The pdf that you got includes all the charts that are available for free around here, so far as I know. More comprehensive sets of GTO preflop ranges are available for sale from a variety of sources, including Peter Clarke's own "carrotcorner" website. (This is where the charts that you saw in episode 17 of his course comes from).
The book Modern Poker Theory by Michael Acevedo also includes charts for the most common preflop spots (including cold-calls, 3-bets and 4-bets) that occur in 6-max NLHM cash games. They are exact GTO solutions assuming 5% rake with a 3bb cap. Some websites allow for access to wide arrays of solved preflop ranges on a subscription basis for training purposes.
March 2, 2021 | 6:19 p.m.
Very instructive from start to finish. This is not my favorite format, in general, but an exception can be made for Peter's videos. His "stream of consciousness" running commentaries are remarkably lucid and articulate. I especially enjoyed the discussion of "air advantage", starting around 33:00, and how considerations of how much air each player has in his range tend to be neglected and must complement considerations regarding the strong and nutty parts of the ranges.
Feb. 21, 2021 | 4 p.m.
In the "Database distribution" window, click on the "check" button, on top, (or on some betting action if OOP has betting options on the flop) to navigate the tree. This moves you to the next node in the tree and the information in the right panel is updated to reflect the IP equity and action frequencies at this new node.
Feb. 15, 2021 | 9:25 p.m.
You most probably are a solidly winning player at 5NL and 10NL. It's unclear if you are or aren't a winning player at 25NL just on the basis of a 15k hands sample. Assuming a standard deviation of 100BB/100, the standard deviation in total winnings is just about the same as the expected total winnings when your true win rate is 8bb/100. This means that if you are break-even at that limit, you have a 16% chance of winning more than a consistent 8bb/100 winner would, on average, and, conversely, if you are a consistent 8bb/100 winner yourself, you have a 16% chance of just breaking even or losing money after just 15,000 hands.
Might there be fundamental leaks in your strategy and/or mental game in spite of your advanced knowledge? Maybe. If you'd like to brush up on the fundamentals, and if you haven't done it already, I would recommend either one or both of (1) Peter Clarke's From the Ground Up course and (2) his newer book "Poker Therapy: Rewire your Mind for Poker" in order to put you back on a productive course.
Feb. 11, 2021 | 11:57 p.m.
This spot somewhat approximates a polar vs bluff-catcher river spot and Brokos's reasoning is spot on. But it's not a pure polar vs bluff-catcher river spot since, as is the case with your specific 88 hand, you do have some nutted hands in your range and, furthermore, if you do lead the turn big as a bluff, Villain has some better hands that may have to fold. (This can't happen in a pure PvBC situation). As a result, when I solve those two scenarios with Pio Solver, leading the river big with 88 on those two runouts isn't a terrible move. It loses some EV compared with checking but it isn't as bad a leading with a small bet. When you lead with a very large bet, Villain still is required to call with marginal holdings to meet a minimum defence frequency (and make you indifferent to bluffing). With a very large river lead, you've effectively turned the tables on him. Nevertheless, checking remains the most profitable option. When you check, Villain only checks back half his range, whereas when you lead big, Villain folds 71% (on Qd7d3h2c8c) to 75% (on Qd7c3h2c9d) of his range at equilibrium.
Your second question raises a very good point. When the front-door flush draw bricks out on the river, Villain has very many busted flush draws in his range. At equilibrium, this doesn't change anything to his bluffing frequency. In both scenarios, my Pio simulations have him betting about 50% of his range in order to make you indifferent to bluff-catching. However, it's true, as you noted, that he is forced to check-back most of his busted flush draws since they are blocking your folding range (that is, your own busted flush draws). The GTO Villain, therefore, has to find other less-natural looking bluffs in his range in order to remain balanced. If actual villains are unable to find those other bluffs, they may have a tendency to under-bluff such a spot, and you can safely overfold your bluff-catchers exploitatively against them.
Feb. 11, 2021 | 6:54 p.m.
Great video. Always great seeing bullies who normally bully me being bullied. (On a related topic, see the Wikipedia page on "Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo".)
Feb. 9, 2021 | 10:02 p.m.
