Matthew Hunt's avatar

Matthew Hunt

57 points

Hey Sam,

Great video, loving this series. You mention splitting ranges on the river in the 65s hand - what kind of factors might make you want to split instead of going with one sizing? If both strategies are valid and we have a range that could support either one, what criteria can push us to the point where we think a more complex strategy might have higher EV? Thanks!

May 22, 2017 | 6:43 p.m.

Really outstanding video, the Yerkes-Dodson Law is such a simple but effective concept. Thanks!

May 22, 2017 | 6:23 p.m.

I've long believed that lacking a consistent process for practice is one of the things holding back a ton of players. Thanks for a great video!

May 19, 2017 | 12:29 a.m.

Yeah, that's fair enough. Probably a little ambitious even if HRC does like it.

May 22, 2016 | 11:22 p.m.

Your original comment wasn't exactly phrased in a helpful way. But thanks, I didn't mean to rant about it.

May 21, 2016 | 12:19 a.m.

Well, leaving aside any unnecessary comments on whether I understand game theory on a decent level, of course there are some spots where a solver would use that strategy, but if you're telling me it's possible to correctly execute a strategy where a solver tells you to limp hand X 43% of the time and raise it 57% of the time, then you're wrong. The essence of remaining less exploitable is to simplify our strategy to the point where we can actually successfully execute a pseudo-GTO strategy, rather than failing to execute an unnecessarily complex one. That's why I reduced this spot to a limp-or-raise scenario.

Just for kicks, I ran this through a solver to test your theory and it turns out that HRC suggests a limping range of 53.2% here (which does include A5o with 100% frequency) and a somewhat polar 8.3% 2.25x raising range (which doesn't include A5o at all obviously). There is literally no hand that is in the raising range with 100% frequency, so it doesn't seem like it's feasible to execute that strategy in-game.

Our EV with a mixed strategy is +0.124% of chips in play; with raise-only it's +0.086%, with limp-only it's +0.113%. So if we're talking GTO lines, limping seems better than raising.

May 20, 2016 | 10 p.m.

That's definitely true that it would be useful, but you can replicate those 3-handed late position spots by going to advanced hand, clicking 'next', and then just clicking 'Player XXXX folds' until you get to the player who's acting first in the hand you're looking at. Then you leave their action undefined, and it'll run the spot with the full pot size calculated but assuming that all the previous players fold. You're welcome. :)

May 20, 2016 | 9:48 p.m.

So you're saying to use a mixed strategy of some limping, some raising? That feels pretty exploitable. I suppose it might be fairly easy to execute though.

May 20, 2016 | 6:24 p.m.

Great idea. I think experimenting with HUNL or 3-handed as a blueprint is great since it gives you an idea of how to play in BTN vs SB vs BB spots, which are super important when ranges are so much wider.

May 20, 2016 | 6:23 p.m.

KTcc does have a BDFD and BDSD though - maybe I misread your point?

What sizing would you recommend on flop? Are you just going like 200 with your whole range?

May 20, 2016 | 6:21 p.m.

Equity realisation all depends on the strength of the range overall, usually. If you run a postflop GTO solution then generally what will happen is that a stronger range will usually be able to realise pretty close to 100% of its equity even OOP, while a weaker range will struggle in position and really suffer when OOP. The profitability of any given hand is only a function of the range of which it forms a part.

It would certainly be very cool if HRC allowed that but I don't think it's realistic - not to mention that it would have to cater for the differences in equity realisation between, say, 44 and QJs and A7s. I think the easiest way to do things is to look at the ranges HRC gives us and make an approximate adjustment based on what we know about our ability to realise equity with certain hands. There are plenty of spots where you can lock your own preflop flatting range and let the program decide your 3-betting range, for example.

That's pretty much what you need to do to get your opening ranges, yeah. Although you'll want to disable open-limping, and you can disable things that don't really happen much (like over-flatting 3-bets, flatting 4-bets and flatting 5-bets) if you want to speed up the calculation. If you don't define any ranges for anyone then HRC will simply give you the Nash equilibrium for the spot given the parameters you set. Hope that helps!

May 20, 2016 | 1:36 p.m.

