Thoughts on playing "unexploitably" (long & theoretical)

[DonT77]

Just a thought that I'd like to get your opinions on-

My thought is that if you play too far to either side of 'normal' that your play can be exploitable.

Here is what I mean:

Say you raise too often from LP, then you will invite resteals.

Say you play too tight in EP, you will be playing with your cards face up.

Say you C-bet too often, you will get check-raised a lot.

Say you bluff too much, you will get called down when you are behind.

And so on...

IOW, if you play any situation too far to one side of 'normal', then it creates a situation that can be exploited by your opponents. So obviously you want to mix up your play - to be "unexploitable".

Now I realize that this is probably not especially enlightening to a lot of you, but I want to take this line of thinking a step further.

Here are some general questions on trying to play "unexploitably" to invoke discussion-
1. How do we define what is 'normal'? (I know Sklansky mentions optimal bluffing/calling% in TOP, but this doesn't apply to all situations.)
2. How do we identify exploitable tendancies in our opponents? (I'm pretty sure that any large deviation from what we define as 'normal' is exploitable. If this is the case, then is 35/20/5 more exploitable or is 5/4/0.5 more exploitable?)
3. If we have raised from LP and won the blinds 4x in a row should we raise again with A9o or 55 when it gets folded to us in LP this orbit when we are pretty sure somebody is ready to resteal-push on us?
4. If we have folded 40 times in a row in EP, do we have to play 27o to avoid playing with our hand face-up next orbit when we get AA?
5. Say you've C-bet the flop the last 7 pots you've been in and you feel a CR coming on, do you C-bet the flop with AJ on a JT4-2suited flop or do you take one off and give up a free card?
6. Say you've been caught bluffing (or showed bluffs) twice in the last 3-4 orbits but now a flush card comes on the river and you are sure that your opponent doesn't have a flush and that he probably has you beat with ~3rd pair. Do you try to bluff him off his hand again?
7. If we have folded our BB 6 times in a row to an LP raise do we have to defend or attempt a resteal with 27o to get our opponents to back off?

I know the answer to most of these is "it depends" - I guess the more general question is how much do you consider your future and perceived exploitability in situations like these?

[NoahSD]

Nice post.. I wish you'd post more.

[ QUOTE ]

1. How do we define what is 'normal'? (I know Sklansky mentions optimal bluffing/calling% in TOP, but this doesn't apply to all situations.)


[/ QUOTE ]

Like you mentioned, in some spots it's easy to find the exact unexploitable strategy. For example, HU with < 9 BBs is pretty easy to solve--you could probably do it in 5 or 6 hours with just twodimes and a calculator if you felt like it, and with a computer program it takes a few seconds.

Other spots are really complicated, like any preflop situation with > 20 BBs and more than 2 people. I have no idea what to think about this except that there are degrees of exploitability and I can lessen my exploitable play by not, for example, raising A2o UTG+1 9-handed with 20 BBs against good players. When I think about AJoish hands in the same spot, I just content myself with the fact that neither raising nor folding is horribly exploitable.

[ QUOTE ]

2. How do we identify exploitable tendancies in our opponents? (I'm pretty sure that any large deviation from what we define as 'normal' is exploitable. If this is the case, then is 35/20/5 more exploitable or is 5/4/0.5 more exploitable?)


[/ QUOTE ]

Good question. I don't really know.

[ QUOTE ]

3. If we have raised from LP and won the blinds 4x in a row should we raise again with A9o or 55 when it gets folded to us in LP this orbit when we are pretty sure somebody is ready to resteal-push on us?


[/ QUOTE ]

This is a very interesting question.

Basically, when we're playing unexploitably our goal is to have a smart player who knows exactly what we do with every hand in every spot have no edge on us. When we do this, a perfect player will respond by playing unexploitably himself.

However, when a perfect player thinks we're playing exploitably, he will shift his strategy to exploit us. Thus, he will be playing unexploitably and we should adapt our strategy to exploit this.

