Hand 37 (page 163):

Hero is UTG and raise 3BB with JsJc.
BB (Reg, VPIP: 26 / PFR: 20 / 3Bet: 9 / Fold to Cbet: 40) calls.

Flop: Qd 7h 6d

BB checks, Hero checks

I'm not convinced Hero should check instead of Cbet.

Well, what about the straight and flush draws? Why they weren’t mentioned here?

I have 2 conflicting arguments.

1- If we Cbet, Villain will only fold weaker hands that probably wouldn’t improve on the Turn and he will call with his draws anyway. So we check.
2- We are probably ahead, so we want to extract some value if Villain try to chase his draw. So we Cbet.

1 or 2?

Not that we want to derail too much, as this thread is more about the book overall, and less about a deep discussion of single hand examples, but I would check back here almost always, for the reasons given in the book; this hand would be one of the first I would have in mind to put in a check-back range in this spot. Possibly bet with the Jd, as that will turn additional equity on diamond turns.

I highly recommend this article in the twoplustwo magazine by Andrew Brokos

http://www.twoplustwo.com/magazine/i...ame-theory.php

It may well be one of the many situations where a mixed strategy is actually correct.

Anyway I've ordered the book. I mostly play tournaments though so I'm not sure I'm qualified to give a review.

JJ is literally the first hand that comes to mind if someone asks me "What hands would you check back on this board?"
It's probably only optimal to bet JJ there if you're betting 1/4 pot with about 75% of your range. If you want to make a standard/larger c-bet with a more polar range, JJ clearly does not belong, since it can't go for multiple streets of value. It 'wants' to get to SD in a small pot. (I think KQ might even be a check-back and bluff-catch hand, tbh).

+1

It's arrived now (both the kindle and pdf were supplied).

I know the 2p2 tradition is to post in the review thread when a book arrives before you've read it "reviewing" some minor aspect like how well bound the cover is. As this is an e-book I'll restrict myself to saying I like how the contents hyperlink to the actual chapter or sub-chapter so you can get where you want quicker than with a standard paper book.

Current solvers allow you to both create a "best play" equilibrium given best counter play and also to find best exploits after adjusting for population reads or specific player reads.

I know it's a lot of work but in 2016 I think a book's hand examples to be thorough have to be based on Solver validation of assumptions. Otherwise it become your opinion versus, for example, Mason's in many spots. And logic is not the winner because verbal discussion is incomplete logic. Before solvers we had no choice but now solvers elegantly resolve this debate by forcing assumptions to be operationalized and then solving for unexploitable play.

As a result, solver's are changing our understanding of the interaction of texture/ranges and betsizes quite a bit. These new understandings are already being adopted in actual online play.

IMO a second edition of Carroter's book is needed to at a minimum update key hand examples with solver analysis and validation. (Not all hands but many key hands). Similarly, pre-flop BB defence versus opening ranges and opener's counter-defence is now solved (but not SB or multi-player spots).

Further, I am not sure but I suspect the next generation of solver-trained poker players will not use what will be seen as archaic terms such as value betting or bluffing. I suspect they will intuitively think in terms of their range and be aghast at the very question of how to play a specific hand and instead think coherently about how to play their range and how it interacts intimately with their opponent's range and tendencies. They will certainly be very at home with proving concepts and specific plays in Solvers. And demand it in any educative material. And be very comfortable with complex percentages of indifferent plays as an essential poker skill.

The bar has been forever raised.

I've read the pre-flop section so far. I'm mostly a tournament player so it's maybe the section that's least applicable but I've tried increasing my iso-sizing OOP on the basis of that chapter and it seems to play better like this so its helped me too.

Why would they bother when there is no point unless the game contains another player who not only has a perfect knowledge of GTO ranges and percentages for indifferent plays, but can move away from them to exploit someone making the right move with the wrong frequency by playing an exploitative game of paper-scissors-stone? - something they presumably learn to do against the specific player by observing tens of thousands of hands shown down - newsflash, at all the growing sites people can change screen names so you will never get a big enough sample, even by datamining.

Also, in the context of a book aimed at micro and small stakes, to me this is like saying the future of the school rounds of the maths olympiad is that everyone will be able to divide a 9 figure number by a 4 figure number in 3 seconds. The national round might be like that but in the school round there is never going to be enough people who are physically capable.

I do agree there might be more GTO in the future at these stakes - for example a greater awareness of what is a compulsory bluff with a specific hand for example where it is currently played as a "give-up" by a lot of players.

I can't agree that Carroter's book is like an academic textbook. I guess it depends on your academic discipline. Carroter's book is academic similar to a sociology textbook. But (online) poker ultimately is mathematics and Carroter's book is not at all like a mathematics textbook. It simply lacks the academic rigour expected of maths. Past poker books have had less rigour because the solver tools did not exist but they do now.

