$80 NLHE MTT Rebuy: SB shoving against BB

[rumsey182]

lots of points so ill address them individually.

1/ you failed to address why we should assume that the calling range on a representative player in this player pool will call sufficiently wide enough to make nash shoving better than exploitably wide shoving.

2/ if he calls optimally he will be +EV, both players will be. owing to the fact there is money in the pot. his worst calling hand will be indifferent between calling and folding meaning the equity of the two will be equal, given folding is 0 EV. therefore his expectation must be positive or he would be indifferent between folding all hands.

Your kind of right,.. but not really. The blinds count in EV, but that is already part of the equation. What happens is range vs range in a frequency sense it is at worst neutral (neither can increase their EV above 0 if both know each others strategy). Your thinking in a hand vs hand situation, and yes you will have hands be +ev but in a true unexploitable situation when you are shoving correctly and they are calling correctly it is neutral. This is the value of it, because if you could tell someone what you are doing and it can't help them to be profitable, it makes your position powerful. This kind of ties back into what i said as it sort of being like a freeroll,.. it isn't exactly a freeroll but it is the closest example I could think of

3/ I am aware of what a nash equilibrium is, I am a masters student studying economics. your assumption about nash being optimal is fundamentally incorrect. provided there are identifiable player pool tendencies it would be optimal to construct a representative agent and make exploitable plays against that model. or better yet we could weight that model appropriately with different calling ranges and create an aggregate.

you can not make any assumptions in a vacuum, which is what we are in. I'm guilty in another thread (KQo wanting to call a 3 bet OOP) of actually making assumptions without having the ability of doing so. No info means we have to assume nothing ( which is when nash is optimal)

4/ let me hammer this point home, this is not a a bluff/nuts v bluff catcher situation. the calling ev is definitely not 0. if it were zero, they could increase their profit by calling with less hands. and the consequence is to imply that they are indifferent between folding 100% and calling nash, which is patently absurd when their range is uncapped and they could choose to call with only KK+ and show (an admittedly infrequent) yet clear profit.

yes but the folding is negative because of the overlay in the pot

5/ just because you shoved a nash range does not mean it was optimal, assuming the representative player calls too tightly, which seems a very very very conservative assumption to make, that means that we are folding hands that have a positive expectation when shoved.

i agree, but we can't assume this, we must establish what an unexploitable range is and make adjustments when we gain info

Poker is a game of limited information, you may honestly never be able to truly "know" anything simply because you need repeated times of being in the same situation to test any hypothesis we do develop about someones tendencies. We are not guaranteed of this more so in MTTs. Least in cash games we have a lot better chance of someone staying for a few hundred hands and or we have a chance of playing against the same people with the same stack sizes many times.

When we can't know anything, we have to basically attempt to assume more or less that the villain will basically know what we are doing as a default. We can make some mild assumptions, but you simply can not make assumptions about someone in a vacuum.

Making assumptions about the "population" of what you are playing is somewhat unreliable as well, we simply do not have a good chance to play the same representative sample of the true population enough times to make our confidence in these assumptions.

Your right that generally people do not call enough in allin situations like this, but from an academic stance you can not assume such. In any situation in poker you can put yourself into a guessing game, but i would not recommend this unless you have a LOT of experience. Most people need to be focusing on more important things anyways, but teaching against a worst case scenario is ideal with lack of info.

Sorry TLDR but I'm sort of agreeing with you but you can't make any assumptions in a vacuum ( and it is good to teach the mechanics of how you avoid assumptions because it leaves room for people to fill in with there own info)

 

[duggs]

This is wrong, I'll come back later and respond

 

[duggs]

Your kind of right,.. but not really. The blinds count in EV, but that is already part of the equation. What happens is range vs range in a frequency sense it is at worst neutral (neither can increase their EV above 0 if both know each others strategy). Your thinking in a hand vs hand situation, and yes you will have hands be +ev but in a true unexploitable situation when you are shoving correctly and they are calling correctly it is neutral. This is the value of it, because if you could tell someone what you are doing and it can't help them to be profitable, it makes your position powerful. This kind of ties back into what i said as it sort of being like a freeroll,.. it isn't exactly a freeroll but it is the closest example I could think of