[ QUOTE ]

4. If we have folded 40 times in a row in EP, do we have to play 27o to avoid playing with our hand face-up next orbit when we get AA?


[/ QUOTE ]

Only if we think our opponents have adjusted enough to make raising 27o profitable there.

[ QUOTE ]

5. Say you've C-bet the flop the last 7 pots you've been in and you feel a CR coming on, do you C-bet the flop with AJ on a JT4-2suited flop or do you take one off and give up a free card?


[/ QUOTE ]

This depends a ton on the exact numbers. If they're bluffing enough that you have >50% equity against their c/ring range, then bet. (The actual number depends on how the turn and river will play too and how many outs hands that fold have, but that's a rough estimate.)

[ QUOTE ]

6. Say you've been caught bluffing (or showed bluffs) twice in the last 3-4 orbits but now a flush card comes on the river and you are sure that your opponent doesn't have a flush and that he probably has you beat with ~3rd pair. Do you try to bluff him off his hand again?


[/ QUOTE ]

Usually not.

[ QUOTE ]

7. If we have folded our BB 6 times in a row to an LP raise do we have to defend or attempt a resteal with 27o to get our opponents to back off?


[/ QUOTE ]

No.. if they're raising our blind too much, then they're playing exploitably and we should encourage this.

However, there are other similar situations in which it's correct to make -EV plays to tighten your opponents up. For example, if you're stealing blinds like mad at a table and a few players figure out that you'll fold to a resteal, making -cEV calls can definitely be correct to shut them up.

[A_PLUS]

For some reason, I feel like diving into an extended game theory discussion. I am coming off the end of a 7 hour session, an 8 hour work day, and I am deep ina rebuy, and I havent had a game theory class in 2 years. So I expect a lot of errors, sue me.


This is pretty much where every 'mathy' player wants to get to as a player. It is a pretty straight forward game theory problem. The problem is, for a poker hand with even moderately deep stacks, it is next to impossible to actually solve. I realize that my last statement sounds a little contradictory, but I will try to explain it with some examples.

The easy one

We have 7xBB in the SB (.5/1.00 blinds, no antes) BB has the same

An easy way to attack the problem is in steps. It doesn't really matter wear we start (I dont think).

-So we assume he calls with 100% of hands.
-Given his range, what hands are +EV to push (easy, but time consuming with pokerstove), that is our push range
-But, he knows this (an assumption we make if we are looking for unexploitable plays). So, he adjusts his range using the same math.
-Then we go back to and adjust our pushing range, given his adjusted calling range, etc, etc.

The idea is, eventually we get a pushing range for us, and a calling range for him, where neither of us can improve by playing differently (given the other one plays optimally).

Then we add a million variables
Lets say we are on the flop with AJo, and a 8 9 J rainbow flop.
As long as we have 2-3x the pot behind, the problem gets thrown to SH!^#

The first step alone is impossible to 'solve'. Why? B/c we dont know his range. Think about it. If we use the same methodology. What is our opening range? If we think he plays 100% what hands should we play? We need to look at every combo of pocket cards. Then, we need to figure out his optimal play, given every possible board, and every possible action at every decision point. So basically we have a practically infinite number of possibilities.

So, I can picture HOW to solve for an optimal play, but I cant see how anyone could actually do it.

So whats the point
Well, it is soooo much easier for late tournament play. Our play choices are usually, bet-fold, bet-call, bet-jam, check-fold, check-jam.

So, we could figure out create a solvable problem for something like. "What hands should I call a 3xBB raise with from the BB, with 20xBB".

The problem is still HUGE and worthy of a PHD thesis is economics, but it is something I think I could solve with enough time.

Oh, well. This got a lot less interesting when I just got cold decked for the 4th time in the stars rebuys for another top 100, non-FT finish (my 17th and 18th of last 2 months).