Blinds: t50/t100 (9 Players) SB: 4,030
BB: 4,952
UTG: 10,769
UTG+1: 4,585
MP: 4,752
MP+1: 9,827
MP+2: 3,483
CO: 4,970
BN: 7,494 (Hero)
BB here is RIO coach Setherson2, Seth Davies. Would be interesting to read his comments if he sees this hand. SB is also a pretty aggressive and good reg.
Preflop (150) Hero is BN with K T
6 folds, Hero raises to 253, SB folds, BB calls 153
I'm opening to ~2.5x here to make it harder for the blinds to realise equity since they're both good players. Range is probably a little tighter than usual, maybe in the 35% region. Setherson defends, which I expect him to do quite wide - his 3-betting range here I would think is quite polar, not sure exactly what he will expect of my response ranges so I don't know if he would be making any exploitative adjustments there.
Flop (664) 6 J J
BB checks, Hero bets 312, BB raises to 762, Hero folds
I think my c-bet sizing is fairly straightforward, I'm c-betting most of my range here but not all of it. Do you guys prefer reducing our sizing to allow ourselves to c-bet our entire range? I would be checking back a mixture of some A high hands, some air, and the occasional KK/AA/Jx+ type hand, although I'm aware that it's questionable as to whether I would actually be checking those back in-game even though I say I would be. It's one thing to say your flop checking range is protected and another thing to actually protect it.

When he check-raises, I'm curious about what you guys expect his range to be (or if he's able to provide input, what his range would actually be). I would tend not to have much of a check-raising range on a board like this so it's hard for me to interpret - I don't know whether he'd be more likely to check-raise a value-heavy range here because he expects me to never give him credit for Jx, or whether he's mostly check-raising a bluff-heavy range because he knows how much air I'm c-betting.

When it comes to building a continuing range, I assume all Jx and all 77+ are in there, so what would the next best hands be? I'm guessing stuff like KQs with a backdoor flush draw, maybe some other backdoor FDs that have live cards vs some of his bluffs...I sort of expect his bluffing hands to be overlapping a little bit with the weakest part of my continuing range, so perhaps there's an argument for folding more often since the times where I have the bottom of my continuing range (I thought about continuing with KTcc here) are perhaps going to be the times when he more often has value, but then I do block his KJ/JT here, so...I'm not sure.
Final Pot BB wins 1,288

May 19, 2016 | 8:29 p.m.

Blinds: t300/t600 (9 Players) SB: 8,958
BB: 13,961
UTG: 25,269
UTG+1: 11,312
MP: 7,401
MP+1: 8,469
MP+2: 13,889
CO: 21,601
BN: 17,342 (Hero)
BB is ZeeJustin here. We have no history and he has no read on me, but I expect him to play very well vs everybody obviously. SB is a tight, fishy guy who will likely make some mistakes.
Preflop (900) Hero is BN with 5 A
6 folds, Hero calls 600, SB calls 300, BB raises to 1,800, Hero calls 1,200, SB folds
My thinking was that he would probably play more optimally versus a raising strategy than he would versus a limping strategy since limping the BTN is relatively new in tournaments and he's not an online MTT grinder. I would rarely limp A5o here but the SB was fishy enough that I felt there was an advantage to allowing him into the pot more frequently. Curious what you guys think about a limping strategy here vs raising - obviously I'm also limping strong hands and more playable ones here. Range is probably around 35-40% with no raising range.

ZJ raises and his sizing is smaller than I expected - I feel like with the SB being so capped behind me and very likely to fold a lot, it's sort of difficult to fold here even though I know ZJ is going to put me in some tough spots postflop. I would fold some % of my limping range and limp-shove some frequency too. Thoughts on ZJ's raising range here preflop? How wide?
Flop (4,875) K 6 5
BB bets 1,800, Hero calls 1,800
Don't think I have a raising range on a static flop like this. I have a fair amount of Kx in my range and some 87/97, plus a little bit of 6x and 5x. A5 is probably one of the lowest-EV hands I would call the flop bet with - to a bigger bet I might just fold.
Turn (8,475) K 6 5 K
BB bets 3,000, Hero folds
Some of ZJ's hands pick up equity here and I would expect him to perhaps check-shove Kx some % as well, so I thought about continuing with 5x here since on this particular turn it might have higher EV than some other hands, but I ended up folding.
Final Pot BB wins 8,475

May 19, 2016 | 8:16 p.m.

