I've been meaning to try and put together a post/thread on this subject for the FAQ for ages. Please feel free to comment, elaborate, etc.



Posts about folding KK preflop or a set are often met with derision and contempt. This is because the standard answer, the default play, is to not fold.

The only way this changes is with information that is often absent from these types of posts - A read on villain. Without that read, the question becomes unanswerable and we are just left with the default answer.

Do not fold.

And keep in mind that folding in these situations is often so difficult that a read on villain may not even be enough. We may well require some idea of how villain sees us. For example, does villain see us as a nit who would only 3bet KK? If so, then his 4bet range is that much more narrow. OTOH, does villain see us as a LAG so he may be taking a stand with a much wider range?

If you make such a post without this kind of information, do not expect to be met with kindness. And if you do include this information, then answer is generally obvious. So all you're really doing is posting for confirmation and reassurance that you made the right play, not advice. In which case you're just as likely to be called out for your self-indulgence. That, or you're posting a bad beat or cooler. Which no one appreciates. We've all been there. Many, many times.

So avoid making these kinds of posts unless you're really confused, you're prepared for a possibly unfavorable reaction and you're able to provide more than just a hand history.



In order to make folding KK preflop correct, you must put villain on exactly AA to a VERY high degree of certainly. Depending upon the amount of money already in the pot and remaining stack sizes this read may need to be a virtual certainty. Remember, even if he does have exactly AA you still have close to 20% equity in the hand. So if the pot is already big and he just might have AK or QQ sometimes it gets a lot harder to say fold.

equity win tie pots won pots tied
Hand 0: 57.191% 54.62% 02.57% 19640560 924261.00 { KcKd }
Hand 1: 42.809% 40.24% 02.57% 14469302 924261.00 { QQ+, AKs, AKo }

So it comes down to, does he only have AA? The answer is usually, "No he has JJ/QQ/AK/78S sometimes" but use your judgement. You're the one sitting at the table with him, you're the one who knows the table texture, you're the one who knows how aggressively you've been playing. It's unlikely anyone else is going to have a better grasp of the situation. Make you read and go with it.



There are three recurring situations which justify considering folding a set. Once again, they depend on reads.

First and most common is against a likely straight or flush. This is a judgement call based on how likely villain is to have it (ie: board texture, if he chases, if he's passive and never bets without a monster, etc), versus pot odds and our remaining equity in the hand (our chance to boat up if we're not on the river yet) versus implied odds (again if we're not on the river and if villain is likely to stack off if we fill up) versus reverse implied odds if we want to try for pot control and get to showdown cheaply if we don't improve (ie: whether we have to call one bet, two or three to see showdown) or if we decide to just felt it (ie: on a coordinated flop).


Second is if we're set mining, hit a small set and another likely set miner springs to life. For example, UTG raises PF, we call MP with 33, another set miner calls behind us. Flop is T53 rainbow. UTG bets, we raise and the other set miner behind us raises again. Can we fold here?

Once more, the most important piece of information here is our read. Is this raise likely to be anything other than a bigger set? If the raiser is a nitty set miner than this is very likely a set given the flop texture - There's not much in the way of draws or 2pair hands available. On the other hand, if the flop was T93 with two spades and villain is LAG, there are a bunch of draws out there as well as T9. Or he may just love his AT. So we're much more likely to have the best hand.


The third situation is also when we flop a small set but in a limped pot where a TAG or nitty player makes it clear they want to play for stacks. Since these types of players seldom want to play big pots in limped hands with anything less than 2pair this is where we may want to be cautious and slow things down. Again though, this comes down to a mix of reads and flop texture. How many hands are we beating based on flop texture and how good is our read on villain? Could he play an overpair like this?



Deciding to make these big folds should not be common and they should require more than just an uncommented hand history. There are no absolutes in poker and there will be times when folding or calling are not clear. However, keep in mind that most of these situations are actually common, trivial and people get tired of hearing about them. So if you do post such a hand, you may not get the kind of response you're hoping for. Don't be surprised by a negative reaction.

good post, thank you. Can we hand out bans for these now?

Great post. Thanks

it'd be sweet if for a week a thread about folding would lead to a week-long ban

of course no hands would get posted, but it would be entertaining

NH, CMAR!

