Matt Janda Poker Theory Q&A: Wednesday May 22 at 19:00 CET

Originally posted by jules97
Why should we buy your book and who is it aimed at?

I heard about your book from sauces blog and thought I'd pick it up when ti comes out. His canned foreword convinced me

I wrote it aimed at players playing NL$50 to NL$400 (using Stars standards for stakes). It's tough to objectively analyze something I spent so much time on so I'll have to wait for reviews to see if HSNL players, live player, microstakes players, etc. like it too. Sauce liking it was an amazing start that got me very excited.

Originally posted by acetbfish
-How close is the SB's and CO's opening range?

Hard to say exactly, but I'd guess SB should open considerably wider than the CO. This isn't a solvable problem though, I just noticed how wide the BB must defend vs a SB open so I think opening hands like A6o will be profitable (and I would not open that or close to it in the CO). You really have to do some math and play with the ranges to see what you think.

Originally posted by acetbfish
-Could you please explain why our 3bet bluff range should consist of the hands which are too weak to call but too strong to fold? Does it has something to do with the equity we have against villains range?

This is a very old way of thinking which isn't currently recommended. It's discussed why this isn't the best way to approach hands in the book.

It was a rule I think used to be very useful (I loved this rule when someone first taught it to me, and it helped me tremendously), but games are much tougher now and people are much better now that I wouldn't recommend it.

Originally posted by acetbfish
- Since playing in position (especially on the button) will generally earn us more money than playing out of position, why not raise and 3bet bigger IP and smaller OOP?

Because position is more valuable as stacks get deeper, and smaller 3-bets force the opponent to play OOP more frequently and reduce the effectiveness of his 4-bets.

Originally posted by acetbfish
-How playing according to GTO isn't exploitable since our range in some situations (eg being in the SB and calling an open raise from the CO) is pretty clear?

Because the range is balanced. On every flop texture I'll have some strong, weak, and medium strength hands, so even if you know my range there's not much you can do about it.

Originally posted by acetbfish
-Is a HUD really necessary at higher stakes?

Haven't played high stakes, so I wouldn't know first hand. But I imagine it's not, as a very good GTO oriented player wouldn't need a HUD to still win.

Originally posted by acetbfish
-Do you think NL 6max cash will ever be completely solved or the good players will always be able to make money?

I don't think the game will be solved anytime soon, and even if it was, no player could implement the strategy. I think you're better off understanding models and simulations and just trying to emulate GTO, and you don't need to worry about any non-super computer ever implementing it.

Originally posted by MatthewJanda
So for the GTO question where winner gets a book (I believe this is the right place to post it, let me know if I'm wrong) I'd like to ask the following:

When should you consider overbetting, and why do overbets rarely happen on the flop?

-Matt

Overbet for 2 reasons.

1. Bluff when villain will fold too often.

Villain folds too often when he has a weak range and the overbet is believable, i.e. we can represent a strong range.

2. Value when villain will call it too often

Villain calls too often when his range is stronger and hero does not represent a strong range.

Overbets don’t often happen early in a hand because there is less information regarding a villains range than later in the hand.

Originally posted by MatthewJanda
So for the GTO question where winner gets a book (I believe this is the right place to post it, let me know if I'm wrong) I'd like to ask the following:

When should you consider overbetting, and why do overbets rarely happen on the flop?

-Matt

Overbetting when villain's range is capped is effective. It rarely happens on the flop because ranges are usually wide enough to have at least few combos of "nutted" hands on most boards.

EDIT: saw that oblioo said pretty much the same thing, well 1st is 1st

Originally posted by MatthewJanda
So for the GTO question where winner gets a book (I believe this is the right place to post it, let me know if I'm wrong) I'd like to ask the following:

When should you consider overbetting, and why do overbets rarely happen on the flop?

As for the rest of the questions people have already posted, I'll try to start answering them by tomorrow. If anyone else has additional questions that weren't answered today they can post them here as well and I'll try to help.

