Short handed O8 - maniac isolation tactics?

[BlueBear]

Hi,

The game is $5/$10 party shorted-handed 08, say 4 players, there is a maniac to my direct right who raises 95% hands. Assume the rest of the table is loose with a mix of agg/passive players. When the maniac open-raises, is it wise tactics to re-raise with a somewhat marginal hands (for example, 3499, KK47, KQJT, 25QK, etc...) so that we can get heads-up with maniac with a positional advantage?

There are two approaches to deal with this case that is to just call the maniac and get other players to play along OR to isolate the maniac.

I understand the isolation of a maniac is good tactics in short-handed holdem with any two good/marginal hands. Does the benefits of isolation play also apply for 08 (or should I just give up as people are going to call the 3-bets anyway)?

BlueBear

[chaos]

It depends on the nature of your hand and your other opponents. Some hands play better shorthanded while other would like as many callers as possible. The other opponents may be so loose that they call three bets without hesitation or they may realize that you are running an isolation play.

I think you still need a decent hand to try and isolate. In my experience you are not going to get the maniac to lay down very many hands so you will need to showdown the best hand at the river. Most of the hands you listed (other than KQJT) are sub-marginal - they are bad. I don't think a hand like 3499 will perform that well against a random hand. I would wait for a hand with some redeeming qualities to enter the pot.

Keep in mind that you will need to scoop the pot to make money. If you isolate the maniac and split the pot you will gain very little since you would only be splitting the folded blinds minus the rake.

[Gitz]

Bluebear,

I would move to his right if possible. If not I would tighten up and just call in most positions before the flop. It's going to cost you more money to play marginal hands that you describe that almost need a perfect flop to continue.

paul

[Buzz]

Blue Bear - You have more than just one opponent. I think it is a big mistake to concentrate too much on out-playing one opponent in Omaha-8. “Too much” is the key; of course the overly aggressive opponent deserves your attention. But you need to consider your other opponents too. I realize you already know that, but I think it’s the first consideration, and deserves to be noted.

You always need to choose a tactic that will actually accomplish your objective. If your objective is to get one-on-one with the maniac, and if a raise will accomplish that objective, then you can raise for the purpose of getting one-on-one.

Therefore your second consideration, in my humble opinion, is to establish what you can accomplish with a raise. How a raise will affect your opponent’s actions is very dependent on your opponents, also very dependent on their opinions of you, and also very dependent on their opinions of the interaction between you and the maniac. Raising to get rid of any of them who suspect your motive may not accomplish your objective. So at some point you have to test the waters and then go from there.

The third consideration, in my humble opinion, is the actual cards you hold. Starting hands that tend to make the nuts are wasted in one-on-one play. That’s an overstatement, I suppose, but maybe you catch the drift. What I mean is you want lots of customers chasing with second and third best hands when you make the nuts.

You can’t tell ahead of time if you’re going to make a nut straight-flush, quads, full-house, or straight. For example, a pair of kings generally does better than a pair of eights, but if there’s a pair of fives plus an eight but no king on the board, then the pair of eights is clearly better than the pair of kings... (I’m not suggesting that you include hands that feature a pair of eights in your A-list of starting hands).

However, with a suited hand or a hand with low cards, you should know when you are likely to end up with a nut flush draw or a nut low draw... It’s fairly simplistic. When you have a suited ace and flop a flush draw, you’ll have the nut flush draw. When you have acey deucey and flop a low draw, you’ll have the nut low draw - or if you have ace-trey and flop a low draw that includes a deuce, you’ll have the nut low draw - or if you have deuce-trey and flop a low draw that includes an ace, you’ll have the nut low draw.

Thus the third consideration is that some starting hands do well when one-on-one while other starting hands do better against as many opponents as possible.

My feeling, but I don’t have simulation data to back me up, is that hands like the 3499, KK47, and 25QK you have listed do reasonably well one-on-one against random hands. These hands are all basically two two-card hands with a shot at low and also high. And I personally don't like rainbow hands much (with some exceptions like A234n) - but maybe that's just my hang-up.