I don't think there's any way that 3-betting range is accurate for villain. Bare minimum, you need to add in some variable frequencies to almost all the hands - to assume villain is 3-betting all those hands with 100% frequency also assumes they have basically no flatting range on the BTN here, which is a bit of a weird assumption to make from only a 30 hand sample.

I think the likelihood of a random villain 3-betting hands like 66 or A5o here with no read on you (remember that you're probably a random to him as well) when you open UTG is close to zero. I don't play on Bovada but the games would have to be insane for that to be true.

That said, that doesn't necessarily mean this isn't a good 4-bet. The problem is that you've put hands like AJs and TT in their 5-bet jamming range, and then ended up folding to the jam getting pretty decent odds! If you think they're jamming that wide, this is probably a call.

However, I disagree with the logic of "villain might be a bit tighter with his 3-bets here, so we're going to try to account for that by widening his 5-bet shoving range" - this does account for lowering his folding to 4-bet frequency so it does help us calculate the bluff success % in a vacuum, but it doesn't tell us what our actual bluffing frequency should be here, and it also is probably going to end up putting us in spots where we should actually be 4-bet/calling hands like 88 and AJ if villain is 3-betting wide and folding a ton to 4-bets, because if he's folding enough to the 4-bet then it becomes profitable even if we sometimes price ourselves into a marginal call vs his 5-bet.

You can run these spots on HRC fairly straightforwardly - I think this is one where it can probably tell you with more accuracy than we can what your ranges should be here.

May 19, 2016 | 7:31 p.m.

I think running 'advanced hand' scenarios on HRC is a pretty good way to improve. HRC assumes that everyone realises 100% of their equity postflop which seems like it would complicate things a little, but when it comes to opening ranges I think it's reasonable to assume that a stronger player is likely to realise at least 100% of their equity in most situations, particularly in position or against weaker players.

If you run scenarios with all of the flat-calling options enabled and allow HRC to tell you your appropriate opening ranges, you'll find you get a good blueprint for GTO play. I find it useful to construct fictitious scenarios and test what our range should be in each spot - this can be even more useful than using real hands in some instances.

May 19, 2016 | 7:16 p.m.

Re: above comments, thanks man, that's helpful. Appreciate the insight.

May 17, 2016 | 5:42 p.m.

Okay, cool. So let's say we're UTG 9-handed, we're okay with a super-exploitative strategy of raising top 10% or so, open-limping like the next 20% below that, and then folding the rest? Are we mixing at all and occasionally limping hands we might want to limp-3-bet?

May 17, 2016 | 3:40 p.m.

Your Jc blocker is interesting here. Villain is definitely repping super thin because he can't really have A5 or 55. It's basically a few combos of JT (if he doesn't raise flop), a few combos of 22 (if he doesn't raise flop), or maybe like, 4h3h? Any ideas about how wide he peels pre/how wide he 3-bets? A lot depends on his preflop tendencies here.

May 16, 2016 | 10:13 p.m.

Wow, interesting hand. I figure he assumes that his range for peeling a 4-bet OOP here is so heavy on strong Aces that this board is pretty strong for him, making it good for him to have some kind of a leading range.

If that's the case, KxQh is one of the better bluffing hands in his range since it blocks KK and AK plus blocking the nut flush, but it isn't strong versus villain's range for betting this flop for more than one street, and he won't get paid often if turn comes a heart.

He mentions his other bluffs are QQ and JJ with a heart - I imagine his value range is some % of his QhJh/JhTh, 88, some % AK, and some % AA/KK. I actually like his line once he peels the 4-bet.

May 16, 2016 | 10:09 p.m.

Great video! Going to be playing a lot of live cash this summer and I never thought about any of these ideas before.

Question: do you think there's an argument for using a limp-only strategy in early/middle position at extremely passive $1/2 tables where most pots are limped multi-way? I'm thinking about those tables where everyone plays with a 50% VPIP and 5% PFR.

I was thinking that we might be able to open-limp something like 30-35% of hands profitably in a lot of spots given we're so rarely going to be isolated, and that might be better versus a table of bad players than raising a range of 15% and getting called in 3-4 spots every time anyway. It also minimises pot size to keep us from getting raked too much, and minimises preflop variance to an extent also.