-Matt

You should consider overbetting the river when your range range is really weak, i.e. you have a lot of bluffs in your range.

Doesn't happen much on the flop bc categorizing hands into values and bluffs before the river is flawed.

Originally posted by MatthewJanda
So for the GTO question where winner gets a book (I believe this is the right place to post it, let me know if I'm wrong) I'd like to ask the following:

When should you consider overbetting, and why do overbets rarely happen on the flop?

As for the rest of the questions people have already posted, I'll try to start answering them by tomorrow. If anyone else has additional questions that weren't answered today they can post them here as well and I'll try to help.

-Matt

Reasons for overbetting:
1. When we have a read that villain's range is capped (full of weak showdown value hands including low PPs and A high) on dry boards, we overbet bluff because they won't continue peeling with these hands without a draw.
2. When we know that villain never folds to flop cbets with hands like weak draws, mid pair and weak top pair and has high WTSD after calling flop and we have TPTK+, we can go for a pot size bet on the turn and overbet on the river to get maximum value from villain's worse hands.
3. When villain's range is capped we can donk overbet bluff shove river when a draw completes and get tons of folds.
4. When we rep a very narrow range after draws missed (for example repping nut full house on 47T7T two tone) we can overbet with the nuts to get some hero calls from 7x or overpairs.
5. When villain's range is polarised, for example on a board with flush and straight draws completed, villain either has relatively weak hands that he will usually fold, a straight or a flush. We can overbet shove nut flush to get calls from straights and weaker flushes.

Point 1. only works on villains who can't hand read well. Vs better players they might suspect a bluff a large amount of the time and call because they know their sdv hand will be good.

Point 2. only works on huge stations who overplay their hands.

Why do overbets rarely happen on the flop?
On the flop the villain's range is widest compared to the turn and river. When we overbet we need sufficient information about villain's range. We have no idea whether villain has a monster on 78T flop or a draw or a pair until we see some action. By overbetting this early we are gambling and thus making unnecessary risks.

Also,
1. If our overbet bluff on the flop failed, we still have to worry about 2 more streets. A mistake on the flop might lead to mistakes on subsequent streets and cause you to spew a lot of chips especially if playing with a deep stack.
2. There is less fold equity overbetting the flop even if we narrowed down villain's range. Strong draws will call due to implied odds(since pot is smaller on the flop) and most strong made hands will call but on the turn most draws fold and more made hands will fold to an overbet because their relative strength is usually weaker on the turn/river than on the flop, and draws now have bad odds.

Originally posted by pleno1

They were answered buddy.

Thanks, missed the first half of coaching glad it was recorded.

All that stuff about villains range and what villain might do has nothing to do with GTO.

Originally posted by gadget51
I'm here because I know very little about GTO play, so please keep the words small.

If I play [close to] GTO versus someone else who plays as well as me, aren't we then just wasting our time and blinds to the rake? Or is that not possible?

Mal.

If two players are equally skilled at GTO, then you'll break even minus the rake. Just like if two players were equally skilled and played using any other style.

Originally posted by oblioo
Could you give one or two practical examples of using multiple opening sizes from a given position (MP, let's say)?

Unfortunately, not really. It's way too difficult to tell how all the positions left to act should respond to a min-raise vs 3x open, and even if we could, there are so many streets left to act there's no way we could tell the different in EV between the two raise sizings.

It's probably possible make a few different bet sizing ranges for a button or SB open though, since there are fewer players left to act. But even then, it'll be difficult to do without a drastic change in the EV of the hand. In other words, if you're opening every hand 2.25 BB in the button, it's not going to change your EV much if you open it for 2 or 2.5BB even if one of those sizings is slightly better (because maybe it makes a few hands which you dominate call, or a few hands which dominate you fold).

Originally posted by spliffstar11
All that stuff about villains range and what villain might do has nothing to do with GTO.

This is hugely important to GTO play, since GTO play assumes the opponent is playing perfectly and thus knows the opponent's ranges and frequencies. So a GTO player will take into account the opponent's preflop range as well as the board texture and figure out if betting, check-raising, check-calling etc is best.