Hi Peter, outstanding video, as usual! No idea if you still read new comments about months old videos... This is just a nit-pick regarding the interpretation of the EVs of various options in a solver output. (41:40) You may have already known this but have kind of forgotten: GTO strategies only are mixed when the EVs of all the options with non-zero frequency are strictly identical. (Hence, it is never GTO to take an action with any finite (non-zero) frequency at all when there exists another action that has an higher EV at this exact node). It's only because we're looking at solutions that are solved with limited accuracy (and hence are still exploitable) that different options in a mixed strategy have slightly different EVs.
When we proceed to solve with an ever higher accuracy, either one of two things is guaranteed to happen: (1) The frequency of the options with lower EV converges towards zero, or (2) the different options remain mixed but their EV-differences tend towards zero. Hence the EV difference visualization tool in Pio must mainly be used to spot the hands with pure strategies (or that are just missing some of the allowable actions) that would lose the most to a simplification for the whole range that would force them to take an option not actually taken in the more complex mixed GTO strategy. The smallish EV differences of the truly mixed options, by contrast, merely testify to the lack of accuracy of the solve.
Feb. 9, 2021 | 9:53 p.m.
Yes, seeing you continuing with the same player pool would be nice. Hopefully, you can demonstrate practically how you are putting together matching puzzle pieces. But even if the opportunity for doing exactly that doesn't arise, because the same players don't alway show up or we don't very often butt heads with then in relevantly similar spots, it might still be extremely valuable just to hear you explain how, based on your past experience, you are usually fitting those specific kinds of puzzle pieces.
Feb. 6, 2021 | 9:12 p.m.
Tip #4 Pay mostly attention to the first three tips.
Tip #5 Disregard Tip #4
Tip #6 Disregard Tip #5 (and hence pay mostly attention to the first three tips).
Tip #7 Disregard Tip #6
Tip #8 Disregard Tip #7 (and hence pay mostly attention to the first three tips).
Tip #9 Disregard Tip #8
Tip #10 Disregard Tip #9 (and hence pay mostly attention to the first three tips).
Feb. 6, 2021 | 5:46 p.m.
Another issue, in addition to what CrappyTimeSlot just said, is that solvers know how to correctly split their flush draw combos between their passive action vs active action ranges (i.e. checking-back vs betting, or calling vs raising, or checking vs betting). Human players, by contrast, especially at low- to mid-stakes, tend to favor the active line with their flush draws since they believe they must nearly always semi-bluff. So, when they do take a passive line, this removes a lot of flush draws from their range and your blockers to flush draws are much less relevant. This means that your hand isn't necessarily a bad one to bluff with in spite of its flush draw blockers.
Feb. 6, 2021 | 3:47 a.m.
Hey, so glad to see you back producing for RIO, Uri! I had already watched all of your videos -- all three of them -- and only wished there were more! People who also follow your YouTube channel will understand you haven't been lazy.
This newer video of yours has great value. It is true that, as you mentioned, opportunities for taking notes didn't come up as often as you'd wished, but you explained very well your note-taking philosophy and you rationale for taking them. And there were no downtimes since you provided so many good strategic advices along the way. Hopefully, this is the first one of a two or three part series.
Feb. 5, 2021 | 7:16 p.m.
It's a very good idea to take a short break. Maybe Peter Clarke's new book "Poker Therapy" is what you need right now. I haven't read it but I can guess some of its content from the occasional comments on the mental games of poker that he sprinkles his instructional videos with. One of his best advices, in addition never to be result oriented as you play, is to rejoice whenever you lose a hand to a player who played badly and got lucky. It is testimony to the fact that the games still are easily beatable and your prospects for the future remain good provided only you keep working on your game.
Jan. 29, 2021 | 3:10 a.m.
I'd like to see the range you gave to MP. Does he have some AA traps and some 76s?
I also did those two sims, guessing MP's range, and I get somewhat similar results. Frequencies of large-bet/small-bet/check on Q76r are 15%/18%/67%, and on Q72r they are 23%/62%/15%.
MP's nut advantage on Q76r, and our range advantage on Q72r, explain why we can range bet small on the latter and why we must polarize on the former (while protecting our check-back range with everything except 77), I guess.
Jan. 25, 2021 | 2:44 a.m.