May 16, 2016 | 6:44 p.m.

Blinds: t400/t800 (6 Players) SB: 14,564
BB: 54,442 (Hero)
UTG: 43,375
MP: 46,219
CO: 44,641
BN: 36,759
I've been playing a lot of hands, pretty active and doing a little bit of snapfolding in some spots where I made it obvious to villains I was just screwing around. So far it's gotten me paid a couple times. Villain is unknown but seems competent just from his sizings etc, I played a couple BvB pots with him where he seemed to know what he was doing.
Preflop (1,200) Hero is BB with Q Q
UTG raises to 1,600, 3 folds, SB calls 1,200, Hero raises to 5,333, UTG calls 3,733, SB folds
Think my sizing is reasonably good here. SB is fairly splashy and I thought there was a chance he might just flat the 3-bet so I didn't want to go too big.
Flop (12,746) 4 A Q
Hero bets 5,608, UTG calls 5,608
Awesome flop obviously, don't think there's much I can do differently here. With his Ax hands unblocked I think this is a clear bet.
Turn (23,962) 4 A Q 6
Hero bets 10,543, UTG calls 10,543
Not really sure what I was doing if he just shoved on me but I figure I can't really check this turn, he's likely to check behind a lot of Ax and that makes it hard for me to get value on a lot of rivers.
River (45,048) 4 A Q 6 7
Hero checks, UTG bets 16,400, Hero calls 16,400
I tanked this one for a while. In the end the conclusion I came to was that there could only be so many combinations of offsuit hands in his flatting range versus the 3-bet, he would 4-bet AKo preflop, and he probably wouldn't peel turn often with KdQx. I don't know if he would valuebet the Jd or Td on the river and again, there are very few combos including those cards. I think he would likely shove turn with non-nut flushes for protection (and he can't have many of those anyway), so that leaves his only nut flush combos as KdJd and maybe KdTd. I also thought there was a reasonable chance he was enough of a thinking player to turn a hand like AhTh into a bluff here, so considering that he's very unlikely to have more than maybe 4-5 value combos here, he doesn't need to have much in the way of bluffs to make this call okay.

The thing that was leaning me the other direction was that people almost never make it this sizing as a bluff, where they bet like 75% of what their remaining stack is. I really felt like that was stronger, but I just couldn't find many hands in his range strong enough.
Final Pot UTG wins and shows a flush, Ace high.
UTG wins 77,848

Nov. 5, 2015 | 7:35 a.m.

Okay, thanks, felt pretty okay about this one in the end. I'm curious how far we go with our bluffcatching range, though - would we call a hand like 66? That unblocks all of their bluffs, but also all their value (except 76hh if they have it). Interesting.

Oct. 27, 2015 | 12:15 a.m.

I actually disagree that the K is a better card for his range. I don't expect him to check back many King highs at all, so really the only occasions where he hits the K are when he slowplays KK or has AK, maybe he occasionally checks back KQ on the flop too. I, on the other hand, am defending a huge number of combos of Kx. It seems like the K improves me more than it improves him.

I can get on board with not necessarily value-betting any pair, but if we're not, then what's our value-betting threshold on the turn?

I agree about river. He's unlikely to bluff it much so there's no real incentive to check-call, you're right about that.

Oct. 25, 2015 | 4:50 p.m.

Blinds: t250/t500 (9 Players) BN: 19,062
SB: 11,478
BB: 16,292
UTG: 21,229 (Hero)
UTG+1: 5,394
MP: 13,852
MP+1: 24,087
MP+2: 7,026
CO: 16,535
Only 5 hands on villain, no real info. Assume they're a complete random.
Preflop (750) Hero is UTG with Q Q
Hero raises to 1,125, 6 folds, SB calls 875, BB folds
Imagine villains in these tourneys are going flat ranges that are too pair-heavy and too ace-heavy, and just generally too wide. I feel comfortable that there are no traps in there as well.
Flop (3,290) T K 3
SB checks, Hero checks
Think checking here is reasonably straightforward, tough to get value from much when we block heart draws and straight draws. Planning to call at least one future street.
Turn (3,290) T K 3 4
SB bets 3,290, Hero calls 3,290
Alarm bells are kinda ringing here since we block so many of his bluffs but I also feel like since QQ is pretty much the top of our range it's super tough to just instantly fold. I think call and re-evaluate is okay.
River (9,870) T K 3 4 5
SB bets 7,003 and is all in, Hero folds
Here's where I kinda throw up...he's got so few QJ or heart combos to do this with now, but I also block his KQ. I feel awkward folding the top of my range in a spot where I don't know much about villain but I don't feel like I have much of a choice here. Anyone have any other thoughts? I think I might call with QQ that didn't have the Qh.
Final Pot SB wins 9,870

Oct. 24, 2015 | 8:05 p.m.

