKn since 23TJn makes more nut straights than 23QKn. But it isn’t. When you are playing a hand with two wheel cards and two honor cards heads-up, the simple rule seems to be: the higher the honor cards, the better the starting hand. Pairs are better than two different ranks for high. And obviously you generally prefer to have a hand with an ace.
_
I have
• 23TJn at 47.49% for one-on-one play and
• 23QKn at 50.11% for one-on-one play.
Both are for 10000 deals against random hands and random boards, using Wilson. Note that these are both rainbow hands. Hands do better if they're single suited, and even better if double suited. Replacing the deuce and trey with other non-paired wheel cards doesn't make as much difference as suitedness. That is, 45TJd is a better heads-up starting hand than 23QKn. (But 45QKd is better than either).
_
Recently I ran a whole series of sims with gradually decreasing numbers of players. I was looking for hands that did better short handed and one-on-one than they did in full games. To make the comparison, I multiplied the winning totals I got by the number of players.
_
There are some hands that do better one-on-one than in a full game, but they're mostly hands that don't do very well in full game play and they only turn out to be a bit above average in one-on-one play.
_
What happens mostly is hands get closer in value when played one-on-one against random hands.
The best one-on-one hand is still AA23d, but what a waste to play that fine hand heads-up, at least in a limit game! (And of course then the flop is 999, you'd prefer to be holding anything with a nine).
_
I fiddled with twodimes.net for a while before I started using Wilson. One of the advantages of Wilson is that you can run partial hands or run a hand against unknown (and random) cards instead of always making up particular cards for an opponent.
_
Anyhow, interesting post. I'll look it over some more and might (or might not) reply again to it.

Buzz

gergery · #3 11-04-2005, 05:23 PM

EV ala twodimes only tells you how often you’ll win the hand or parts of it.

It doesn’t tell you how much you’ll win when you win, and how much you’ll lose when you lose. The reason A245 will be MUCH better than 7765 is that you are much less likely to make wrong decisions postflop

Wintermute · #4 11-04-2005, 05:33 PM

I agree with what Buzz and Gerg wrote. Another issue is postflop bettability. Many of these interesting EV matchups will never play out because the flop will allow an A2xx hand to bet liberally while a 5677 hand will feel great pressure to fold. If we got all-in preflop every time, then sure 5677 would be a more reasonable playing hand, but the way the game is structured it's junk because you'll have a hard time doing anything post-flop with it.

For example, A2QK rainbow could easily bet a 38J flop with two to a suit, while 5677 would take a look at that flop and get the hell out of dodge.

Yet, 2dimes tells us:

Omaha Hi/Low 8-or-better: 820 enumerated boards containing Js 8s 3h
cards scoop HIwin HIlos HItie LOwin LOlos LOtie EV
Ks Ac 2d Qh 193 325 495 0 342 72 0 0.488
7s 7c 6d 5h 213 495 325 0 225 342 0 0.512

So these 2dimes results have to be taken w/ a grain of salt.

gergery · #5 11-04-2005, 05:47 PM

[ QUOTE ]
I agree with what Buzz and Gerg wrote. Another issue is postflop bettability. Many of these interesting EV matchups will never play out because the flop will allow an A2xx hand to bet liberally while a 5677 hand will feel great pressure to fold. If we got all-in preflop every time, then sure 5677 would be a more reasonable playing hand, but the way the game is structured it's junk because you'll have a hard time doing anything post-flop with it.

For example, A2QK rainbow could easily bet a 38J flop with two to a suit, while 5677 would take a look at that flop and get the hell out of dodge.

Yet, 2dimes tells us:

Omaha Hi/Low 8-or-better: 820 enumerated boards containing Js 8s 3h
cards scoop HIwin HIlos HItie LOwin LOlos LOtie EV
Ks Ac 2d Qh 193 325 495 0 342 72 0 0.488
7s 7c 6d 5h 213 495 325 0 225 342 0 0.512

So these 2dimes results have to be taken w/ a grain of salt.

[/ QUOTE ]

Yet this is also why "maniacs" like ZAPPUD and others can do better than we expect them to.

Wintermute · #6 11-04-2005, 05:53 PM

[ QUOTE ]
[ QUOTE ]
I agree with what Buzz and Gerg wrote. Another issue is postflop bettability. Many of these interesting EV matchups will never play out because the flop will allow an A2xx hand to bet liberally while a 5677 hand will feel great pressure to fold. If we got all-in preflop every time, then sure 5677 would be a more reasonable playing hand, but the way the game is structured it's junk because you'll have a hard time doing anything post-flop with it.

For example, A2QK rainbow could easily bet a 38J flop with two to a suit, while 5677 would take a look at that flop and get the hell out of dodge.

Yet, 2dimes tells us:

Omaha Hi/Low 8-or-better: 820 enumerated boards containing Js 8s 3h
cards scoop HIwin HIlos HItie LOwin LOlos LOtie EV
Ks Ac 2d Qh 193 325 495 0 342 72 0 0.488
7s 7c 6d 5h 213 495 325 0 225 342 0 0.512

So these 2dimes results have to be taken w/ a grain of salt.

[/ QUOTE ]

Yet this is also why "maniacs" like ZAPPUD and others can do better than we expect them to.

[/ QUOTE ]
Yes, but the jury's still out on ZAP and other like him, IMO. He's taking some huge swings lately (110 BB/100 std dev) and is a losing player after 12k hands in my database.

The problem with his style as I see it is this... it is true that often he is getting correct odds to call or whatever on the flop, as in this 5677 case. Winning in those instances makes his opponents tilt, probably the biggest factor to his success.

But when he's not getting correct odds to call with his 5677 (say the A2 is A2JK), he's crushed:

Omaha Hi/Low 8-or-better: 820 enumerated boards containing Js 8s 3h
cards scoop HIwin HIlos HItie LOwin LOlos LOtie EV
Ks Ac 2d Jh 443 644 176 0 342 72 0 0.742
7s 7c 6d 5h 46 176 644 0 225 342 0 0.258

Over time, good opponents will make him pay, I think he'll have to adjust to win long term.

But, again, the jury is definitely still out (I'm watching this guy pretty closely).

gergery · #7 11-04-2005, 06:02 PM

Completely agree. Overall I think that style is a losing one.

I’m just pointing out that it is a more successful style than one might initially think, precisely because he can be in fine shape against many hands that look like they should have him crushed.

-g

Wintermute · #8 11-04-2005, 06:08 PM

[ QUOTE ]
Completely agree. Overall I think that style is a losing one.

I’m just pointing out that it is a more successful style than one might initially think, precisely because he can be in fine shape against many hands that look like they should have him crushed.

-g

[/ QUOTE ]
Definitely. I think the question to really mull over is whether it's possible to make this style successful. If you want to get the image/tilt benefits, you have to do this so much that I think it's probably tough to escape -EV in sum. I mean, you can't selectively decide when you're going to call off a stack with 5677 on the flop because you might be a coinflip, right? Seems like an all-or-nothing proposition, and "all" is probably not +EV.

emptyshell · #9 11-04-2005, 06:09 PM

I think it's a winning style if done correctly. I think we are really missing out on opportunities that a weighted-hand probability approach could provide. I must have missed a swath of hands WM caught, but ZAPPUD is still a big winner for the 14k hands in my database (not to say that he's implementing the strategy perfectly.)

Wintermute · #10 11-04-2005, 06:11 PM

He's getting absolutely butt-raped at the 1k/2k levels. 10 buyins at one 1k table the other night. I see a fiery implosion at the end for this guy. But who knows...