One thing that pops out to me about folding KK (not something I have studied in detail TBH) is that say your not sure which range the villan 4-bets (Assuming 100BB)
(A) KK+
(B) QQ+ AK

Against the worst case range (KK+) your crushed 80-20
Against the range your hoping for your not a huge favorite anyway - 57-43. So folding against range "B" isnt as bad a mistake as shoving into range "A".

Am I making any sense?
Edit - I forgot dead money does provide an overlay but does my thinking stand correctly if we adjusted for that?

Thanks CMAR,

Like the new Avatar. Can you just make this a sticky on its own once people are done commenting? So few actually read the FAQ before they post. Maybe Fold KK preflop in the title will get their attention.

Although I am not as good as most people here when it comes to theory, but to me you make sense.

Usually I would advocate shoving KK for 100BB 100% of the time without any very solid information that I am up against AA. I even could not imagine what this information might be up to where I played the game, which is 10NL.

Still, your point has something worth thinking about I guess. You lose some value, but also avoid some very expensive pots you will lose. So you reduce variance in these spots.

fold kings for 100 is usually (unless he is a real nit or smth like that) a no go (did it once - what a surprise - he had aces)

Well - but fold kings for let's say 200+ bb's is another game - usually the shove range for 200+bb's become to AA only - so a call with KK here is just throwing the money out of the window (unless he's a maniac)

Thanks CMAR!

Any one else find it slightly comical that CMAR talks about never posting on folding KK, and now here we all are discussing its theory, which we're not supposed to talk about...

+++
lock&stickyftw

I think that is exactly what we are supposed to be talking about, the very narrow range of times you can fold KK. Basically making a reference for anyone who wants to ask about folding KK so we don't have to have keep having the conversation that pretty much leads back to the same answer over and over again, which is no.

A+

I'ma just post the link to this thread in the "fold KK/set" threads now. Thanks.

Yeah, that's why I left this thread open rather than just locking and stickying it. So if there was more to say, objections to cover, etc we could get it all out in one comprehensive thread.

Yes, you make sense, except folding against QQ+/AK range is actually a pretty huge mistake given all the dead money that is likely in the pot.

Let's take a simple example. 100bb stacks. Villain raises to 4bb. We 3bet to 12bb. Villain pushes (note the more complex situation where villain 4bets and it's up to us whether to push or fold. We have to account for some fold equity to go along with our pot equity).

Including the blinds, there is 113.5bb in the pot. We need to call 88bb. If we call, we win 57.191% of the time and we loose 42.809% of the time.

So the expected value of calling is: 113.5x.57191-88x.42809

The ev of calling is: +27.239865 bb


Against KK+, our equity is 22.618%

So the expected value of calling is: 113.5x.22618-88x.77382

The ev of calling is: -42.42473 bb


Let's compare that to KK vs AA (We're better off if the AA has none of our suits, but worse off if it has both so I've chosen the middle road). In this case, our equity is 18.054%.

So the expected value of calling is: 113.5x.18054-88x.81946

The ev of calling is: -51.62119 bb



So yes, the amount you loose when you're up against KK+ or AA is more than what you gain when you're up against QQ+/AK. However, when you're up against KK+ or AA you're not loosing as much as you probably expect. And when you're up against QQ+/AK you're actually winning a considerable amount. And how much more likely are you to be up against QQ+/AK than KK+ or AA?

Also keep in mind meta considerations and Shania. There are many situations (ie: steals and squeezes or if you've been battling each other for hours) where villains range is going to be much wider than QQ+/AK. And if you fold too much preflop you become exploitable (not something you need to worry about much in the micros but you will eventually).

And this is the simplest of examples, there are no other callers of the initial raise. Any more dead money makes it that much more profitable to call when we're ahead and that much less of a loss when we're behind.

So against a 4/2/.5 nit, sure, go ahead and fold to his push. But against more aggressive opponents folding that KK can be quite costly unless you're really sure you're right.

Thanks CMAR, I think thats the maths of this one cracked.

Lets take your example and throw in something "sklankyesque" to make sure we on the same song sheet here.