Hm, would you say that a perfectly playing opponent wouldn't be capable of avoiding of capped to bluffcatchers range?

Originally posted by NightFrostaSS
Hm, would you say that a perfectly playing opponent wouldn't be capable of avoiding of capped to bluffcatchers range?

Yup, that happens pretty often. It even happens on the turn.

Say the flop comes 9h 8h 6c and your opponent CB and you're IP. It's pretty hard to justify slowplaying a set or two pair here, since there are so many turn cards which either lose you action or give your opponent the best hand (since most of his "bluffs" have outs to very strong hands). So it's very difficult to argue slowplaying 88 is more profitable than raising it here even if your opponent will sometimes get an awesome turn (like the 2s) and sometimes overbet.

In fact, not overbetting the turn in spots like this I think is one of those leaks that the vast majority of players have whereas in a few years they'll look back and say "I can't believe I used to never overbet here, the spot is perfect for it and the line is super effective against almost anyone."

Originally posted by MatthewJanda

Originally posted by spliffstar11
All that stuff about villains range and what villain might do has nothing to do with GTO.

This is hugely important to GTO play, since GTO play assumes the opponent is playing perfectly and thus knows the opponent's ranges and frequencies. So a GTO player will take into account the opponent's preflop range as well as the board texture and figure out if betting, check-raising, check-calling etc is best.

This ofc is important, but i meant the usual soulreading you see in most handdiscussions.

Also, i understood your question as to when it is best to overbet to be set in a gto scenario. This means we tell villain our strategy and he reacts perfectly against it. If we are close to gto he can at best be at 0ev however. But a lot of lot of the guys answers were based on villains range. If villains range looks like this it best to do this, etc. Well if we know villains range then we don t want to play gto, but i think people overestimate their ability to handread tremendously let alone the fact that we still don't know how villain plays that range.
So basing the answer on villains range is the opposite of what we have to do if we want to answer the question gto-wise. Instead we have to look at our own range.

Hope i explained it somewhat clearly what i meant. English is not my first language. Great to have you here btw.

Originally posted by MatthewJanda
When should you consider overbetting, and why do overbets rarely happen on the flop?

Unfortunately my response turned into an essay so apologies to anyone who doesn't like reading.

I would like to separate my answer into two parts, I shall in a way be attacking the question in reverse order as you shall see:

Part 1)

Before going into reasons for bet-sizing I would like that state that in my view whichever bet sizing we use needs to be a result of our analysis of villains range in any given spot, and is also somewhat an extension of the process we use to construct our own range.

In the natural course of a hand any given players range will start out at its widest, and taper both with each player action, but also in respect to the number of cards available to each player. Let us not forget that on the river everyone can see 5 out of each players 7 possible cards that are used to make a hand and this narrows down the list of best possible hands he or she can have. If I were to make up an example I would give an AAA flop with A on the turn and K on the river. The number of possible 5-card hands it is possible to construct on the board is basically 1.

So using these points I would like to suggest that we will be far more likely to use large (or small) bet-sizes the later we are in the hand and, if i relate this to the question, less likely to use an overbet on flop, than on turn, or river. We see a villains range become smaller as we continue and can thus use more and more specific ranges and bet sizes to make the optimal play. This is not to say that we could not theoretically use multiple sizings with different parts of our range on the flop and still be potentially balanced/GTO, but it is certainly going to be far more difficult for us and thus rarer for us to be able to put in these actions on the relatively undeveloped flop.

I would like to hopefully illustrate my points by saying you are far less likely to see a flop overbet when BU opens and BB calls preflop and you see a dry T52 rainbow flop than you would be if MP opens, calls a SB 3bet and sees a ThJh3c flop. Put simply the second example will have far more defined ranges for each player and I hope from your poker experience you will agree with me on this.