For example, in the 3499, the 34 go together for low and the 99 go together for high. Neither 34 nor 99 is a good combination against a full field. But one-on-one the 34 may win for low and the 99 may win for high - especially against a random hand. (There are some other possibilities as well - eg. the 34 may win for high - but this is a long shot).

The KQJT you have listed is a nut making hand. For this one you want customers when you make the nuts... And when you have customers, although Big Dave may disagree with me, I strongly feel you also want the hand to have a possible flush re-draw. People have made a case for raising before the flop with this hand to collect from opponents who play tightly after the flop - just in case the hand has a fit with the flop... I think whether or not you raise before the flop with this hand depends on how many will be calling the raise and how they play after the flop, both when there’s a pre-flop raise and when there isn’t. At any rate, I would not try to isolate with KQJT.

[ QUOTE ]
I understand the isolation of a maniac is good tactics in short-handed holdem with any two good/marginal hands. Does the benefits of isolation play also apply for 08 (or should I just give up as people are going to call the 3-bets anyway)?

[/ QUOTE ]

Pretty hard to say without knowing the individuals involved - and not an easy question even when you do. I think you might be concerned about individuals targeting you and/or the maniac by selectivly calling the 3-bets.

Just my opinion.

Buzz

[Drunk Bob]

Set on his right and check raise him at every oppertunity.

Be prepared for big losses or big wins.


Personally I would look for another table.



But that is just weak tight me.

[Drunk Bob]

Thank you Buzz

[BlueBear]

Hi,

Thanks for all your replies. I guess I got it wrong when trying too hard to isolate with hands that really should isolate. I ran an simulation of those stated hands against any random hand and they don't really perform that well head-ups. I guess I'll be raising preflop mostly for value (and maybe to win the button/position) from now on.

[Buzz]

[ QUOTE ]
I ran an simulation of those stated hands against any random hand and they don't really perform that well head-ups.

[/ QUOTE ]

Blue Bear - My impression is that although some hands are much better than random hands in a full game, no hand dominates random hands one-on-one.

I'd be very interested in seeing the results of your simulations.

Buzz

[chaos]

My gut feeling is that a pair of aces with either a 2 or a 3 would dominate a random hand.

[Phat Mack]

Here's a sim for 9943r against a random hand. It looks as though it's a push, but on examining the results, it might be a good hand to play as long as you flopped a set or higher for high:

<font class="small">Code:</font><hr /><pre> Monte carlo simulation results from Poker Calculator 1.1.4.1
Omaha Hold'em hi/lo 8/b, 100000 combinations tested.

Hand | 9h9s4d3c | xxxx |
------+--------------+--------------+
High | 12203 | 15669 |
Draw | 3061 | 3061 |
Lose | 35976 | 35358 |
Scoop | 35358 | 35976 |
Low | 14679 | 11560 |
------+--------------+--------------+
Win% | 49.6% | 50.4% |
------+--------------+--------------+


9h9s4d3c:
Pair win: 6902 draw: 87 lose: 22149
Two Pair win: 14686 draw: 199 lose: 22597
Three of a Kind win: 10885 draw: 6 lose: 4362
Straight win: 5516 draw: 483 lose: 1593
Full House win: 8581 draw: 19 lose: 943
Quads win: 991 draw: 0 lose: 1


xxxx:
High Card win: 0 draw: 0 lose: 3011
Pair win: 5954 draw: 87 lose: 20872
Two Pair win: 18318 draw: 199 lose: 18536
Three of a Kind win: 6060 draw: 6 lose: 2930
Straight win: 9069 draw: 483 lose: 896
Flush win: 5907 draw: 0 lose: 486
Full House win: 5785 draw: 19 lose: 830
Quads win: 474 draw: 0 lose: 0
Straight Flush win: 78 draw: 0 lose: 0
</pre><hr />

Here's the 9943 hand double suited. It picks up 4%+ in win rate. The flushes it makes seem to be &gt;4:1 favorites. Interesting.

<font class="small">Code:</font><hr /><pre>Monte carlo simulation results from Poker Calculator 1.1.4.1
Omaha Hold'em hi/lo 8/b, 100000 combinations tested.