It very much depends on the complexity of the game tree and how much precision you aim at (as a percentage of the pot). I have a ryzen 3900x CPU with 12 cores, but since I am using PioSOLVER basic (rather than pro or edge), only 6 CPU threads are being used. Simple trees with tight ranges can take as little as 15 seconds to solve. More complex trees defined with several bet sizes on many streets, and wide ranges, can take 10 minutes or more to solve. But then, when you define fairly complex trees, you can also exceed memory requirements. (I have 64GB ram, so seldom hit the limit).
You can also consider using GTO+ that has most of Pio's features, is much cheaper, has a faster solving algorithm than Pio, and uses all of your computer's threads. GTO+ also interfaces beautifully with Flopzilla Pro. I still use Pio the most mainly because of my greater familiarity with it and the very clear way in which the interface displays ranges and frequencies for individual hands.
Jan. 23, 2021 | 6:50 p.m.
Just wanted to mention: good thinking on your part to notice that your blocking the missed backdoor flush draw doesn't have much relevance since he frequently would have bet it on the turn. I just recently had my attention drawn to this concept at 26:50 in Peter Clarke's video "$50nl 6-Max Zoom Session Review (Part 2)". If Villain had bet the turn and barreled the river again, it would be very significant that we have two heart blockers, but when the turn checks through, it's indeed much less relevant.
It may actually be beneficial for us since we are blocking part of his own bluff-catching range. Since missed backdoor heart draws are part of our own range, the ace of heart would be a good card for him to bluff-catch with, but he can't have it.
Jan. 23, 2021 | 1:51 a.m.
You raise many good points. I decided to do my own solves (with Pio) of all three of those flops in this SB v BU 3-bet spot. I was surprised that, indeed, the JT7ss is the one that most favors the SB, equity-wise and EV-wise. Some of Peter's remarks remain true, though, especially as regard the nuts advantage of the SB on the 744r flop. Here are some of my results with a few comments to follow:
744r, two bet sizes (66% and 33%) and check
OOP EV 133.68; OOP Equity 54.76
OOP Strategy on flop: 51% 66, 26% 33, 23% check
When SB bets 33, BU is supposed to call 66% raise 17% and fold 18%
OOP EV 130.97
OOP EV 130.67
When SB bets 33, BU is supposed to call 85.6%, raise 6.8% and fold 7.6% !
OOP EV 123.54
JT7ss, two bet sizes (66% and 33%) and check
OOP EV 136.53; OOP Equity 57.344
OOP strategy on flop: 55% 66, 17% 33, 27% check
When SB bets 33, BU is supposed to call 62% raise 13% and fold 25%
OOP EV 134.05
When SB range-bets-33, BU is supposed to call 75%, raise 5% and fold 20%
A87r , two bet sizes (66% and 33%) and check
EV 134.78, OOP Equity 56.47
OOP strategy on flop: 21% 66 , 69% 33 and 10% check
When SB bets 33, BU is supposed to call 68%, raise 4% and fold 27%
OOP EV 133.45
So, it appears that Pio likes to polarize the flop c-betting strategy the most on JT7ss, where it likes the small betting option the least. It comes the closest to the range betting strategy on the A87r flop, where it most frequently chooses the small betting option. Finally, although it doesn't favor range-betting on 744r, it nevertheless checks this flop at a lower frequency (23%) than it does on the most favorable JT7ss flop (27%)
I think what enables the SB to bet so very frequently on 744r, in spite of the fact that this flop does not favor the SB's range nearly as much as JT7ss, may be the fact that, as Peter notices, the SB has such a high number of undiluted nutty combos in the form of big over-pairs. If SB plays an optimal mixed strategy with two bet sizes and checks, BU is supposed to fold very few missed combos and find a fairly large number of raises and calls, although those calls and raises make sense for the most part in terms of equity and backdoor nutty potential. The BU optimal response when SB range-bets-33 is much more counterintuitive since he is supposed call almost everything (85.6%) and fold or raise almost nothing. (See picture below). The only pure fold in BU's range is QJo! And the only hands that raise at any significant frequency are 99, TT, A8s, A7s, A2s and A3s (some of them preferably hearts, with no backdoor flush potential!)
So, the fact that 744r make it so hard for BU to properly respond to SB's range betting strategy, and the fact that even at equilibrium, a mixed strategy doesn't require much polarization, may justify Peter's advice on partially exploitative grounds.
