5 posts / 0 new
Last post
coffeeyay's picture
Limping in the SB (when BU folds)

I wanted to open up some discussion on something I didn't cover in my Beating Spin and Go Poker videopack: limping in the SB when the BU folds. I believe that it's quite close in terms of readless EV to min-raising with most hands so it's something that you can often get away with not using. However, in some places it may be a very good option for us. What kind of hands and in what situations do you guys believe limping may be be better than min-raising?

To give us a starting place I did a little bit of math to help us compare the two options. First let's look at limp/folding. Let's look at the EV of limp/folding assuming that open folding is 0bb in EV:
EV(Limp/fold) = -.5bb*R + EV(when villain checks)*(1-R)
Where R is villain's raise frequency versus our limp. We can expand the new EV term by thinking about a number F which represents the fraction of the pot we expect to win on average post-flop. This lets us write:
EV(when villain checks) = 2bb*F - .5bb
This is simply the fraction of the limped pot we expect to win minus our investment in that pot from our decision point.

Now with some algebra we can solve for the F we need to make limping better than folding. I'll spare you the gory details and just post the result, which due to some cancellation ends up simple.

F = .25/(1-R)

A common raise frequency I've seen is 30%. This would lead to F = .357. To give you a reference point, when you flat a min-raise OOP in a HUSNG you need F = .25. However, in this case ranges are quite different since a min-raising range is uncapped but also doesn't have some bottom hands, while a check back range will be capped and include more junk.

To look at min-raise/folding we need to know how often villain raises and folds. We can write the EV the same way as for limping in terms of a new F term. R is the 3b frequency and C is the calling frequency.
EV(minr) = 1bb*(1-R-C) – 1.5bb*R + (4bb*F – 1.5bb)*C
Solving for F gives a slightly more messy:
F = .375*(1/C)*R - .25*(1/C)*(1-R-C) + .375

Let's again plug some numbers into this. A common raise frequency I've seen is 20%, and a common call frequency I've seen is 45%. Then we get F = .347, so a bit lower than for limping. It's important to remember though that hand villain's range will be stronger than when he checks back, so it may be harder to win a higher fraction of the pot. We can now see why limping and min-raising are likely pretty similar in EV in many situations. We can now start thinking about hands that maybe benefit from the larger pot post-flop caused by min-raising, or that have similar equity versus both the flatting and check back range and so appreciate not needing quite as much F to make the min-raise profitable. You can also examine other frequencies and see how F will vary in each case given different player types.

Try to work out some combination of min-raise and limp ranges for different player types. What kind of results are you getting? Post them here and we can all discuss them and build the best SB range we can.

it1111's picture
Hi Coffeeyay,I just checked

 

Hi Coffeeyay,

I just checked my database for pop tendencies and my numbers (ISO, flat and 3B) are very close to yours at 30s. At lower stakes (7s and 15s), ISO and 3B are bit lower but only about 1% to 2% difference so this does not change the final results of your formula much.

I would also like to point out to another thing that you didn't include into your calculations, and this is fold to cbet percentage on flop or is this already included in F?
Fold to cbet  on flop  in LP is higher than fold to cbet in MR pot.  
I checked pop tendencies for this stat too and it varies slightly at different stakes (not sure about this, it could be sample size issue) but at all stakes that I played (7s, 15s, 30s) FvCB(LP) > FvCB(MP).
 

I think that this changes the final result slightly in favor of limping for all marginal hands that benefit from fold equity postflop.
Good hands still prefer raising i think. Have to think bit more where to draw a line between good and marginal hands though.

I'm sure we can benefit for having 2 different ranges in SB (limping and min raising), but we need more calculations or empirical data to see which hands are better limped, minraised or folded...

What about raising to 2.5bb in SB? Our good hands would benefit from bigger pot and weak hands would benefit from additional fold equity?

---------
UPDATE:

I was thinking about this overnight and found another reason why limping marginal hands could be a good option. 

When I checked my database for chipEV winnings in 3-Handed play and HU play it looks like HU is much more profitable for me than 3-Handed, so in order to get to HU with a weak player and maximize my edge I want to lower the variance in 3-Handed game and limping marginal hands in SB will have lower variance than min-raising them, so if equity between 2 actions is very close why not choose lower variance one and get a chance to make more EV later if we get a chance to get to HU play?

 

 

 

 

 

 

 

 

"If you want to win, you must not lose!"

rahold's picture
Hi coffeeyay, An interesting

Hi coffeeyay,

An interesting topic. With the fold% I found in my database to CB and LCB both are probably profitable. 

If open folding is 0EV is your reference point shouldn't the formula for minraise/folding be:

EV(minr) = 1.5bb*(1-R-C) – 1.5bb*R + (4bb*F – 1.5bb)*C

If our steal is successful (1-R-C) we win the pot of 1.5bb.

In that case solving for F gives:

F=0.375*(1/C)*R-0.375*(1/C)*(1-R-C)+0.375

With the numbers you used we get F = 0.25 which makes raising a clear favorite over limping.

My database however gave a higher flat% (55%) which gives F = 0.35 which is again close to the required F for limping.

4 card brett's picture
should you not also account

should you not also account for the villains 3 bet % when hero raises from SB and compare that to villains ISO when hero limps from SB

if the ISO is greater than 3 bet then this can have an effect on profitability can it not?

rahold's picture
That is taken into

That is taken into account.

EV(Limp/fold) = -.5bb*R + EV(when villain checks)*(1-R)

The part -.5bb*R is handling the ISO.

EV(minr) = 1.5bb*(1-R-C) – 1.5bb*R + (4bb*F – 1.5bb)*C

The part -1.5bb*R is handling the 3bet.