100bb stacks. Villain raises to 4bb. We 3bet to 12bb
We have a decision to make depending on our read of the villan.
Lets say we know that the villan is a bot and is programmed to push with either range A 50% and range B 50%.
Range A = KK+
Range B = QQ+,AK

If we knew that was his "range distribution" then we should fold because our EV is
+27.24 bb - 42.42 bb = -15.18bb

Again, shout out if Im wrong here but that means that we can still be cautious even if his range is QQ+,AK half the time??

I folded Ks PF once at NL10 about three years ago. Dude showed me aces. No reads and I sucked worse than I do now. I can rightfully say that I have folded Ks PF correctly 100% of the time. Having said that I will never do it again.

yo, can we also make it mandatory that you post your stats in the OP? I played a small session of like 300 hands at nl50 last week and people just spew all over the place, but then people post a hand asking if they can fold top 2 in a spot where I've seen plenty of people show down 2nd pair and it doesn't make sense to me. I just don't get why people are only interested in folding in this forum.

I think we need to clarify what "range" means.

I think you're trying to say that when he pushes, 50% of the time he has KK+ and the other 50% of the time he has QQ+/AK.

That doesn't really make any sense because when he has a range of QQ+/AK that includes KK+. So we're actually counting the KK+ portion of his range twice. You're trying to take a range of ranges, that doesn't really work very well when those ranges overlap.

What you're really trying to do is assign villain a dynamic range. We want to model the assumption that he doesn't always play his hands the same way. Most of the time when he has AA he pushes, sometimes he just calls and traps. OTOH, when he has QQ he's more likely to call and see if he likes the flop, he only pushes QQ sometimes.

Normal ranges assume an even distribution and consistent play. As modelled up thread, he always plays AA the same and he always plays QQ the same.

We want to say he's more likely to push AA than he is to push QQ.

So I'm going to pull some numbers out of my ass to represent just how often he pushes a particular hand.

Lets say he pushes the following hands the following percentage of times when he holds them:

AA 80%
KK 70%
AKs 60%
AK 50%
QQ 50%

When he's not pushing he's just calling our 3bet or folding (obv he's never folding AA but he may fold QQ or AK/AKs sometimes).

So then we need the distribution of hands to figure out how often he will hold each of them. Since we have KK we know there are the following ways he can have these hands:

AA 6 ways x 80% = 4.8
KK 1 way x 70% = .7
AKs 2 ways x 60% = 1.2
AK 6 ways x 50% = 3
QQ 6 ways x 50% = 3

So when he pushes, he has

AA 37.7% of the time (4.8/12.7)
KK 5.5% of the time (.7/12.7)
AKs 9.45% of the time (1.2/12.7)
AK 23.62% of the time (3/12.7)
QQ 23.62% of the time (3/12.7)

Our equity against these hands are:

AA 18.054%
KK 50%
AKs 65.894%
AK 69.881%
QQ 81.935%

To find our equity:

(We're figuring a weighted average of equity given the likelyhood of him holding any particular hand)

(37.7%)(18.054%)+(5.5%)(50%)+(9.45%)(65.894%)+(23. 62%)(69.881%)+(23.62%)(81.935%)

51.64%

So, given the dynamic range outlined above, where he pushes AA 80% of the time, KK 70% of the time, etc, we have 51.64% equity against that range.

With 113.5bb in the pot and needing to call 88 bb, the break even point is 43.67% equity.

Now you can use this relatively simple formula and adjust his range - If you think he ALWAYS pushes AA and KK but is less likely to push QQ then you can easily recalculate our equity. However, also keep in mind that we've assume a pretty limited range. We aren't allowing him to EVER have JJ or 89s here. Adding those kinds of hands, even as a small possibility, will improve our equity dramatically because we're so far ahead.

And keep in mind how dead money affects EV. The more dead money in the pot, the less equity we need to be profitable.



The bottom line is that folding KK preflop is difficult because while we are crushed if villain has exactly AA, if his range is anything more than just AA we are crushing the rest of his range. So unless we can put him on AA with a high degree of certainty felting is going to be a +EV proposition.



(Somebody please tell me all that made sense! )

Posting your own stats is a really good idea and we rarely ever see it. The image that YOU have at the table makes a huge difference in how your opponents will be playing/reacting against you.



Makes sense to me, CMAR. Looks goot I think.

+++
OP's image relative to the villain is critical info.
If there is no history it should at least be stated.