My favourite NLHE books are Tipton's Expert HU NLHE V1,2 that narrowly predate commercially available solvers (2012 and 2014). Tipton wrote his own solver and also sold a software course to create a solver to solve the exercises in the book.

Now, to me, Tipton's book is an academic textbook in the mathematics textbook sense.

I expected Carroter's book to be a 2016 6max academic textbook now 4 years more advanced than Tipton's work. I expected the rigour of Tipton's work applied to online 6max cash games. I expected many solver screenshots of solved pre-flop scenarios. I expected exploitative solver screenshots of how, for example turn barreling frequency and hand selection is affected by stationing villain who overcalls and under X/R's flops. Ditto versus aggro villain who over X/R's.

On the contrary, the value of Solvers is not in learning GTO. The value of the Solvers is confidence in one's play, which in turn leads to more hands per hour at a higher EV/hand = more total $. And BTW, equilibrium play (GTO) is much more aggressive than even midstakes players play so Solvers give the confidence to dramatically ramp up aggression.

Pre-solvers, players had to rely on pseudo-science about "charging draws" and checking "showdown value". This pseudo-science knowledge was held by successful players and somewhat kept secret.

Post-solvers, anyone (with logic and computer skills) can know exactly how to play a particular texture given the player positions and player tendencies. No more secrets. Post-solvers anyone can know how to maximally exploit.

This is not necessarily rote learning, it is more about pattern observation. It does take skill but the skill is very different from the skill required of earlier poker generations.

Note that many of these solver conclusions about how we should play against, say, for example, a calling station, differ from conventional pseudo-science poker wisdom. The Solver is right if assumptions are right. The conventional wisdom is provably wrong. The assumptions are transparent and can be disputed, debated and improved.

As regards indifference probabilities, should we bet or check 2nd pair on Q82r versus an unknown? If we know that it is 50/50 we can more quickly make a decision and play more hands per hour.

Anything that causes decision delay while playing can later be analysed in a solver. It is like having a world expert always available to mentor us.

No more secrets. Greater confidence. Higher winnings.

To what extent? I thought pre-flop was only 'solved' for HU up to about 18bb deep and for 184 representative flops or something like that. Maybe I've missed a memo, but I thought that the optimal defence for the BB vs an UTG 3x is still somewhat debatable.

FWIW, I agree with much of what you wrote. My "ideal" poker book would be something that detailed the GTO solution for several spots, and then gave advice on how/why to deviate from it in order to maximally exploit certain player types, known imbalances, or population tendencies.
It seems to me that one of the problems with publishing poker books is that they are often outdated before they even get printed. This isn't really a criticism of Clarke's book though. I belief he "uploaded" as much contemporary/relevant advice as he could.

Well how is that different from knowing a particular hand is 74/26 bet vs. check, or 10/90 or it's just "mixed"?

In the part of your earlier post that I quoted you seemed to suggest that knowing these percentages was an essential skill - or would soon be at the stakes we are discussing.

You can buy solution packs for pre-flop, for deep stacks as well as for 100BB stacks (and shorter stacks) for BB vs SB and for BB vs D for minopen, 2.5BB open, 3BB open and for common progression of 3B/4B/5B betsizes. D R of 35%, 40%, 45%, 50%.

If BB is closing the action and UTG 3x range is known then that is trivial to pre-flop solve, too, for say, 12%, 15%, 18%, 20%, 25% or for other raise sizes. The scripts are the same as above. It just takes server time on a gigantic cloud server with many Gb of RAM because the tree is huge.

What is not solved is if the intervening players act prior to BB as solver theory only works in two player situations.

The pragmatics of how many flop subsets to use is well understood. You don't need that many to get an accuracy that doesn't deviate with more flops added. The flop sets have to satisfy certain mathematical properties. They aren't just random.

The percentages vary by flop texture and range vs range comparison and by some other stuff but the patterns are clear enough to learn. The patterns are important because just because a hand is mixed you can't just say I'll always bet it because the EV of bet and check is the same. If you do that you screw up the rest of the range. All the hands in a range interact to maximise EV for the whole range.

Another example of mixed percentages is in the preflop 3B wars between BB and D, hands like TT, JJ, AJo, even AA are not 100% any action. The pattern is clear. It is not random. Applying this is important because it yields higher EV versus any opponent but especially against a reg that thinks you can't have a certain part of your range when you always can. Against any opponent the ranges are simply more playable post-flop and unexploitable if all-in pre-flop.

Think of poker as more like a mixed martial arts contest where whatever the opponent does he exposes a weakness that you are prepared to exploit maximally all the time. In Zoom games in particular population-read-based exploitive strategy of unknowns is essential.

I think players from earlier generations underestimate how easily talented players from the emerging generation are going to be able to be comfortable with 70% for this and 40% for that and so on endlessly. It is no harder than learning opening ranges or 3B ranges or adjusting for table conditions. It is just another dimension to adjust. But learning it strangles the opponent into submission because there is nothing they can do.

Ok here's one thing I like but with an asterisk.