Blinds: t125/t250 (8 Players) BB: 20,160
UTG: 11,570
UTG+1: 9,771
MP: 7,079
MP+1: 25,324
CO: 3,982
BN: 34,236 (Hero)
SB: 9,649
Villain is 13/8 over 25 hands. No real reads, not sure if he's a reg or anything but he seems pretty tight. Not querying what to do with my actual hand, more just looking at how to construct a river bluffing range.
Preflop (375) Hero is BN with K 9
5 folds, Hero raises to 617, SB calls 492, BB folds
Seems very straightforward so far. I put villain on a fairly pair-heavy range, I imagine they would 3-bet most Ax hands here that they're VPIPing with.
Flop (1,684) A T Q
SB checks, Hero bets 959, SB calls 959
Obviously if villain's range is pair-heavy then this is a flop they're not going to call very much. Once they do call I figure it's not a great spot for future aggression since their range is heavy on Ax/Qx and mine is wide.
Turn (3,602) A T Q 5
SB checks, Hero checks
Pretty much giving up on the pot here. Villain has some flushes now (although not many), and will have some KQ etc that contains a spade and therefore will not fold here as often.
River (3,602) A T Q 5 Q
SB bets 1,512, Hero folds
This is where the idea of bluffing starts to be interesting. Qs blocks a lot of villain's flushes (QsJs, KsQs), and since it's hard for them to have AxKs or KsJs (would have either 3-bet preflop or check-raised flop some %), they're a little bit capped at AxJs here in terms of flushes. They also can't have many FH combos either - AQ or QT would have check-raised the flop some %.

We can't really have many FH combos here ourselves since we would have bet turn with 2 pair or a set, so my question is, would we be value-raising the Ks on the river? If so, which hands are the best bluffs to balance? We have all combos of KsQx/KsKx/KsTx in our range, plus maybe even Ks5x, but I can't figure out how to choose a bluffing range. I guess we need blockers to his bet/call range on the river, so maybe something that blocks the combos of hands that have the Ks, but I can't even find many combos in his range that have the Ks. Perhaps a T blocker to block QT? I'm a little stumped. Maybe someone can help me out.
Final Pot SB wins 3,602

Oct. 24, 2015 | 8 p.m.

Blinds: t150/t300 (9 Players) BB: 18,975
UTG: 10,594
UTG+1: 9,171
MP: 5,127
MP+1: 6,429
MP+2: 27,145
CO: 3,332
BN: 34,975 (Hero)
SB: 11,175
Villain is very active at these stacks (40/32 with 3.0 AF) but only over 25 hands. No specific read, I assume they are a random fish.
Preflop (450) Hero is BN with Q K
6 folds, Hero raises to 741, SB folds, BB calls 441
Very straightforward, nothing to say here.
Flop (1,857) T J 5
BB checks, Hero bets 1,058, BB calls 1,058
Sizing is designed to set up future barrels on good runouts.
Turn (3,973) T J 5 3
BB checks, Hero checks
This feels like one of the worst cards for me to barrel. I was torn between choosing not to barrel because I have no blockers to villain's diamond combos and it's a card that's totally un-scary for the wide flop calling range of a loose villain, versus barrelling here because it sets up a good triple barrel spot against a fairly weak range. The thing that always runs through my mind is that there's a difference between knowing a villain has a wide/weak range, and being able to get them to fold it.
River (3,973) T J 5 3 Q
BB bets 4,200, Hero folds
I can't really figure out what villain's bluffing range could be here. Obviously 98 got there, QJ/QT got there, AQ got there, Q5 got there...villain would have to be turning 5x into a bluff here or check-calling flop with a hand like 87 for me to be good, and I'm not convinced the frequency of that is particularly high.