Jan. 22, 2021 | 11:25 p.m.
After the preflop action in this single raised 3-way pot, in a live setting, none of the ranges are especially strong. Theoretically, the SB's range may be the strongest. UTG's range when he limps and then just calls after SB called your raise, remains fairly wide when playing that deep. When SB shoves the flop with his very shallow remaining stack, he may very well be protecting any overpair from 55 to TT (and even JJ-KK that live players often like to trap with) or semi-bluffing some straight draw. For sure, he also has sets. Regardless of your chances of winning the main pot, you certainly want to over-call and extract as much value as possible from UTG+1, with whom you are very deep and who you have position on. If he check/raises you at some point on the turn or river, and thereby shows willingness to bring stacks in, you can consider folding, but not before then!
Jan. 21, 2021 | 7:07 p.m.
May he not also play KQ or KJ this way? Also, isn't he likely to bet larger on the turn and river with flopped sets with the hope of stacking you? (Even if he's trying to be balanced, he may have more than one bet size and put sets plus some bluffs in the large one; and TPGK types of hands plus some bluffs in the smaller one). Your call doesn't seem bad to me, but I may be wrong.
Jan. 20, 2021 | 1:42 a.m.
Figuring out exactly the math of what those calling or betting frequencies ought to be can't be any easier than computing the Nash equilibria (GTO strategies). The algorithms for approximating those aren't very complex (although they require many iterations) but they don't reduce to simple formulas except for the case of some simplified toy games. At best we can find some heuristics, or simplified rules, for determining how many bluffs, or semi-bluffs, or bluff-catchers, we can have on the flop or turn, and what hands are the best candidates because of their implied odds, blockers, need for protection, etc.
I'm pretty sure when strong players give advice about over-calling or over-bluffing (or under-bluffing) they don't base that on mathematical formulas but rather on their experience toying with solvers and anticipating how their range constructions will carry them to reasonably well understood and favorable river spots. Ben Sulsky and Uri Peleg (among others) have some videos that explain how to use solvers efficiently to build such intuitions. If your subscription is 'Essential', look at the videos of Peter Clarke and Qing Yang. For the theory, you can delve into Andrew Brokos "Plays Optimal Poker 1 & 2", or Will Tipton's "Expert Heads Up No Limit Hold'em 1 & 2". In both of those you will find formulas that apply to more complex toy games giving insight into real hold'em poker situation if that's what you're looking for.
Jan. 18, 2021 | 4:44 a.m.
There isn't any useful minimum defense frequency on the flop or turn since range interactions with the board evolve when more community cards are dealt. Range and nuts advantage shift or get equalized, and there is no useful approximation to a multi-street static polar v bluff catcher situation. This is why solver study is useful for trying to devise simplifying heuristics applicable to broadly characterized spots : e.g. how to double barrel on paired board in single raised pots in position against loose callers, etc. A short and enlightening explanation why the minimum defense frequency formula is very limited in its applications apart from some specific river spots (or some toy games with numbers, with no card removal effects) has been given by punter11235 (who is a developer of Pio Solver) on the 2+2 support thread for the software:
Jan. 17, 2021 | 6:36 p.m.
What lIlCitanul said.
I watched the video by Doug Polk that you referenced. I'm not sure exactly why Polk thought that he should bluff with 4 or 5 combos in this river spot. However, his main concern (at this point in his explanation) was to avoid ending up at the river with a range that contains many more bluffs than he can profitably bet. In that case, he would have to give up on most of them (and hence waste his turn bet) or, else, bet them all and potentially become very exploitable. So, the bet to pot ratio on this turn isn't the relevant consideration. Instead, Polk is concerned with his anticipated range composition on the river. It's on the river, when the ranges of both players approximate closely enough an ideal "polar versus bluff catcher" situation, that the minimum defense frequency formula can be applied to determining the proportion of your range that you can bluff while staying balanced (and hence, unexploitable).
Jan. 17, 2021 | 4:49 p.m.
Zenith Poker has a bunch for free (and some preflop too). You must register but that's free also. I grabbed them because I'm a hoarder but I don't use them much since they must be viewed with MonkerViewer and I found this software a little crude and cumbersome to use. Also, I find it profitable to solve spots individually while playing with the game tree parameters to figure out what's important and what simplifications preserve EV.