Hand | 9h9s4s3h | xxxx |
------+--------------+--------------+
High | 13465 | 13665 |
Draw | 2980 | 2980 |
Lose | 31948 | 40261 |
Scoop | 40261 | 31948 |
Low | 12780 | 12568 |
------+--------------+--------------+
Win% | 54.16% | 45.84% |
------+--------------+--------------+


9h9s4s3h:
Pair win: 5862 draw: 69 lose: 17875
Two Pair win: 12659 draw: 132 lose: 19185
Three of a Kind win: 9685 draw: 9 lose: 4005
Straight win: 4711 draw: 429 lose: 1243
Flush win: 11333 draw: 0 lose: 2346
Full House win: 8486 draw: 22 lose: 957
Quads win: 990 draw: 0 lose: 2


xxxx:
High Card win: 0 draw: 0 lose: 2874
Pair win: 4980 draw: 69 lose: 21861
Two Pair win: 15436 draw: 132 lose: 21144
Three of a Kind win: 5123 draw: 9 lose: 3726
Straight win: 7838 draw: 429 lose: 2216
Flush win: 5794 draw: 0 lose: 1127
Full House win: 5831 draw: 22 lose: 778
Quads win: 532 draw: 0 lose: 0
Straight Flush win: 79 draw: 0 lose: 0
</pre><hr />

[iblucky4u2]

[ QUOTE ]
I think you might be concerned about individuals targeting you and/or the maniac by selectivly calling the 3-bets.

[/ QUOTE ]

Buzz,
I think this is the most important thing in your response.
When a manica hits my table, I like to tighten up and nail him whenever I can by being one of the players to selectivly call his raises. I will three bet with hands that play well heads-up but normally, in O/8, premium hands want as many callers as possible. I may not play as many hands, but wind up winning much larger pots when I do enter.

[Phat Mack]

When a manica hits my table, I like to tighten up and nail him whenever I can by being one of the players to selectivly call his raises.

This is one of the reasons that I don't mind a maniac sitting on my left in an O8 game. If a catch a big one, he'll bet it for me and I'll appear to be sucked in...at least 'til late in the hand.

In other formats, I want to sit on a maniac's left so I can re-raise and isolate. But in O8, that's not always the best play.

[Buzz]

[ QUOTE ]
Flush
win: 11333
draw: 0
lose: 2346

[/ QUOTE ]

Hi Mack - Thanks. I love your simulations. Interesting how well a nine high flush does, heads-up, winning better than four times out of five (or almost five times out of six). That would make it enough of a favorite to generally push in a heads-up limit game, hoping to collect from someone who would chase with a weaker hand. But pot limit is a different story.

My original gut feeling was most losses of the nine-high flush would be to full houses (with the board obviously flushed and also paired - which would happen a lot, roughly half the time, when you held
9/images/graemlins/heart.gif 9/images/graemlins/spade.gif 4/images/graemlins/spade.gif 3/images/graemlins/heart.gif, but after some crude approximations, I think most of the losses are actually due to higher flushes.

Buzz

[Buzz]

[ QUOTE ]
My gut feeling is that a pair of aces with either a 2 or a 3 would dominate a random hand.

[/ QUOTE ]

Chaos - Thanks. Good point. Mack could tell us better, but a premium hand like
A/images/graemlins/club.gif A/images/graemlins/diamond.gif 2/images/graemlins/diamond.gif Q/images/graemlins/heart.gif or A/images/graemlins/club.gif A/images/graemlins/diamond.gif 3/images/graemlins/diamond.gif K/images/graemlins/heart.gif
is probably something in the neighborhood of a two to one favorite over a random hand.

Buzz

[BlueBear]

http://momothedog.tripod.com/omaha.gif

The above is my simulation data of various omaha hands against a number of random opponents.

Random win is defined as the percentage the hand I hold wins assuming that all hands are unknown and played to showdown.

Advantage is defined as the percentage win over percentage of a random win. Eg, for my K4sK7o, my advantage over my holding of a random hand is 55.57/50 = 111.14%

The data provides some interesting results.