This is fundamentally incorrect, like really really really wrong. the definition of a nash equilibrium is that neither player can increase his payoff by deviating, no where does it state nor impose that this must be equal to 0. the maximum any player can win in a nash situation is the size of the pot, what is because if they won more than that their opponent could increase their payoff by folding 100% and that violates the definition,

both ranges will profit, if you shove nash and he calls nash you will each get a share of the pot which neither can improve upon by deviating, both will clearly show a profit.

further more it is logically inconsistent that a range comprised of only hands that profit or break even can equal zero, the other element of a nash equilibrium that doesn't usually get mentioned is that each hand has to be played for maximal profit given opponents strategy, that means no hand can have an EV of less than 0.

you are conflating EV with exploitation, being unexploitable doesn't imply anything about the EV of a situation.

you can not make any assumptions in a vacuum, which is what we are in. I'm guilty in another thread (KQo wanting to call a 3 bet OOP) of actually making assumptions without having the ability of doing so. No info means we have to assume nothing ( which is when nash is optimal)

this would only be a true vacuum if we had no information about buy in, stage of tournament, location (online site or physical location) timing. we can and absolutely should apply population tendencies, your assertion that we should play nash implies that we need to assume a symmetric distribution of players around a nash calling range.

yes but the folding is negative because of the overlay in the pot

folding can never yield -EV, and never because of the size of the pot. (some count losing a blind and ante as -1.2BB EV but then its much more logical to add this amount to plays and use folding as zero, it is also computationally simpler)

i agree, but we can't assume this, we must establish what an unexploitable range is and make adjustments when we gain info

why? for our exploitable range to be inferior we have to be worried about being exploited, that means a random in this tournament has to call exploitably wide, and has to do it more often than they call exploitatively tight.

Poker is a game of limited information, you may honestly never be able to truly "know" anything simply because you need repeated times of being in the same situation to test any hypothesis we do develop about someones tendencies. We are not guaranteed of this more so in MTTs. Least in cash games we have a lot better chance of someone staying for a few hundred hands and or we have a chance of playing against the same people with the same stack sizes many times.

this is why we use bayesian probabilities to weight our opponents tendencies, I'm sure you do it in practice but you have ignored my earlier question about whether you use balanced b/c and c/r c/c ranges on all streets to avoid exploitation or you make assumptions about general opponent tendencies based on other things you have seen them do.

because the extrapolation of your argument is that if we are playing a 100/0 who is infinitely passive on flop, we should have a balance b/c range on the river because we can't make assumptions

When we can't know anything, we have to basically attempt to assume more or less that the villain will basically know what we are doing as a default. We can make some mild assumptions, but you simply can not make assumptions about someone in a vacuum.

see above point

Making assumptions about the "population" of what you are playing is somewhat unreliable as well, we simply do not have a good chance to play the same representative sample of the true population enough times to make our confidence in these assumptions.

it doesn't have to be very reliable, it just has to be a strictly better assumption than no assumption, you know, those small edges you were talking about?

Your right that generally people do not call enough in allin situations like this, but from an academic stance you can not assume such. In any situation in poker you can put yourself into a guessing game, but i would not recommend this unless you have a LOT of experience. Most people need to be focusing on more important things anyways, but teaching against a worst case scenario is ideal with lack of info.

Sorry TLDR but I'm sort of agreeing with you but you can't make any assumptions in a vacuum ( and it is good to teach the mechanics of how you avoid assumptions because it leaves room for people to fill in with there own info)

some of your assertions on nash equilibrium and nash push folds are fundamentally incorrect, i don't see a good argument against using representative agents, and you are hedging all of your earlier positions.

 

[duggs]

Unexploitable range can still be way way from optimal

and just to further point out that you liked attacking this quote,

Unexploitable is only optimal in the face of a nash calling range.

 

[LeGenie]

Let's analyze the math.