I do have some better hands in my range here on the river - AK, AQ, QT, some frequency of QJ maybe, perhaps even some Q5 if I bet those on flop - but KQ is obviously beating any bluffs. I'm curious what you guys think of villain's river sizing and whether we're ahead often enough to make the call here.
Final Pot BB wins 3,973

Oct. 24, 2015 | 7:48 p.m.

Blinds: t60/t120 (9 Players) MP: 6,039
UTG+1: 16,312
MP+1: 4,656
MP+2: 1,596
CO: 17,919
BN: 3,268
SB: 1,224
BB: 3,998 (Hero)
UTG: 12,110
Villain here is a pretty good winning regular, runs 21/18 at these stack sizes with 30% RFI and 50% steal. We have a limited amount of history, I expect him to view me as the pretty stationy and not especially tricky reg that I am. Not really concerned with the hand I actually have here, more just with how to choose my overall ranges.
Preflop (180) Hero is BB with Q 9
6 folds, BN raises to 240, SB folds, Hero calls 120
I think this is obviously extremely standard. I expect him to open 40-50% of buttons here and I expect I'm probably defending 50-60% of the time, as well as 3-bet/calling 88+ AJs+ AQo+ and 3-bet bluffing with perhaps a few A and K blocker hands or some trashy suited stuff.
Flop (648) 2 2 3
Hero checks, BN checks
Checking is kind of obvious with our whole range here. I expect villain to check back A9+ with a high frequency, A4-A8 with a slightly lower frequency, AA/KK/33/22 with some frequency, and the occasional JTdd/98ss as well.
Turn (648) 2 2 3 K
Hero checks, BN checks
With villain's range being somewhat capped, am I right in saying our sizing should be larger as a consequence? And does the need for a roughly 1-to-1 ratio of value hands to bluffs hold up in this spot, or is it not applicable?

Let's say we're betting any pair for value. According to Flopzilla we have around 150 combos of value here in that case. 40 combos of flush draws, 52 combos of gutshots and 16 combos of OESDs gives us 108 combos of bluffs. That means we need to find another 42 combos of bluffs from somewhere. What types of hands would you guys use here? Hands containing the Qs or Js that would block his turned AJss/AQss combos? Blockers to Ace-high? I'm not sure what to go with.
River (648) 2 2 3 K T
Hero checks, BN bets 303, Hero folds
River is an easy check-fold with this specific hand, but I run into trouble here because if I'm betting any pair+ for value on turn, then it becomes tough to construct a betting/check-calling range here. Maybe I bet with AT/QT/JT and check-call with T9 or worse? And if I have a betting range like that, which are the best combos to use as bluffs?
Final Pot BN wins 648

Oct. 24, 2015 | 5:24 p.m.

Well, 'reasonably straightforwardly' doesn't mean 'never bluffing this river'. My main question was about the frequency with which we should expect him to either play his bluffs differently by betting the turn, or widen his value range on the river to include 9x hands. If I knew the answer, I wouldn't have posted.

Oct. 1, 2015 | 4:17 p.m.

Comment | Matthew Hunt commented on WSOP MAIN DAY 3

You're probably the only one in the hand who can still have Qx after it checks around on flop, so I think leading turn with the plan to shove most rivers is pretty sound. I don't particularly like leading turn and giving up on river because I think you give up the opportunity to put either one of these guys to a tough decision with a hand like AJ which is very close to the top of either of their ranges.

Check-shoving is okay too, but my concern is that most of the hands that these guys might consider betting on this turn are reasonably unlikely to fold. The kinds of hands you want to be able to get to fold, like 88 or JTs, are much more likely to check back turn, but if you lead turn and river they will fold on a high number of rivers.

Sept. 30, 2015 | 3:32 a.m.

Comment | Matthew Hunt commented on Big 109 value bet

The likely equilibrium on this hand is to check back KQ on the 80th percentile.

Can you elaborate on your process for estimating this, please? I'm not disputing it at all, just curious about how you got to the point of being able to be so specific about the equilibrium here.

Have you just turned yourself into a GTO cyborg by spending hundreds of hours running equilibrium calculations? :)

Sept. 26, 2015 | 11:37 p.m.

Load more uses cookies to give you the best experience. Learn more about our Cookie Policy