- Most hands (with the exception of AAxx) are not clear dominating winners heads-up before the flop, and unlike holdem, the overall strength of the hand is VERY dependent on the flop.

- Hands containing A2 perform MUCH better with a high number of callers/limpers.

- K4sK7o is a good hand heads-up but only by very slightly. Hence, in heads-up play, it may not be worth it to raise out of position preflop against a typical loose play.

- KJsQT is actually a weak(er) hand heads-up and performs better with a high number of callers/limpers.

- I must emphasize that these hands are close to unplayable but it is interesting to note that the hands 25oQKs and 9394s have a very unusual property of performing relatively better with a medium number of players. (25oQKs performs well with 5 players, while 93s94s with 7 players). Studying the reason for this would make an interesting future investigation.

Fire away with any comments.

BlueBear

[Phat Mack]

Good guess. AcAd2dQh is close to 2:1 fave:

<font class="small">Code:</font><hr /><pre>Monte carlo simulation results from Poker Calculator 1.1.4.1
Omaha Hold'em hi/lo 8/b, 100000 combinations tested.

Hand | AdAcQh2d | xxxx |
------+--------------+--------------+
High | 11410 | 14865 |
Draw | 2482 | 2482 |
Lose | 19607 | 53428 |
Scoop | 53428 | 19607 |
Low | 14499 | 10283 |
------+--------------+--------------+
Win% | 67.1% | 32.9% |
------+--------------+--------------+


AdAcQh2d:
Pair win: 10530 draw: 147 lose: 16514
Two Pair win: 22236 draw: 234 lose: 12820
Three of a Kind win: 11209 draw: 27 lose: 3221
Straight win: 4032 draw: 255 lose: 1179
Flush win: 6497 draw: 0 lose: 429
Full House win: 9249 draw: 27 lose: 308
Quads win: 1024 draw: 0 lose: 1
Straight Flush win: 61 draw: 0 lose: 0


xxxx:
High Card win: 0 draw: 0 lose: 2697
Pair win: 0 draw: 147 lose: 25581
Two Pair win: 7603 draw: 234 lose: 27971
Three of a Kind win: 5145 draw: 27 lose: 3691
Straight win: 10958 draw: 255 lose: 1866
Flush win: 5010 draw: 0 lose: 1560
Full House win: 5170 draw: 27 lose: 1471
Quads win: 495 draw: 0 lose: 1
Straight Flush win: 91 draw: 0 lose: 0
</pre><hr />

Here's AcAd3dKh:

<font class="small">Code:</font><hr /><pre> Monte carlo simulation results from Poker Calculator 1.1.4.1
Omaha Hold'em hi/lo 8/b, 100000 combinations tested.

Hand | AdAcKh3d | xxxx |
------+--------------+--------------+
High | 11724 | 14110 |
Draw | 2315 | 2315 |
Lose | 19868 | 53648 |
Scoop | 53648 | 19868 |
Low | 13779 | 10677 |
------+--------------+--------------+
Win% | 67.07% | 32.93% |
------+--------------+--------------+


AdAcKh3d:
Pair win: 10612 draw: 124 lose: 16512
Two Pair win: 22378 draw: 225 lose: 12592
Three of a Kind win: 11483 draw: 21 lose: 3124
Straight win: 4155 draw: 252 lose: 987
Flush win: 6409 draw: 0 lose: 420
Full House win: 9244 draw: 28 lose: 343
Quads win: 1028 draw: 0 lose: 0
Straight Flush win: 63 draw: 0 lose: 0


xxxx:
High Card win: 0 draw: 0 lose: 2831
Pair win: 0 draw: 124 lose: 25614
Two Pair win: 7288 draw: 225 lose: 28455
Three of a Kind win: 5009 draw: 21 lose: 3639
Straight win: 10907 draw: 252 lose: 1911
Flush win: 5097 draw: 0 lose: 1530
Full House win: 5043 draw: 28 lose: 1388
Quads win: 525 draw: 0 lose: 4
Straight Flush win: 109 draw: 0 lose: 0
</pre><hr />

The same. I can't read Blue Bear's image, but it might be interesting to see if you make or lose money by trying to get these hands heads up. Probably very dependent on game conditions. I can imagine scenarios where these hands would get paid off more and more readily in a heads up situation.