For the rest of your post, most preflop decisions can be solved with a little effort, with < 15xBB. A common application, is most people dont call a resteal anywhere near optimally. Which is why resteals are so powerful. So, you need to know where there mistake is (some players make the mistake of never raise-folding, the only way to exploit them, is to fold hands in your equilibrium pushing rnage). If they fold too much, you push hands outside of your equilibrium range.

FWIW, the only time I actually use game theory in my play is in blind battles. Most of the times, it is by pushing any 2 from the SB, or folding a hand to a push that is 'unexploitable' b/c I know the player is pushing way too tight. Nothing too sexy, and something all good players do intuitively, but I am different, is that I have a very specific equilibrium range that I am using (thanks to Eastbay).

[mornelth]

[ QUOTE ]
However, when a perfect player thinks we're playing exploitably, he will shift his strategy to exploit us. Thus, he will be playing unexploitably and we should adapt our strategy to exploit this.

[/ QUOTE ]

Any strategy that EXPLOITS us is also EXPLOITABLE. If our opponent adjusts to us BEHAVING weak-tight - we can trap him.

By unexploitable strategy we usually understand a strategy where an opponent cannot get a significant edge by KNOWING what our strategy is. For example - when you are HU with a certain stack size at certain blind level - there exists a Nash Equilibrium which dictates what hands you can profitably (or at least break-even) move in with. If you move in with that range - your play is unexploitable. However, it DOES NOT give you an edge over an opponent who understands NE and plays the perfect calling strategy. It just prevents you from being too tight (or too loose) for the current blinds/stack ratio. You get advantage if your opponent deviates from NE and becomes either too tight (folds too much) or too loose (calls too much).

To the OP - If you are raising only QQ+, AQs+ and AKo in EP - you are exploitable (opponent can fold with confidence all but strongest hands). If you add suited connectors in the range 56s - 89s - you are no longer exploitable (or at least LESS exploitable). Your opponent would be wrong to either always call or to always fold. This is all game theory stuff and has been covered to some extent in TOP, why are we going over it again?...

[A_PLUS]

"Thus, he will be playing unexploitably and we should adapt our strategy to exploit this. "
-NoahSD

"Nobody goes there anymore; it's too crowded."
- Yogi Berra

[A_PLUS]

Another quick ramble on actually using this stuff rather than theory.

In tournaments, a common situation is where you call a raise PF, and reraise push, or C/R push on the flop. (result of having 20xBB, not from any tournament theory).

So, you can figure out 'types' of hands that you make this play with. To aproximate the equilibrium (unexploitable), you want to make it so that your opponent is indifferent between calling and folding given his range of hands. (in terms of expectation, not in terms of %)

You use the same methodology, only you start with an assumption for both of your ranges (and perceptions of each others ranges) on the flop. You make it solvable by just grouping flops (quads, set, OE straight draw, etc. ) and figuring out how often you and your opponent flop each given your preflop range.

so, it goes...
-I only push the nuts
-He bets everything, folds to a push
-I adjust and push a few more
-he adjusts and calls with 2nd nuts, etc, etc.

[DonT77]

[ QUOTE ]
This is all game theory stuff and has been covered to some extent in TOP, why are we going over it again?...

[/ QUOTE ]

Some of us skipped class that day? [img]/images/graemlins/blush.gif[/img]

If your advice is to re-read TOP, then thanks - I probably should do that.

OTOH, if you are expecting to see previously undisclosed to mankind advanced poker theory in my posts (or many of the other posts here), you may be disappointed. [img]/images/graemlins/tongue.gif[/img] [img]/images/graemlins/grin.gif[/img]

[NHFunkii]

[ QUOTE ]
"Thus, he will be playing unexploitably and we should adapt our strategy to exploit this. "
-NoahSD

"Nobody goes there anymore; it's too crowded."
- Yogi Berra

[/ QUOTE ]

I believe that was supposed to say exploitably =]
btw that word starts to sound funny after you read it 30 times in one thread