He rightly points out the position names should be thought of as relative to the button, and the position names are BB, SB, BU, CO, HJ, UTG and that if we are 5-handed the missing position is UTG, because the distinguishing characteristic of that position is that there are 5 left to act. I always want to ask e.g. TV commentators who talk BS about pre-flop, if when they play 3-handed on the button their range is wider because they are "short-handed" or narrower because they are "under the gun" - Carroters knows the important thing is the number of players left to act after us rather than the number of people who have already tossed their cards. Later he says

The term 'UTG' in full-ring poker (9 or 10 handed) carries with it some ideas about being incredibly tight indeed and not having very much fun at all. Fortunately, UTG in 6-max is not such a bleak land of endless folding.

That's right, because UTG with 8 players left to act isn't even the same position as "our" UTG with 5 left to act. The thing is this confusion between UTG in full ring and UTG in 6-max and at short-handed tables is an inevitable result of using a position-name based on counting forward from the BB rather than back from the BU and is actually a reason not to use the term at all. I personally prefer to call this position LJ for lo-jack and call the positions to it's right (in full ring) U7, U8, U9 etc. So the asterisk is using a suboptimal name for that position. Though it's less important in 6-max, there are no decent tournament books so a lot of us read materials aimed at 6-max when that is the closest thing with fairly developed literature and it matters more to us (and also in the text aimed at 6-max grinders it's felt necessary to teach them that UTG just disappears rather than moving round the table as players sit out - which could be shortcutted by renaming the position).

That only applies if villains adjust though. If they play against you as if you had played the correct percentages with the mixed hands then they don't benefit from your not doing so; their EV is the same - if they do adjust then of course you're playing exploit/counter-exploit.

BTW I'm only quoting a small part of your post each time because I more or less agree with all the rest.

I'm about half way through the book so far, and I think it's pretty solid. I've been away from poker from some time, the book has been really good for me to brush up on some concepts and fix some leaks (FWIW I'm only playing NL25z atm)

Guess it might be slightly off-topic, but if I wanted to get started working with solvers, could you perhaps give me some pointers of where I could get started?

start by looking at common flop spots, they're where people make pretty big mistakes frequently, and will give you a better basis for understanding turn/river play.

Makes sense, but is there any specific solver you can recommend?

I use Piosolver and I'm happy, but it's tough to recommend one because I doubt many people have experience with more than one of them.

Don't know if this has been said but is this a typo on the EV of a pretty flop shove at the end of the book. Branch 2 comes out at negative 4.42bb but the figure entered in the parenthesis for overall EV is negative 4.84?

Sent from my MotoG3 using 2+2 Forums

Hi Carroters, I have a question about the c-bet sizing part.

In the book, you suggest that when your range seems very strong, you should c-bet larger. For example when you raise from UTG and a reg defends from the BU, on the KJ6 flop your perceived range hits this flop so hard whereas his range (mostly medium pairs) hits less, so you should bet big for value.

I kinda disagree with this. I think in this spot, the correct way to play is to c-bet more often, but with smaller size. A competent villain will recognize that your range hits the flop, so he will fold very often. Why would you bet big when he will fold a lot? It makes him play correctly (as he will need to defend less often vs your c-bet when you bet big).

FWIW in my experience, in most CCIP spots, a lot of strats have largely similar EV for OOP.

Hi, I am almost done with the book and think it's excellent. One of the best I've ever read. I have understood/agreed with all the hand examples so far but have a question regarding hand 125:

CO - Aggro Reg(26/22/64/59)
BU - Hero Ah4h

CO raises to 2.5bb, Hero calls.
Flop: Jd7h4d
CO bets 4bb, Hero calls 4bb
Turn: Qh
CO bets 12BB, Hero calls 12BB.
River 5c
CO bets 27BB, Hero raises to 81.5BB(all in)

I am not convinced that this is good play. I agree that on the flop we could raise our strong hands but since we only have 6 combos of sets we don't want to cap our calling range vs such an aggro villain even though some turns should put some strong hands in our range. But on the Qh turn shouldn't we be raising our sets and QJ for value given how many bad river cards there are that it should be higher EV than slow playing? So from villain's perspective, on the river when we raise we no longer have the value hands that we are representing and villain's value range that are AQ+ become a clear call given how many busted draws we also have? Also I think this hand has some showdown value against a likely polarized range that we should consider calling or folding instead of raising? So i guess my question is why delaying a raise until the river is better than having a balanced raising range on the turn instead?

Thanks!

Shouldn't it be a 3-bet pre?

I think 3bet pre is standard as well. This was in the chapter on bluff raising the turn and river so what happens pre isn't that important as long as it's not clearly bad. The author says this is a spot where it's fine to wait until the river before having a raising range which is what I don't really understand/agree with. Obviously preflop play dictates our postflop range but say if we have 3-6 sets in our range surely we should be raising them on the turn if we slowed played the flop?