Part 2)

My best shot (and final go) at an answer:

We cannot bet purely for value vs a capped range as an overbet as a perfect opponent can just change a "never folding" decision into a "never calling" one and visa versa with respect to making an overbet purely for fold equity. This is in essence because they are exploitative plays and exploitative play is by definition exploitable.

That said we want to make it as difficult as possible for our opponents to make the correct decisions, so where his range is very capped and narrow yet still strong, (mostly on river), we can still be motivated by value for example and make large overbets with the nuts balancing this out with a purebluff to make his decision more difficult.

When I try to think of an example of this, my mind generally sends me in the direction of having quads or a full house etc where we put villain on a range containing many flush-draws that will become completed. Perhaps something similar to a deceptively played 77 on a 788hh flop, Jh turn.

Now, when villains range is full of weak-middlish bluffcatchers that will call a potsize bet but have a difficult decision against an overbet we can base our range on this initial premise yet balance it with some hands that are likely ahead so that again he does not have an easy decision.

Hand History example of capped villain range:

I shall now try and walk you through an example of where a villains range can become capped through his actions on an appropriate board texture, This villain is an aggressive regular on BB and we are on BU. It is simplest to say effective stacks are 100bb for the time being.

Preflop: Hero Opens "Hand" on BU, Small Blind folds, Big Blind Calls.

- By electing not to 3bet villain has already provided us with the info he is unlikely to have a very strong hand, we can remove TT+ AK, and very likely AQ, AJs perhaps AT as well as combos of other broadways like KQs he may want to 3bet for value.

Flop: A 8 3 . Villain checks, Hero bets, Villain calls.

- Villain further restricts his range by being passive. We know he likes to bluff and semi-bluff a lot so it becomes unlikely he has a set, a two-pair, Top pair better than AJo or ATo and He is aggressive enough that we can start crossing out some backdoor flush-draw type hands from his range also.

Turn: Q . Villain checks, Hero bets, Villain calls.

- It seems that villain caps and restricts his range even further on an even more drawy board than before. If he slowplayed flop its now even more unlikely he slowplays turn.

note- I miss out the betsizing info as we can already make an overbet versus his capped range on this turn.

River: 7 . Villain checks, Hero???

- As I hope I have explained, villains range is capped to weakish top pairs, passively played draws with or without a middle pair. If he were to face an overbet from hero here it might be very nasty for him as our range has all nut combos in it and his most certainly does not, This should be a great spot for our hero.

I'll try and of a range hero might use: If in this position hero decided to bet 2.5xpot for which his opponent to win 42% to break-even (ignoring rake) and he wants to bluff some suited broadway type hands that he barrelled and missed. "KhQh, KhJh, QhJh, KhTh, QhTh, JhTh, Kh9h, Qh9h, Jh9h, Th9h" for example is 10 combos. He has got to the river with the 9 two-pair combos "A8s-A7s, A3s, Q8s". Fairly easy hands to balance with although perhaps we'd prefer to take a different line with them. (AK may be better to balance with)

Anyway, this is almost enough 'value' combos in our range that if we so desired we could balance this range with minimal tweaking to stop the villain ever having a +ev call even with the best hand (AT rather than A7) we predict he might hold. More bluffs than we can make if we just bet potsized

edit: When I come back to this, I want to somehow justify this statement "more tricky for our opponent". It isn't immediately clear to me how I can explain or prove this to be true, but I believe it often, but not always, is. I also would like to mention that just because a decision is harder for our opponent may not mean it improves our expected value. An overbet for the sake of being difficult to play against seems stupid to me at the moment if it is not the maxEV line.
- Edit9000: I think (maybe) I worked it out! Its not that villains decision would become more difficult because that doesn't matter when we are facing an optimal opponent. Its purely to do with the calling success frequencies required for each of our betsizings! Not only do we make more money when villain calls our strong hands, but we can bluff more hands as well! This is not to say making overbets doesn't make it trickier for non-optimal opponents however.

edit 2: I added an example to try and show a villains range becoming capped at the end.

edit 3: Unfortunately I now feel like It is now important for me to work out how bet-sizing for the value portion of our range will correlate to our EV. Starting at a minbet, at first it seems like common sense that we would make more money the larger we bet, but as soon as we bet as large as pot we become very aware the villain will not call as much and thus we reduce our EV. I don't know how to solve this, is there some sort of bell-curve for it or is an over-bet actually going to give us more value.....?

