Skip to forum
Notifications
Clear all

Bluffcatching -- how, when and why?

6 Posts
5 Users
1 Reactions
1,761 Views
Super Moderator
VorpalF2F
Joined: 02.09.2010

Here is a fairly typical hand:

Feral Cow Poker Hand Converter
PokerStars Limit 5 Card Draw $0.10/$0.20 - 6 players

UTG: $3.29
UTG+1: $3.62 (Hero)
CO: $1.33
Button: $4.79
SB: $3.90
BB: $2.87

Dealing Hands: ($0.15) A♥ A♠ K♦ 5♦ 2♠ (6 players)
UTG calls $0.10, Hero raises to $0.20, 4 folds, UTG calls $0.10

First Draw: ($0.55) (2 players)
UTG discards 1, Hero discards 3,
A♥ A♠ || K♠ 7♣ 9♥
UTG bets $0.20, Hero ???

It really doesn't matter much for this purpose whether villain draws 1 or three.

If D1, then could be a busted draw.
if D3 or even two could be a worse pair.

In 117K hands, I have faced this decision 2100 times,

Here are the results:

Actn      BB  Count
   X
    444.82 1072
   B    147.87  394
   C   -176.14  202
   F   -198.50  198
  XF   -116.50  116
  XC    -97.88  110
  BF    -70.00   35
  BC    -71.62   30
  na     13.38   19
   R     -3.50    3
  RF     -3.50    1

 TTL   -131.57 2180

This table shows those times I raised pre with a pair.
It does not consider postion

I can't cut out calling altogether of course.

How do we decide when to call?

What criteria should be used?

I presume at least some of the time I folded the best hand, but considering my calling station tendencies, I doubt that.

I'm also guessing that at least SOME of the time I checked, a worse hand would have called.

Cheers,
--VS


Reply
Quote
5 replies
jbpatzer
Joined: 23.11.2009

Oh well it's game theory innit.

Assuming he has a flush draw, he'll get there about 18% of the time, so he needs to bluff 18% / (1+0.55/0.2) = 5% of the time. If he bluffs more often than that you should always call, and less often you should always fold. Given that people who know anything about this sort of thing tend not to limp/call their flush draws, you can be sure he's not bluffing optimally, so you just need to know which side of the fence he sits on. If you want to bluff catch unexploitably, you need to call about (0.55/0.2-1)/(0.55/0.2+1) = 50% (roughly) of the time that you don't improve, so you should call with the top half of your range, which is all bluff catchers in this spot. If you're raising, say JJ+, that's probably a call with AA+.

This is all a bit rough and ready, and I haven't done the sums properly, but you get the idea.


Reply
Quote

Assuming he has a flush draw...

:rolleyes:

It would be much easier if they will only limp/call with draws, but they can have anything from quads to 3322 also.

Since It's FL 0.2$ (big players rotation), you will rarerly have enough hands to notice how far they are from GTO.

Basic math tell us that we need to put 0.2$ to win ~0.91$ pot (0.96*0.55+2*0.2),
so if we win more than ~22% (0.2/0.91*100%) of the time calling is EV+

EV(call)=0.22*(+0.71$)+0.78*(-0.2$)=0

Sometimes pot is slightly lower (e.g. Hero open raise, BB call)
Pot=0.45$
0.2/(0.96*(0.45+2*0.2))*100%=~24.5%, so we need to win slightly more offen.

The only reasonable approach to this is to take all of your hands where:

:spade: pot was single raised by you;
:spade: you were against one opponent;
:spade: op draw 1, while you draw 3;
:spade: you call or check/call his bet;

and check out if standard fish bluff more offen than 22-24% of the time.
If it's doable please post the results here, I really would like to see them.


Reply
Quote
Super Moderator
VorpalF2F
Joined: 02.09.2010

Originally posted by Yadumalchtoona18let

Assuming he has a flush draw...

:rolleyes:

It would be much easier if they will only limp/call with draws, but they can have anything from quads to 3322 also.

The only reasonable approach to this is to take all of your hands where:

:spade: pot was single raised by you;
:spade: you were against one opponent;
:spade: op draw 1, while you draw 3;
:spade: you call or check/call his bet;

and check out if standard fish bluff more offen than 22-24% of the time.
If it's doable please post the results here, I really would like to see them.

I don't know about standard players, but there are a few I've marked who like to bluff FDs and even SDs

My 2180 hands don't take into account how many opponents saw draw, nor how many each drew. They are simply all instances where I raised 1 pr and did not improve.

It is rather clear that I'm calling too often, or in the wrong spots.

Or both...

Tks,
--VS


Reply
Quote

I would guess that a typical UTG limper at microstakes plays passively. This player may have a hand like two pairs or a flush/straight draw. As jbpatzer mentioned, if you want to play unexploitably, you must call roughly half of the time you don't improve and it's considered quite standard to take the top of that range; after all, there should be a tiny chance your opponent started with an A-high flush draw that paired on aces and decided to value bet them, so you may as well call with your best one pair hands. Even if your range predraw is only AA, your AAK falls at exactly the top 1/4 of your unimproved range, so IMHO, it's an autopayoff given that you don't know this player. If you do not call with this hand, you'd be folding unimproved quite often; then observing regular players will notice over time that you can be quite exploitable in this spot and try to determine other spots where you can be easily exploitable. The plus side to calling is that you have more data for this opponent ( at the very least, you can type in a note ) and observing players will notice that you are willing to call down, so it's less likely players will try to exploit you in the future which simply means more long-term EV against the observant players.


Reply
Quote
YohanN7
Joined: 15.06.2009

You can also STOP bluffs and INDUCE bluffs. For instance, IP against a player who bluffs too little, stand pat with a reasonably strong hand (say AAA) when he draws one. Now he is not going to bluff for sure. The value of these things is questionable (because either you can't improve or you reveal your hand when drawing 3 to induce), but I can imagine it is worth in in NL versus unbalanced players.


Reply
Quote