JMO

[Phat Mack]

Hi, Blue Bear. I can't seem to read your data. I don't know if it's my browser, but I just see a sign that says "Image hosted by Tripod." Clicking on it doesn't seem to do anything.

Advantage is defined as the percentage win over percentage of a random win. Eg, for my K4sK7o, my advantage over my holding of a random hand is 55.57/50 = 111.14%

This confuses me. If a random hand won 50% of the time, and K4sK7o won 55.57%, wouldn't its advantage be 5.57%?

[BlueBear]

My apologies, the reason why the data can't be seen is because I'm posting an external link to the picture. In the meantime before I fix it, it can be seen at

momothedog.tripod.com/omaha.gif (need to paste this into your address manually).

The reason why I divided the percentages up is so that the "relative improvement" of a hand over a random hand can be directly compared.

For example, compare this two extreme cases
Case A: Consider a 4 handed game - the chances of winning without any known cards is 25%. Assume your cards are XXXX and has a winning chance of 35%. Your "relative improvement" over a random hand is 35/25 = 140%
Case B: Consider a 10 handed game - the chances of winning without any known cards is 10%. Assume your cards are YYYY and has a winning chance of 20%. Your "relative improvement" over a random hand is 20/10 = 200%.

Hence, although both hands have the equal improvement over a random hand (10%), the "relative improvement" over a random hand states that hand B has far more improved.

For the purpose of this discussion, relative improvement and advantage are used interchangably.

Mind you, this are just sketchy ideas and I'm not very good in analysing all of this and I'm sure a competent statistician will tear me to pieces. /images/graemlins/smile.gif

[chaos]

This “relative improvement” figure also shows how much you would stand to win.

Assume each player contributes $100 to the pot. In a four handed game, a 35% win rate would net you $400 * .35 = $140 for a $40 profit. In a ten handed game a 20% win rate would net you $1000 * .20 = $200 for a $100 profit.

[BlueBear]

That's a keen and excellent observation.

[Phat Mack]

Very interesting!

93s94s maintains a fairly constant advantage as the number of players increases (peaking at 5-6 opponents), while A2sA3s steadily gains advantage as the number of opponents increases. KK47's advantage steadily decreases.

This seems to support the age-old advice that high hands want the pot heads up.

[Phat Mack]

This seems to support the age-old advice that high hands want the pot heads up.

Well, maybe not. Here are some sims for KsKhQsQh:

<font class="small">Code:</font><hr /><pre> KsKhQsQh advantage
0.5461 109.22
0.3878 116.34
0.3154 126.16
0.2703 135.15
0.2396 143.76
0.2128 148.96
0.2026 162.08
0.1887 169.83 </pre><hr />

[Buzz]

Hi Mack - Thanks. Beautiful work, as always. Very helpful. Much appreciated.

Saying these hands are 2 to 1 favorites over random hands in heads-up competition, while correct, is somewhat misleading. Your simulation data gives a much clearer picture.

Roughly 26% of the time, AcAd2dQh and AdAcKh3d split with random hands. Nothing won, nothing lost.

But the other 74% of the time somebody scoops, and here AcAd2dQh and AdAcKh3d outscore random hands by about 2.7 to 1.

On a slightly different note, it's interesting how similarly well the two hands fare in heads-up play against random hands.

And along still different lines, these simulations gives us a fix on how often a nine high flush loses to a higher flush. (Remember 3499d)? The ace-high flush and the nine high flush must have about the same chance of encountering an opponent with a full house, or quads.

It really is a different ball game, heads-up, compared to a full table.

Highest Regards,

Buzz

Thanks again.

[Buzz]

Wish I could read it too. I get a blank page (except for the logo) when I paste momothedog.tripod.com/omaha.gif
into my web address box.

Buzz