Great thread, and good suggestions about stats/image inclusion (not that either will be payed attn to.. I'm such a cynic). I'll add that while at the micros, valid spots to fold KKPF/Sets are so rare that even if you did find them, it would have a completely negligible effect on your winrate in the long-term. Use the brain power and study time to focus on the likely many other holes in your game.

not that a discussion around these spots is worthless as it often brings up incorrect mathematical methods and wrong ways of thinking that affect all poker decisions.

ok this thread is making me really confused. i understand that folding kings is difficult/not smart at 100bb almost ever etc, but about all the maths and such you are doing... what about the fact that many times you know your opponent will 3-bet/4-bet/push AA and KK, and more often just flatcall after a 3 bet or 4bet with qq and ak? i play micros and played a lot of 250bb+ at stars, and my impression is that people almost never push over your 4bet with qq and ak. lets assume i got a huge amount of stats and found a random villain at the stakes i am playing will do the following:

99% of the time push with aa
97% of the time push with kk
30% of the time push with ak
50% of the time push with qq

the rest call.

if this is the case it should be incorrect to asume a range of hands he will play the same way and do maths based on this? i dont know how to explain this right but i KNOW people shoves AA much more often than qq in this scenarios. even when i think qq is a possible holding after a shove, AA will always be a more -likely- holding. a much more likely holding no matter what range i feel is POSSIBLE for him to be doing the shove with. so range = x+y+z, math = +ev, = gg put money in with kk.... it doesnt add up for me.

im also thinking about stack sizes. i dont know for higher limits, but in nl5 at stars you can buy in for 250 bb and i have met aa with kk and kk with aa several times with 250-400 bbs. when this happends it gets very clear that this is AA vs KK. and people very often tend to call the 3. raise with kk, probably because they are worried about aa. (yeah i know thats a sucker-strategy). anyway i have yet to see somebody go broke for that many bb preflop with qq and ak. when i get all the chips in with my two aces at 250 bb+ it is almost never anything except KK villain has. i am thinking "how can i take 3,7 seconds of time before i make the next raise now to make him believe i have pocket 77 and go broke with his jacks or queens"... just to find out he had kings. because people dont go broke with qq and ak preflop for that many bb. so how can i ever trust such a huge range as qq ak+?

as i see it, when playing deepstack a good player should be able to put people on AA with a really HIGH degree of certainty many times. and not go broke.

i would like to see some good players play for 100K bb+ and watch how the strategy with kk vs aa and set vs set scenarios develope. could be really interesting. would be a totally different game. KK will for sure never ever go broke preflop! so it is a question about stacks. the real question about kings vs aces should be, how many bb deep should i be before considering folding kings?

uhm this was a lot. sorry if this does not make sense to you guys, with your own community with your own language and way to think about poker, but i had to get my thoughts out... now enlighten me!

and btw i have never thought seriously about folding kings yet.

Some of these numbers are just too optimistic. I especially hate the part where it is said that villain will push AKo 50% of the time to our 3-bet for 100BB. Most players don´t even 3-bet AKs, and a fair share or players never push AA. If you make a FAQ/Sticky please prove your numbers to be true.

You don´t have 55% equity when villain pushes 100BB over your 14BB 3-bet.

With due respect, this is a mistake. PAHUD allows us to get away from KK at least in the most obvious situations. Last Tuesday or so I was playing against an amazing nit. His stats were 4/0.7/1.2 through ~4300 hands. He raised UTG, I had KK OTB. Is there really any doubt what he had? Now, if I had played 4300 hands against him without PAHUD, maybe I would have known he was nit that almost never raised, but I couldn't put him on AA precisely and I would have lost a stack. In this case, however, because I had such a precise read and a large sample, I mucked without even throwing in the obligatory 3-bet and I haven't thought twice about it since.

Today I was playing against an uber-nit whose stats were 12/1/2 through 1846 hands. He raised UTG, I put him on AA exactly, but I smooth called in the CO and then called him down to confirm my read. He had QQ, and I made a note so I would play differently next time.

Not throwing away KK sometimes preflop is just a bad habit for online players. I agree with the main point of the OP, which is that doing it is highly read dependent. But PT and PAHUD allow us to make those reads in some situations.