Last edit: So I just saw that Matt is looking over the answers today, I think this means I should stop coming back to try and improve this post now. Hope my longwinded answer was reasonable, I found it very interesting and enjoyable making this and its made me think quite a bit on several different aspects of the game

Conclusion:

With overbets we are targeting clear and/or capped ranges in our opponents and putting them in tricky spots where it is difficult to make the correct decision. Although this difficulty may not matter vs an optimal opponent we can still bluff more hands and remain balanced when we see his range is weak due to our bet-sizing. We can clearly base our initial betting decision on the strength of villains range whether for value or as a bluff, and as mentioned balance it out correctly for the bet-sizing we choose.

The reason overbets rarely happen on the flop is that ranges are generally still wide enough that we cannot yet finely tailor either our range or betsizing to the situation and thus we see such actions only from time to time, normally on the more defined flops where it is a lot more possibly to define a hand-range.

Ok, thats it, no more edits. Thanks Matt for this opportunity to confuse myself

Matthew, I prefer playing at anonymous tables. I am not sure, which basic strategy is better if you never have reads or a history based on some thousand hands with a player: Either you want to play GTO or you still want to play exploitative based on assumptions you have on the player population on a certain level. Will your book give answers to both of those approaches and which one will be more +ev in the longterm?

Originally posted by MatthewJanda

When should you consider overbetting, and why do overbets rarely happen on the flop?

-Matt

We should consider it when the opponent's range is capped and consist mainly of bluff catchers and our range still has nuts in it. It happens rarely on the flop because the ranges are not as defined as on the later streets.

Originally posted by VCRulez
Matthew, I prefer playing at anonymous tables. I am not sure, which basic strategy is better if you never have reads or a history based on some thousand hands with a player: Either you want to play GTO or you still want to play exploitative based on assumptions you have on the player population on a certain level. Will your book give answers to both of those approaches and which one will be more +ev in the longterm?

The best strategy will probably be one that's close to GTO with slight deviations to exploit the opponent. So at small stakes, if people don't raise the flop aggressively enough, then we'll exploit the opponent by folding a bit too much to flop raises (whereas in theory we probably need to float our opponent's raises on some board textures). Likewise maybe we open the button too wide and don't defend aggressively enough against 3-bets. But for the most part, we'll be playing pretty close to GTO, because most hands that work best as bets, checks, calls, folds, etc against a GTO opponent will also work best against a weak opponent.

The book just showcases certain spots you can model to estimate what ranges and frequencies are likely good starting points and hopefully close to GTO. It will hopefully give you the tools necessary to both exploit opponents and play an overall more balanced and profitable game against good opponents.

In the next few days I'll look at the responses you guys gave to my question, as well as answer it myself. A lot of the answers look really good at a glance.

Hey Matthew,

I am a FR-Player and I am struggeling figuring out, what would be a good approximation for the frequencies in the different positions to defend against an Openraise.

For example:

10 handed game UTG opens 3bb

In order to make the UTG indifferent to openraise or not with his worsed hands the other players have to defend combined 33% by 3-betting. This means, everybody has to defend around 5% by 3-betting.

My problem is now:
UTG+1 for example has to fear, that there are 8 persons behind him, that could wake up with strong hands. So there is a good chance, that his worsed 3-betting hands will become -EV. Therefore he should not 3-bet 5% in my opinion and should 3-bet fewer hands. Therefore somebody else has to 3-bet a little bit more. I guess the positions, which are closer to the blinds, because they have position and fewer people left to act or already money in the pot and get a discount.

My question is now:
How can I figure out or get a good approximation of the frequencies in the different positions?

Would be cool, if you could give me a hind, which lead me in the right direction.

Thanks.