Pot contains 300*9= 2,700+ 1500 +3000=7,200

Which picking up the pot preflop increases our stack by 18%

How much FE do we have? With unknown players I just assign them my own range and I'd be calling here fairly wide because I only have 10bbs. Something like 44+, A2+, KJ+, K9s+, QTs+. That is a 23% calling range.

So 77% of the time he'll just fold and you'll win the pot which is 7,200*.77=5,544 is your fold equity

The other 23% of the time he'll call which costs you 28,500 more than just folding.

Facing his calling range 97o has 35% equity meaning:

Of the 23% he calls 65% of the time you'll lose 28,500
28500* 0.23 * 0.65= 4,260 lost

Of the 23% he calls 35% of the time you'll win 28,500+7,200=35,700
35,700 *0.23*0.35=2,873 won

now add your fold equity plus your win equity and subtract your loss
5,544+2,873=8417 - 4260= +4157 is the profitability of the move as a whole which is better than 0 (0 is always the equity of folding)

An average increase to your stack of roughly 10%.

Is that an adequate reward to overcome the ICM reality? Depends on where you are in the tourney, but clearly if he's only calling 23% of the time pushing here is never WRONG.

Notice most if your profit in that move comes from fold equity; so increasing his calling range drastically changes the profitability of the move.

Obviously the tighter the BB the wider you should push and the looser the BB you should be pushing for value (hands you suspect are ahead). For instance if you think he is calling wide with hands like K3 or 68 then increase you high card pushing hands and decrease your low card pushing hands.
Ex: Push with K6 and Q7 instead of 98s or 67s

Hope this helps. Plug in different calling ranges to see how it affects the profitability of your shove. Plug n Play.

Thank you for the feedback! It's time to plug and play

 

[Everybodylovesdeuces]

I will add that if you had been at this table for a while you would also want to factor some meta game into your decision. If you have been pushing quite a bit lately you are going to be more likely to get called which isn't great for 97. If you have been a little tighter then go ahead and push away. At least you're often live when you get called. It's tough when ave M is 5. You're going to have to push a lot of hands and get lucky.

 

[rumsey182]

nash doesn't mean your never shoving negative EV hands actually ( you can have a net positive shoving range that has negative -ev hands in it)


You know the more i have been thinking about this,.. i have to say i was wrong both can be +ev sorry which was the crux of it

hope someone learned something useful with me running around in circles some

 

[Aces2w1n]

Yes I loved this thread....

Keep arguing!!!!!!

 

[duggs]

nash doesn't mean your never shoving negative EV hands actually ( you can have a net positive shoving range that has negative -ev hands in it)


You know the more i have been thinking about this,.. i have to say i was wrong both can be +ev sorry which was the crux of it

hope someone learned something useful with me running around in circles some

Nash does mean you are never shoving negative EV hands. You could unilaterally increase your EV overall by not shoving those hands and it violates the definition. You can find texts in both janda and tiptons books that verify that claim. It also is somewhat similar to the sub game perfection requirements outside of poker game theory. You can find similar requirements in any signalling, screening assymetric game, or in any perfect Bayesian Nash equilibrium result.

 

[rumsey182]

Nash does mean you are never shoving negative EV hands. You could unilaterally increase your EV overall by not shoving those hands and it violates the definition. You can find texts in both janda and tiptons books that verify that claim. It also is somewhat similar to the sub game perfection requirements outside of poker game theory. You can find similar requirements in any signalling, screening assymetric game, or in any perfect Bayesian Nash equilibrium result.

your right, im confusing the fact that a +ev shove range can include -ev hands, but it would not be nash

 

[duggs]

a +EV shove could have its EV increased further by removing said hands tho?

 

[duggs]

It very rarely makes sense to knowingly make a -EV play with any hand

 

[Matt Vaughan]

It very rarely makes sense to knowingly make a -EV play with any hand

Never, in fact, unless in so doing, you increase your future EV.

... In which case, was it really -EV??

 

[duggs]

Never, in fact, unless in so doing, you increase your future EV.



... In which case, was it really -EV??

Or unless your opponents strategy is contingent on your strategy, buts that's a weird condition